Datasets:

sswt

/

arxiv_sample_50K

like 0

CAMPARI CameraAware Decomposed Generative Neural Radiance Fields ; Tremendous progress in deep generative models has led to photorealistic image synthesis. While achieving compelling results, most approaches operate in the twodimensional image domain, ignoring the threedimensional nature of our world. Several recent works therefore propose generative models which are 3Daware, i.e., scenes are modeled in 3D and then rendered differentiably to the image plane. This leads to impressive 3D consistency, but incorporating such a bias comes at a price the camera needs to be modeled as well. Current approaches assume fixed intrinsics and a predefined prior over camera pose ranges. As a result, parameter tuning is typically required for realworld data, and results degrade if the data distribution is not matched. Our key hypothesis is that learning a camera generator jointly with the image generator leads to a more principled approach to 3Daware image synthesis. Further, we propose to decompose the scene into a background and foreground model, leading to more efficient and disentangled scene representations. While training from raw, unposed image collections, we learn a 3D and cameraaware generative model which faithfully recovers not only the image but also the camera data distribution. At test time, our model generates images with explicit control over the camera as well as the shape and appearance of the scene.

OodGAN Generative Adversarial Network for OutofDomain Data Generation ; Detecting an OutofDomain OOD utterance is crucial for a robust dialog system. Most dialog systems are trained on a pool of annotated OOD data to achieve this goal. However, collecting the annotated OOD data for a given domain is an expensive process. To mitigate this issue, previous works have proposed generative adversarial networks GAN based models to generate OOD data for a given domain automatically. However, these proposed models do not work directly with the text. They work with the text's latent space instead, enforcing these models to include components responsible for encoding text into latent space and decoding it back, such as autoencoder. These components increase the model complexity, making it difficult to train. We propose OodGAN, a sequential generative adversarial network SeqGAN based model for OOD data generation. Our proposed model works directly on the text and hence eliminates the need to include an autoencoder. OOD data generated using OodGAN model outperforms stateoftheart in OOD detection metrics for ROSTD 67 relative improvement in FPR 0.95 and OSQ datasets 28 relative improvement in FPR 0.95 Zheng et al., 2020.

A Graph VAE and Graph Transformer Approach to Generating Molecular Graphs ; We propose a combination of a variational autoencoder and a transformer based model which fully utilises graph convolutional and graph pooling layers to operate directly on graphs. The transformer model implements a novel node encoding layer, replacing the position encoding typically used in transformers, to create a transformer with no position information that operates on graphs, encoding adjacent node properties into the edge generation process. The proposed model builds on graph generative work operating on graphs with edge features, creating a model that offers improved scalability with the number of nodes in a graph. In addition, our model is capable of learning a disentangled, interpretable latent space that represents graph properties through a mapping between latent variables and graph properties. In experiments we chose a benchmark task of molecular generation, given the importance of both generated node and edge features. Using the QM9 dataset we demonstrate that our model performs strongly across the task of generating valid, unique and novel molecules. Finally, we demonstrate that the model is interpretable by generating molecules controlled by molecular properties, and we then analyse and visualise the learned latent representation.

A Tunable Model for Graph Generation Using LSTM and Conditional VAE ; With the development of graph applications, generative models for graphs have been more crucial. Classically, stochastic models that generate graphs with a predefined probability of edges and nodes have been studied. Recently, some models that reproduce the structural features of graphs by learning from actual graph data using machine learning have been studied. However, in these conventional studies based on machine learning, structural features of graphs can be learned from data, but it is not possible to tune features and generate graphs with specific features. In this paper, we propose a generative model that can tune specific features, while learning structural features of a graph from data. With a dataset of graphs with various features generated by a stochastic model, we confirm that our model can generate a graph with specific features.

DeepCAD A Deep Generative Network for ComputerAided Design Models ; Deep generative models of 3D shapes have received a great deal of research interest. Yet, almost all of them generate discrete shape representations, such as voxels, point clouds, and polygon meshes. We present the first 3D generative model for a drastically different shape representation describing a shape as a sequence of computeraided design CAD operations. Unlike meshes and point clouds, CAD models encode the user creation process of 3D shapes, widely used in numerous industrial and engineering design tasks. However, the sequential and irregular structure of CAD operations poses significant challenges for existing 3D generative models. Drawing an analogy between CAD operations and natural language, we propose a CAD generative network based on the Transformer. We demonstrate the performance of our model for both shape autoencoding and random shape generation. To train our network, we create a new CAD dataset consisting of 178,238 models and their CAD construction sequences. We have made this dataset publicly available to promote future research on this topic.

Learning from Perturbations Diverse and Informative Dialogue Generation with Inverse Adversarial Training ; In this paper, we propose Inverse Adversarial Training IAT algorithm for training neural dialogue systems to avoid generic responses and model dialogue history better. In contrast to standard adversarial training algorithms, IAT encourages the model to be sensitive to the perturbation in the dialogue history and therefore learning from perturbations. By giving higher rewards for responses whose output probability reduces more significantly when dialogue history is perturbed, the model is encouraged to generate more diverse and consistent responses. By penalizing the model when generating the same response given perturbed dialogue history, the model is forced to better capture dialogue history and generate more informative responses. Experimental results on two benchmark datasets show that our approach can better model dialogue history and generate more diverse and consistent responses. In addition, we point out a problem of the widely used maximum mutual information MMI based methods for improving the diversity of dialogue response generation models and demonstrate it empirically.

Reinforced Generative Adversarial Network for Abstractive Text Summarization ; Sequencetosequence models provide a viable new approach to generative summarization, allowing models that are no longer limited to simply selecting and recombining sentences from the original text. However, these models have three drawbacks their grasp of the details of the original text is often inaccurate, and the text generated by such models often has repetitions, while it is difficult to handle words that are beyond the word list. In this paper, we propose a new architecture that combines reinforcement learning and adversarial generative networks to enhance the sequencetosequence attention model. First, we use a hybrid pointergenerator network that copies words directly from the source text, contributing to accurate reproduction of information without sacrificing the ability of generators to generate new words. Second, we use both intratemporal and intradecoder attention to penalize summarized content and thus discourage repetition. We apply our model to our own proposed COVID19 paper title summarization task and achieve close approximations to the current model on ROUEG, while bringing better readability.

Evaluation Metrics for Graph Generative Models Problems, Pitfalls, and Practical Solutions ; Graph generative models are a highly active branch of machine learning. Given the steady development of new models of everincreasing complexity, it is necessary to provide a principled way to evaluate and compare them. In this paper, we enumerate the desirable criteria for such a comparison metric and provide an overview of the status quo of graph generative model comparison in use today, which predominantly relies on the maximum mean discrepancy MMD. We perform a systematic evaluation of MMD in the context of graph generative model comparison, highlighting some of the challenges and pitfalls researchers inadvertently may encounter. After conducting a thorough analysis of the behaviour of MMD on syntheticallygenerated perturbed graphs as well as on recentlyproposed graph generative models, we are able to provide a suitable procedure to mitigate these challenges and pitfalls. We aggregate our findings into a list of practical recommendations for researchers to use when evaluating graph generative models.

DualTeacher ClassIncremental Learning With DataFree Generative Replay ; This paper proposes two novel knowledge transfer techniques for classincremental learning CIL. First, we propose datafree generative replay DFGR to mitigate catastrophic forgetting in CIL by using synthetic samples from a generative model. In the conventional generative replay, the generative model is pretrained for old data and shared in extra memory for later incremental learning. In our proposed DFGR, we train a generative model from scratch without using any training data, based on the pretrained classification model from the past, so we curtail the cost of sharing pretrained generative models. Second, we introduce dualteacher information distillation DTID for knowledge distillation from two teachers to one student. In CIL, we use DTID to learn new classes incrementally based on the pretrained model for old classes and another model pretrained on the new data for new classes. We implemented the proposed schemes on top of one of the stateoftheart CIL methods and showed the performance improvement on CIFAR100 and ImageNet datasets.

Teach Me What to Say and I Will Learn What to Pick Unsupervised Knowledge Selection Through Response Generation with Pretrained Generative Models ; Knowledge Grounded Conversation Models KGCM are usually based on a selectionretrieval module and a generation module, trained separately or simultaneously, with or without having access to a gold knowledge option. With the introduction of large pretrained generative models, the selection and generation part have become more and more entangled, shifting the focus towards enhancing knowledge incorporation from multiple sources instead of trying to pick the best knowledge option. These approaches however depend on knowledge labels andor a separate dense retriever for their best performance. In this work we study the unsupervised selection abilities of pretrained generative models e.g. BART and show that by adding a scoreandaggregate module between encoder and decoder, they are capable of learning to pick the proper knowledge through minimising the language modelling loss i.e. without having access to knowledge labels. Trained as such, our model KMine shows competitive selection and generation performance against models that benefit from knowledge labels andor separate dense retriever.

How much do language models copy from their training data Evaluating linguistic novelty in text generation using RAVEN ; Current language models can generate highquality text. Are they simply copying text they have seen before, or have they learned generalizable linguistic abstractions To tease apart these possibilities, we introduce RAVEN, a suite of analyses for assessing the novelty of generated text, focusing on sequential structure ngrams and syntactic structure. We apply these analyses to four neural language models an LSTM, a Transformer, TransformerXL, and GPT2. For local structure e.g., individual dependencies modelgenerated text is substantially less novel than our baseline of humangenerated text from each model's test set. For largerscale structure e.g., overall sentence structure modelgenerated text is as novel or even more novel than the humangenerated baseline, but models still sometimes copy substantially, in some cases duplicating passages over 1,000 words long from the training set. We also perform extensive manual analysis showing that GPT2's novel text is usually wellformed morphologically and syntactically but has reasonably frequent semantic issues e.g., being selfcontradictory.

Controlling Conditional Language Models without Catastrophic Forgetting ; Machine learning is shifting towards generalpurpose pretrained generative models, trained in a selfsupervised manner on large amounts of data, which can then be applied to solve a large number of tasks. However, due to their generic training methodology, these models often fail to meet some of the downstream requirements e.g., hallucinations in abstractive summarization or style violations in code generation. This raises the important question of how to adapt pretrained generative models to meet all requirements without destroying their general capabilities catastrophic forgetting. Recent work has proposed to solve this problem by representing taskspecific requirements through energybased models EBMs and approximating these EBMs using distributional policy gradients DPG. Despite its effectiveness, this approach is however limited to unconditional distributions. In this paper, we extend DPG to conditional tasks by proposing Conditional DPG CDPG. We evaluate CDPG on four different control objectives across three tasks translation, summarization and code generation and two pretrained models T5 and GPTNeo. Our results show that finetuning using CDPG robustly moves these pretrained models closer towards meeting control objectives and in contrast with baseline approaches does not result in catastrophic forgetting.

A survey of multimodal deep generative models ; Multimodal learning is a framework for building models that make predictions based on different types of modalities. Important challenges in multimodal learning are the inference of shared representations from arbitrary modalities and crossmodal generation via these representations; however, achieving this requires taking the heterogeneous nature of multimodal data into account. In recent years, deep generative models, i.e., generative models in which distributions are parameterized by deep neural networks, have attracted much attention, especially variational autoencoders, which are suitable for accomplishing the above challenges because they can consider heterogeneity and infer good representations of data. Therefore, various multimodal generative models based on variational autoencoders, called multimodal deep generative models, have been proposed in recent years. In this paper, we provide a categorized survey of studies on multimodal deep generative models.

Analog Bits Generating Discrete Data using Diffusion Models with SelfConditioning ; We present Bit Diffusion a simple and generic approach for generating discrete data with continuous state and continuous time diffusion models. The main idea behind our approach is to first represent the discrete data as binary bits, and then train a continuous diffusion model to model these bits as real numbers which we call analog bits. To generate samples, the model first generates the analog bits, which are then thresholded to obtain the bits that represent the discrete variables. We further propose two simple techniques, namely SelfConditioning and Asymmetric Time Intervals, which lead to a significant improvement in sample quality. Despite its simplicity, the proposed approach can achieve strong performance in both discrete image generation and image captioning tasks. For discrete image generation, we significantly improve previous stateoftheart on both CIFAR10 which has 3K discrete 8bit tokens and ImageNet64x64 which has 12K discrete 8bit tokens, outperforming the best autoregressive model in both sample quality measured by FID and efficiency. For image captioning on MSCOCO dataset, our approach achieves competitive results compared to autoregressive models.

Probabilistic Generative Transformer Language models for Generative Design of Molecules ; Selfsupervised neural language models have recently found wide applications in generative design of organic molecules and protein sequences as well as representation learning for downstream structure classification and functional prediction. However, most of the existing deep learning models for molecule design usually require a big dataset and have a blackbox architecture, which makes it difficult to interpret their design logic. Here we propose Generative Molecular Transformer GMTransformer, a probabilistic neural network model for generative design of molecules. Our model is built on the blank filling language model originally developed for text processing, which has demonstrated unique advantages in learning the molecules grammars with highquality generation, interpretability, and data efficiency. Benchmarked on the MOSES datasets, our models achieve high novelty and Scaf compared to other baselines. The probabilistic generation steps have the potential in tinkering molecule design due to their capability of recommending how to modify existing molecules with explanation, guided by the learned implicit molecule chemistry. The source code and datasets can be accessed freely at httpsgithub.comusccolumbiaGMTransformer

A Generative Approach for ProductionAware Industrial Network Traffic Modeling ; The new wave of digitization induced by Industry 4.0 calls for ubiquitous and reliable connectivity to perform and automate industrial operations. 5G networks can afford the extreme requirements of heterogeneous vertical applications, but the lack of real data and realistic traffic statistics poses many challenges for the optimization and configuration of the network for industrial environments. In this paper, we investigate the network traffic data generated from a laser cutting machine deployed in a Trumpf factory in Germany. We analyze the traffic statistics, capture the dependencies between the internal states of the machine, and model the network traffic as a production state dependent stochastic process. The twostep model is proposed as follows first, we model the production process as a multistate semiMarkov process, then we learn the conditional distributions of the production state dependent packet interarrival time and packet size with generative models. We compare the performance of various generative models including variational autoencoder VAE, conditional variational autoencoder CVAE, and generative adversarial network GAN. The numerical results show a good approximation of the traffic arrival statistics depending on the production state. Among all generative models, CVAE provides in general the best performance in terms of the smallest KullbackLeibler divergence.

ArchiSound Audio Generation with Diffusion ; The recent surge in popularity of diffusion models for image generation has brought new attention to the potential of these models in other areas of media generation. One area that has yet to be fully explored is the application of diffusion models to audio generation. Audio generation requires an understanding of multiple aspects, such as the temporal dimension, long term structure, multiple layers of overlapping sounds, and the nuances that only trained listeners can detect. In this work, we investigate the potential of diffusion models for audio generation. We propose a set of models to tackle multiple aspects, including a new method for textconditional latent audio diffusion with stacked 1D UNets, that can generate multiple minutes of music from a textual description. For each model, we make an effort to maintain reasonable inference speed, targeting realtime on a single consumer GPU. In addition to trained models, we provide a collection of open source libraries with the hope of simplifying future work in the field. Samples can be found at httpsbit.lyaudiodiffusion. Codes are at httpsgithub.comarchinetaiaudiodiffusionpytorch.

In What Languages are Generative Language Models the Most Formal Analyzing Formality Distribution across Languages ; Multilingual generative language models LMs are increasingly fluent in a large variety of languages. Trained on the concatenation of corpora in multiple languages, they enable powerful transfer from highresource languages to lowresource ones. However, it is still unknown what cultural biases are induced in the predictions of these models. In this work, we focus on one language property highly influenced by culture formality. We analyze the formality distributions of XGLM and BLOOM's predictions, two popular generative multilingual language models, in 5 languages. We classify 1,200 generations per language as formal, informal, or incohesive and measure the impact of the prompt formality on the predictions. Overall, we observe a diversity of behaviors across the models and languages. For instance, XGLM generates informal text in Arabic and Bengali when conditioned with informal prompts, much more than BLOOM. In addition, even though both models are highly biased toward the formal style when prompted neutrally, we find that the models generate a significant amount of informal predictions even when prompted with formal text. We release with this work 6,000 annotated samples, paving the way for future work on the formality of generative multilingual LMs.

Human Preference Score Better Aligning TexttoImage Models with Human Preference ; Recent years have witnessed a rapid growth of deep generative models, with texttoimage models gaining significant attention from the public. However, existing models often generate images that do not align well with human preferences, such as awkward combinations of limbs and facial expressions. To address this issue, we collect a dataset of human choices on generated images from the Stable Foundation Discord channel. Our experiments demonstrate that current evaluation metrics for generative models do not correlate well with human choices. Thus, we train a human preference classifier with the collected dataset and derive a Human Preference Score HPS based on the classifier. Using HPS, we propose a simple yet effective method to adapt Stable Diffusion to better align with human preferences. Our experiments show that HPS outperforms CLIP in predicting human choices and has good generalization capability toward images generated from other models. By tuning Stable Diffusion with the guidance of HPS, the adapted model is able to generate images that are more preferred by human users. The project page is available here httpstgxs002.github.ioalignsdweb .

Memory Efficient Diffusion Probabilistic Models via Patchbased Generation ; Diffusion probabilistic models have been successful in generating highquality and diverse images. However, traditional models, whose input and output are highresolution images, suffer from excessive memory requirements, making them less practical for edge devices. Previous approaches for generative adversarial networks proposed a patchbased method that uses positional encoding and global content information. Nevertheless, designing a patchbased approach for diffusion probabilistic models is nontrivial. In this paper, we resent a diffusion probabilistic model that generates images on a patchbypatch basis. We propose two conditioning methods for a patchbased generation. First, we propose positionwise conditioning using onehot representation to ensure patches are in proper positions. Second, we propose Global Content Conditioning GCC to ensure patches have coherent content when concatenated together. We evaluate our model qualitatively and quantitatively on CelebA and LSUN bedroom datasets and demonstrate a moderate tradeoff between maximum memory consumption and generated image quality. Specifically, when an entire image is divided into 2 x 2 patches, our proposed approach can reduce the maximum memory consumption by half while maintaining comparable image quality.

XIQE eXplainable Image Quality Evaluation for TexttoImage Generation with Visual Large Language Models ; This paper introduces a novel explainable image quality evaluation approach called XIQE, which leverages visual large language models LLMs to evaluate texttoimage generation methods by generating textual explanations. XIQE utilizes a hierarchical Chain of Thought CoT to enable MiniGPT4 to produce selfconsistent, unbiased texts that are highly correlated with human evaluation. It offers several advantages, including the ability to distinguish between real and generated images, evaluate textimage alignment, and assess image aesthetics without requiring model training or finetuning. XIQE is more costeffective and efficient compared to human evaluation, while significantly enhancing the transparency and explainability of deep image quality evaluation models. We validate the effectiveness of our method as a benchmark using images generated by prevalent diffusion models. XIQE demonstrates similar performance to stateoftheart SOTA evaluation methods on COCO Caption, while overcoming the limitations of previous evaluation models on DrawBench, particularly in handling ambiguous generation prompts and text recognition in generated images. Project website httpsgithub.comSchutureBenchmarkingAwesomeDiffusionModels

Improved Visual Story Generation with Adaptive Context Modeling ; Diffusion models developed on top of powerful texttoimage generation models like Stable Diffusion achieve remarkable success in visual story generation. However, the bestperforming approach considers historically generated results as flattened memory cells, ignoring the fact that not all preceding images contribute equally to the generation of the characters and scenes at the current stage. To address this, we present a simple method that improves the leading system with adaptive context modeling, which is not only incorporated in the encoder but also adopted as additional guidance in the sampling stage to boost the global consistency of the generated story. We evaluate our model on PororoSV and FlintstonesSV datasets and show that our approach achieves stateoftheart FID scores on both story visualization and continuation scenarios. We conduct detailed model analysis and show that our model excels at generating semantically consistent images for stories.

Diffusion probabilistic models enhance variational autoencoder for crystal structure generative modeling ; The crystal diffusion variational autoencoder CDVAE is a machine learning model that leverages score matching to generate realistic crystal structures that preserve crystal symmetry. In this study, we leverage novel diffusion probabilistic DP models to denoise atomic coordinates rather than adopting the standard score matching approach in CDVAE. Our proposed DPCDVAE model can reconstruct and generate crystal structures whose qualities are statistically comparable to those of the original CDVAE. Furthermore, notably, when comparing the carbon structures generated by the DPCDVAE model with relaxed structures obtained from density functional theory calculations, we find that the DPCDVAE generated structures are remarkably closer to their respective ground states. The energy differences between these structures and the true ground states are, on average, 68.1 meVatom lower than those generated by the original CDVAE. This significant improvement in the energy accuracy highlights the effectiveness of the DPCDVAE model in generating crystal structures that better represent their groundstate configurations.

Evaluating the diversity and utility of materials proposed by generative models ; Generative machine learning models can use data generated by scientific modeling to create large quantities of novel material structures. Here, we assess how one stateoftheart generative model, the physicsguided crystal generation model PGCGM, can be used as part of the inverse design process. We show that the default PGCGM's input space is not smooth with respect to parameter variation, making material optimization difficult and limited. We also demonstrate that most generated structures are predicted to be thermodynamically unstable by a separate propertyprediction model, partially due to outofdomain data challenges. Our findings suggest how generative models might be improved to enable better inverse design.

Applications of generalized special functions in stellar astrophysics ; This article gives an brief outline of the applications of generalized special functions such as generalized hypergeometric functions, Gfunctions and Hfunctions into the general area of nuclear energy generation and reaction rate theory such as the energy generation in a simple stellar model and nuclear reaction rates in nonresonant and resonant as well as screened nonresonant

Generic separable metric structures ; We compare three notions of genericity of separable metric structures. Our analysis provides a general model theoretic technique of showing that structures are generic in descriptive set theoretic topological sense and in measure theoretic sense. In particular, it gives a new perspective on Vershik's theorems on genericity and randomness of Urysohn's space among separable metric spaces.

RLDuet Online Music Accompaniment Generation Using Deep Reinforcement Learning ; This paper presents a deep reinforcement learning algorithm for online accompaniment generation, with potential for realtime interactive humanmachine duet improvisation. Different from offline music generation and harmonization, online music accompaniment requires the algorithm to respond to human input and generate the machine counterpart in a sequential order. We cast this as a reinforcement learning problem, where the generation agent learns a policy to generate a musical note action based on previously generated context state. The key of this algorithm is the wellfunctioning reward model. Instead of defining it using music composition rules, we learn this model from monophonic and polyphonic training data. This model considers the compatibility of the machinegenerated note with both the machinegenerated context and the humangenerated context. Experiments show that this algorithm is able to respond to the human part and generate a melodic, harmonic and diverse machine part. Subjective evaluations on preferences show that the proposed algorithm generates music pieces of higher quality than the baseline method.

Temporal Generative Adversarial Nets with Singular Value Clipping ; In this paper, we propose a generative model, Temporal Generative Adversarial Nets TGAN, which can learn a semantic representation of unlabeled videos, and is capable of generating videos. Unlike existing Generative Adversarial Nets GANbased methods that generate videos with a single generator consisting of 3D deconvolutional layers, our model exploits two different types of generators a temporal generator and an image generator. The temporal generator takes a single latent variable as input and outputs a set of latent variables, each of which corresponds to an image frame in a video. The image generator transforms a set of such latent variables into a video. To deal with instability in training of GAN with such advanced networks, we adopt a recently proposed model, Wasserstein GAN, and propose a novel method to train it stably in an endtoend manner. The experimental results demonstrate the effectiveness of our methods.

GAGAN GeometryAware Generative Adversarial Networks ; Deep generative models learned through adversarial training have become increasingly popular for their ability to generate naturalistic image textures. However, aside from their texture, the visual appearance of objects is significantly influenced by their shape geometry; information which is not taken into account by existing generative models. This paper introduces the GeometryAware Generative Adversarial Networks GAGAN for incorporating geometric information into the image generation process. Specifically, in GAGAN the generator samples latent variables from the probability space of a statistical shape model. By mapping the output of the generator to a canonical coordinate frame through a differentiable geometric transformation, we enforce the geometry of the objects and add an implicit connection from the prior to the generated object. Experimental results on face generation indicate that the GAGAN can generate realistic images of faces with arbitrary facial attributes such as facial expression, pose, and morphology, that are of better quality than current GANbased methods. Our method can be used to augment any existing GAN architecture and improve the quality of the images generated.

Recurrent Deconvolutional Generative Adversarial Networks with Application to Text Guided Video Generation ; This paper proposes a novel model for video generation and especially makes the attempt to deal with the problem of video generation from text descriptions, i.e., synthesizing realistic videos conditioned on given texts. Existing video generation methods cannot be easily adapted to handle this task well, due to the frame discontinuity issue and their textfree generation schemes. To address these problems, we propose a recurrent deconvolutional generative adversarial network RDGAN, which includes a recurrent deconvolutional network RDN as the generator and a 3D convolutional neural network 3DCNN as the discriminator. The RDN is a deconvolutional version of conventional recurrent neural network, which can well model the longrange temporal dependency of generated video frames and make good use of conditional information. The proposed model can be jointly trained by pushing the RDN to generate realistic videos so that the 3DCNN cannot distinguish them from real ones. We apply the proposed RDGAN to a series of tasks including conventional video generation, conditional video generation, video prediction and video classification, and demonstrate its effectiveness by achieving well performance.

Privacypreserving Spatiotemporal Scenario Generation of Renewable Energies A Federated Deep Generative Learning Approach ; Scenario generation is a fundamental and crucial tool for decisionmaking in power systems with highpenetration renewables. Based on big historical data, a novel federated deep generative learning framework, called FedLSGAN, is proposed by integrating federated learning and least square generative adversarial networks LSGANs for renewable scenario generation. Specifically, federated learning learns a shared global model in a central server from renewable sites at network edges, which enables the FedLSGAN to generate scenarios in a privacypreserving manner without sacrificing the generation quality by transferring model parameters, rather than all data. Meanwhile, the LSGANsbased deep generative model generates scenarios that conform to the distribution of historical data through fully capturing the spatialtemporal characteristics of renewable powers, which leverages the least squares loss function to improve the training stability and generation quality. The simulation results demonstrate that the proposal manages to generate highquality renewable scenarios and outperforms the stateoftheart centralized methods. Besides, an experiment with different federated learning settings is designed and conducted to verify the robustness of our method.

Toward Spatially Unbiased Generative Models ; Recent image generation models show remarkable generation performance. However, they mirror strong location preference in datasets, which we call spatial bias. Therefore, generators render poor samples at unseen locations and scales. We argue that the generators rely on their implicit positional encoding to render spatial content. From our observations, the generator's implicit positional encoding is translationvariant, making the generator spatially biased. To address this issue, we propose injecting explicit positional encoding at each scale of the generator. By learning the spatially unbiased generator, we facilitate the robust use of generators in multiple tasks, such as GAN inversion, multiscale generation, generation of arbitrary sizes and aspect ratios. Furthermore, we show that our method can also be applied to denoising diffusion probabilistic models.

PAGER Progressive AttributeGuided Extendable Robust Image Generation ; This work presents a generative modeling approach based on successive subspace learning SSL. Unlike most generative models in the literature, our method does not utilize neural networks to analyze the underlying source distribution and synthesize images. The resulting method, called the progressive attributeguided extendable robust image generative PAGER model, has advantages in mathematical transparency, progressive content generation, lower training time, robust performance with fewer training samples, and extendibility to conditional image generation. PAGER consists of three modules core generator, resolution enhancer, and quality booster. The core generator learns the distribution of lowresolution images and performs unconditional image generation. The resolution enhancer increases image resolution via conditional generation. Finally, the quality booster adds finer details to generated images. Extensive experiments on MNIST, FashionMNIST, and CelebA datasets are conducted to demonstrate generative performance of PAGER.

A General Formulation for Evaluating the Performance of Linear Power Flow Models ; Linear power flow LPF models are essential in power system analysis. Various LPF models are proposed, but some crucial questions are still remained what is the performance bound e.g., the error bound of LPF models, how to know a branch is applicable for LPF models or not, and what is the best LPF model. In this paper, these crucial questions are answered and a general formulation GF for evaluating the performance of LPF models is proposed. The GF actually figure out two core difficulties, the one is how to define the definition range of the LPF models, and the second is how to analytically obtain the best LPF model and evaluate the performance of a given LPF model. Besides, the key factors that affect the performance of LPF models are also analyzed through the proposed framework. The case studies compare the proposed LPF model with the DC power flow model, the physicalmodeldriven LPF model, and the datadriven LPF model, and the results show the effectiveness as well as the superiority of the proposed method.

CAMERO Consistency Regularized Ensemble of Perturbed Language Models with Weight Sharing ; Model ensemble is a popular approach to produce a lowvariance and wellgeneralized model. However, it induces large memory and inference costs, which are often not affordable for realworld deployment. Existing work has resorted to sharing weights among models. However, when increasing the proportion of the shared weights, the resulting models tend to be similar, and the benefits of using model ensemble diminish. To retain ensemble benefits while maintaining a low memory cost, we propose a consistencyregularized ensemble learning approach based on perturbed models, named CAMERO. Specifically, we share the weights of bottom layers across all models and apply different perturbations to the hidden representations for different models, which can effectively promote the model diversity. Meanwhile, we apply a prediction consistency regularizer across the perturbed models to control the variance due to the model diversity. Our experiments using large language models demonstrate that CAMERO significantly improves the generalization performance of the ensemble model. Specifically, CAMERO outperforms the standard ensemble of 8 BERTbase models on the GLUE benchmark by 0.7 with a significantly smaller model size 114.2M vs. 880.6M.

SUnotimes U1c gauge models with spontaneous symmetry breaking ; A possible generalization of the technique of the standard model to SUnotimes U1 gauge models is proposed. A special Higgs mechanism and a new kind of Yukawa couplings in unitary gauge are introduced. These allow us to obtain a general method of deriving boson mass spectrum and coupling coefficients which will be used to find an exact solution of the PisanoPleitez threegeneration SU3otimes U1 model. A new anomalyfree onegeneration model is briefly discussed.

Higgs Boson Mass Bounds in the Standard and Minimal Supersymmetric Standard Model with Four Generations ; We study the question of distinguishability of the Higgs sector between the standard model with four generationsSM4 and the minimal supersymmetric standard model with four generations MSSM4. We find that a gap exists between the SM4 and MSSM4 Higgs boson masses for a range of the fourth generation fermion mass considered in the analysis at a fixed top quark mass. We also compare the Higgs boson mass bounds in these models with those in the standard and the minimal supersymmetric standard models.

Multiloop correlators for rational theories of 2D gravity from the generalized Kontsevich models ; We introduce a parametrization of the coupling constant space of the generalized Kontsevich models in terms of a set of moments equivalent to those introduced recently in the context of topological gravity. For the simplest generalization of the Kontsevich model we express the moments as elementary functions of the susceptibilities and the eigenvalues of the external field. We furthermore use the moment technique to derive a closed expression for the genus zero multiloop correlators for 3,3m1 and 3,3m2 rational matter fields coupled to gravity. We comment on the relation between the twomatrix model and the generalized Kontsevich models

The distribution of Pearson residuals in generalized linear models ; In general, the distribution of residuals cannot be obtained explicitly. We give an asymptotic formula for the density of Pearson residuals in continuous generalized linear models corrected to order n1, where n is the sample size. We define corrected Pearson residuals for these models that, to this order of approximation, have exactly the same distribution of the true Pearson residuals. Applications for important generalized linear models are provided and simulation results for a gamma model illustrate the usefulness of the corrected Pearson residuals.

Random Sequential Generation of Intervals for the Cascade Model of Food Webs ; The cascade model generates a food web at random. In it the species are labeled from 0 to m, and arcs are given at random between pairs of the species. For an arc with endpoints i and j ij, the species i is eaten by the species labeled j. The chain length height, generated at random, models the length of food chain in ecological data. The aim of this note is to introduce the random sequential generation of intervals as a Poisson model which gives naturally an analogous behavior to the cascade model.

The Ising Model on Random Lattices in Arbitrary Dimensions ; We study analytically the Ising model coupled to random lattices in dimension three and higher. The family of random lattices we use is generated by the large N limit of a colored tensor model generalizing the twomatrix model for Ising spins on random surfaces. We show that, in the continuum limit, the spin system does not exhibit a phase transition at finite temperature, in agreement with numerical investigations. Furthermore we outline a general method to study critical behavior in colored tensor models.

Generalized preferential attachment tunable powerlaw degree distribution and clustering coefficient ; We propose a wide class of preferential attachment models of random graphs, generalizing previous approaches. Graphs described by these models obey the powerlaw degree distribution, with the exponent that can be controlled in the models. Moreover, clustering coefficient of these graphs can also be controlled. We propose a concrete flexible model from our class and provide an efficient algorithm for generating graphs in this model. All our theoretical results are demonstrated in practice on examples of graphs obtained using this algorithm. Moreover, observations of generated graphs lead to future questions and hypotheses not yet justified by theory.

Dynamic Term Structure Modelling with Default and Mortality Risk New Results on Existence and Monotonicity ; This paper considers general term structure models like the ones appearing in portfolio credit risk modelling or life insurance. We give a general model starting from families of forward rates driven by infinitely many Brownian motions and an integervalued random measure, generalizing existing approaches in the literature. Then we derive drift conditions which are equivalent to no asymptotic free lunch on the considered market. Existence results are also given. In practice, models possessing a certain monotonicity are favorable and we study general conditions which guarantee this. The setup is illustrated with some examples.

Generalized supersymmetry and sigma models ; In this paper, we discuss the generalizations of exact supersymmetries present in the supersymmetrized sigma models. These generalizations are made by making the supersymmetric transformation parameter fielddependent. Remarkably, the supersymmetric effective actions emerge naturally through the Jacobian associated with the generalized supersymmetry transformations. We explicitly demonstrate these for two different supersymmetric sigma models, namely, one dimensional sigma model and topological sigma model for hyperinstantons on quaternionic manifold.

Inflationary Weak Anisotropic Model with General Dissipation Coefficient ; This paper explores the dynamics of warm intermediate and logamediate inflationary models during weak dissipative regime with a general form of dissipative coefficient. We analyze these models within the framework of locally rotationally symmetric Bianchi type I universe. In both cases, we evaluate solution of inflaton, effective scalar potential, dissipative coefficient, slowroll parameters, scalar and tensor power spectra, scalar spectral index and tensor to scalar ratio under slowroll approximation. We constrain the model parameters using recent data and conclude that anisotropic inflationary universe model with generalized dissipation coefficient remains compatible with WMAP9, Planck and BICEP2 data.

Generative Modeling with Conditional Autoencoders Building an Integrated Cell ; We present a conditional generative model to learn variation in cell and nuclear morphology and the location of subcellular structures from microscopy images. Our model generalizes to a wide range of subcellular localization and allows for a probabilistic interpretation of cell and nuclear morphology and structure localization from fluorescence images. We demonstrate the effectiveness of our approach by producing photorealistic cell images using our generative model. The conditional nature of the model provides the ability to predict the localization of unobserved structures given cell and nuclear morphology.

Combining Generative and Discriminative Approaches to Unsupervised Dependency Parsing via Dual Decomposition ; Unsupervised dependency parsing aims to learn a dependency parser from unannotated sentences. Existing work focuses on either learning generative models using the expectationmaximization algorithm and its variants, or learning discriminative models using the discriminative clustering algorithm. In this paper, we propose a new learning strategy that learns a generative model and a discriminative model jointly based on the dual decomposition method. Our method is simple and general, yet effective to capture the advantages of both models and improve their learning results. We tested our method on the UD treebank and achieved a stateoftheart performance on thirty languages.

A Deep Ensemble Model with Slot Alignment for SequencetoSequence Natural Language Generation ; Natural language generation lies at the core of generative dialogue systems and conversational agents. We describe an ensemble neural language generator, and present several novel methods for data representation and augmentation that yield improved results in our model. We test the model on three datasets in the restaurant, TV and laptop domains, and report both objective and subjective evaluations of our best model. Using a range of automatic metrics, as well as human evaluators, we show that our approach achieves better results than stateoftheart models on the same datasets.

Signature change in loop quantum gravity General midisuperspace models and dilaton gravity ; Models of loop quantum gravity based on real connections have a deformed notion of general covariance, which leads to the phenomenon of signature change. This result is confirmed here in a general analysis of all midisuperspace models without local degrees of freedom. As a subclass of models, 2dimensional theories of dilaton gravity appear, but a larger set of examples is possible based only on the condition of anomaly freedom. While the classical dilaton gravity models are the only such systems without deformed covariance, they do give rise to signature change when holonomy modifications are included.

Comparative Study on Generative Adversarial Networks ; In recent years, there have been tremendous advancements in the field of machine learning. These advancements have been made through both academic as well as industrial research. Lately, a fair amount of research has been dedicated to the usage of generative models in the field of computer vision and image classification. These generative models have been popularized through a new framework called Generative Adversarial Networks. Moreover, many modified versions of this framework have been proposed in the last two years. We study the original model proposed by Goodfellow et al. as well as modifications over the original model and provide a comparative analysis of these models.

Deterministic NonAutoregressive Neural Sequence Modeling by Iterative Refinement ; We propose a conditional nonautoregressive neural sequence model based on iterative refinement. The proposed model is designed based on the principles of latent variable models and denoising autoencoders, and is generally applicable to any sequence generation task. We extensively evaluate the proposed model on machine translation EnDe and EnRo and image caption generation, and observe that it significantly speeds up decoding while maintaining the generation quality comparable to the autoregressive counterpart.

A NonParametric Test to Detect DataCopying in Generative Models ; Detecting overfitting in generative models is an important challenge in machine learning. In this work, we formalize a form of overfitting that we call emdatacopying where the generative model memorizes and outputs training samples or small variations thereof. We provide a three sample nonparametric test for detecting datacopying that uses the training set, a separate sample from the target distribution, and a generated sample from the model, and study the performance of our test on several canonical models and datasets. For code examples, visit httpsgithub.comcaseymeehandatacopying

General Bayesian L2 calibration of mathematical models ; A mathematical model is a representation of a physical system depending on unknown parameters. Calibration refers to attributing values to these parameters, using observations of the physical system, acknowledging that the mathematical model is an inexact representation of the physical system. General Bayesian inference generalizes traditional Bayesian inference by replacing the loglikelihood in Bayes' theorem by a negative loss function. Methodology is proposed for the general Bayesian calibration of mathematical models where the resulting posterior distributions estimate the values of the parameters that minimize the L2 norm of the difference between the mathematical model and true physical system.

Generating Related Work ; Communicating new research ideas involves highlighting similarities and differences with past work. Authors write fluent, often long sections to survey the distinction of a new paper with related work. In this work we model generating related work sections while being cognisant of the motivation behind citing papers. Our content planning model generates a tree of cited papers before a surface realization model lexicalizes this skeleton. Our model outperforms several strong stateoftheart summarization and multidocument summarization models on generating related work on an ACL Anthology AA based dataset which we contribute.

ZeroShot Estimation of Base Models' Weights in Ensemble of Machine Reading Comprehension Systems for Robust Generalization ; One of the main challenges of the machine reading comprehension MRC models is their fragile outofdomain generalization, which makes these models not properly applicable to realworld generalpurpose question answering problems. In this paper, we leverage a zeroshot weighted ensemble method for improving the robustness of outofdomain generalization in MRC models. In the proposed method, a weight estimation module is used to estimate outofdomain weights, and an ensemble module aggregate several base models' predictions based on their weights. The experiments indicate that the proposed method not only improves the final accuracy, but also is robust against domain changes.

Magnetic black holes with generalized ModMax model of nonlinear electrodynamics ; Recently Bandos, Lechner, Sorokin, and Townsend Phys. Rev. D textbf102, 121703 2020 proposed Modified Maxwell ModMax model of nonlinear dualityinvariant conformal electrodynamics. Here, Generalized ModMax GenModMax model of nonlinear electrodynamics coupled to general relativity is studied. The metric and mass functions, and their asymptotic as rrightarrowinfty and rrightarrow 0 of a magnetic black hole are obtained. Corrections to the ReissnerNordstrom solution are found and we show that for some model parameters the black hole is regular. The Hawking temperature and heat capacity of black holes are calculated and phase transitions are investigated. We demonstrate that black holes are not stable for certain model parameters.

Protein Structure and Sequence Generation with Equivariant Denoising Diffusion Probabilistic Models ; Proteins are macromolecules that mediate a significant fraction of the cellular processes that underlie life. An important task in bioengineering is designing proteins with specific 3D structures and chemical properties which enable targeted functions. To this end, we introduce a generative model of both protein structure and sequence that can operate at significantly larger scales than previous molecular generative modeling approaches. The model is learned entirely from experimental data and conditions its generation on a compact specification of protein topology to produce a fullatom backbone configuration as well as sequence and sidechain predictions. We demonstrate the quality of the model via qualitative and quantitative analysis of its samples. Videos of sampling trajectories are available at httpsnanand2.github.ioproteins .

Structural generalization is hard for sequencetosequence models ; Sequencetosequence seq2seq models have been successful across many NLP tasks, including ones that require predicting linguistic structure. However, recent work on compositional generalization has shown that seq2seq models achieve very low accuracy in generalizing to linguistic structures that were not seen in training. We present new evidence that this is a general limitation of seq2seq models that is present not just in semantic parsing, but also in syntactic parsing and in texttotext tasks, and that this limitation can often be overcome by neurosymbolic models that have linguistic knowledge built in. We further report on some experiments that give initial answers on the reasons for these limitations.

Counterfactual Identifiability of Bijective Causal Models ; We study counterfactual identifiability in causal models with bijective generation mechanisms BGM, a class that generalizes several widelyused causal models in the literature. We establish their counterfactual identifiability for three common causal structures with unobserved confounding, and propose a practical learning method that casts learning a BGM as structured generative modeling. Learned BGMs enable efficient counterfactual estimation and can be obtained using a variety of deep conditional generative models. We evaluate our techniques in a visual task and demonstrate its application in a realworld video streaming simulation task.

Generative Models for 3D Point Clouds ; Point clouds are rich geometric data structures, where their three dimensional structure offers an excellent domain for understanding the representation learning and generative modeling in 3D space. In this work, we aim to improve the performance of point cloud latentspace generative models by experimenting with transformer encoders, latentspace flow models, and autoregressive decoders. We analyze and compare both generation and reconstruction performance of these models on various object types.

Calliffusion Chinese Calligraphy Generation and Style Transfer with Diffusion Modeling ; In this paper, we propose Calliffusion, a system for generating highquality Chinese calligraphy using diffusion models. Our model architecture is based on DDPM Denoising Diffusion Probabilistic Models, and it is capable of generating common characters in five different scripts and mimicking the styles of famous calligraphers. Experiments demonstrate that our model can generate calligraphy that is difficult to distinguish from real artworks and that our controls for characters, scripts, and styles are effective. Moreover, we demonstrate oneshot transfer learning, using LoRA LowRank Adaptation to transfer Chinese calligraphy art styles to unseen characters and even outofdomain symbols such as English letters and digits.

Anomaly Detection in Networks via ScoreBased Generative Models ; Node outlier detection in attributed graphs is a challenging problem for which there is no method that would work well across different datasets. Motivated by the stateoftheart results of scorebased models in graph generative modeling, we propose to incorporate them into the aforementioned problem. Our method achieves competitive results on smallscale graphs. We provide an empirical analysis of the Dirichlet energy, and show that generative models might struggle to accurately reconstruct it.

A New Algorithm for Doptimal Designs under General Parametric Statistical Models with Mixed Factors ; In this paper, we consider experiments involving both discrete factors and continuous factors under general parametric statistical models. To search for optimal designs under the Dcriterion, we propose a new algorithm, called the ForLion algorithm, which performs an exhaustive search in a design space with discrete and continuous factors while keeping high efficiency and a reduced number of design points. Its optimality is guaranteed by the general equivalence theorem. We show its advantage using a reallife experiment under multinomial logistic models, and further specialize the algorithm for generalized linear models to show the improved efficiency with modelspecific formulae and iterative steps.

TDG Textguided Domain Generalization ; Domain generalization DG attempts to generalize a model trained on single or multiple source domains to the unseen target domain. Benefiting from the success of VisualandLanguage Pretrained models in recent years, we argue that it is crucial for domain generalization by introducing extra text information. In this paper, we develop a novel Textguided Domain Generalization TDG paradigm for domain generalization, which includes three following aspects. Specifically, we first devise an automatic words generation method to extend the description of current domains with novel domainrelevant words. Then, we embed the generated domain information into the text feature space, by the proposed prompt learningbased text feature generation method, which shares a common representation space with the image feature. Finally, we utilize both input image features and generated text features to train a specially designed classifier that generalizes well on unseen target domains, while the image encoder is also updated under the supervision of gradients back propagated from the classifier. Our experimental results show that the techniques incorporated by TDG contribute to the performance in an easy implementation manner. Experimental results on several domain generalization benchmarks show that our proposed framework achieves superior performance by effectively leveraging generated text information in domain generalization.

Towards Product Lining ModelDriven Development Code Generators ; A code generator systematically transforms compact models to detailed code. Today, code generation is regarded as an integral part of modeldriven development MDD. Despite its relevance, the development of code generators is an inherently complex task and common methodologies and architectures are lacking. Additionally, reuse and extension of existing code generators only exist on individual parts. A systematic development and reuse based on a code generator product line is still in its infancy. Thus, the aim of this paper is to identify the mechanism necessary for a code generator product line by a analyzing the common product line development approach and b mapping those to a code generator specific infrastructure. As a first step towards realizing a code generator product line infrastructure, we present a componentbased implementation approach based on ideas of variabilityaware module systems and point out further research challenges.

Fast Adaptation in Generative Models with Generative Matching Networks ; Despite recent advances, the remaining bottlenecks in deep generative models are necessity of extensive training and difficulties with generalization from small number of training examples. We develop a new generative model called Generative Matching Network which is inspired by the recently proposed matching networks for oneshot learning in discriminative tasks. By conditioning on the additional input dataset, our model can instantly learn new concepts that were not available in the training data but conform to a similar generative process. The proposed framework does not explicitly restrict diversity of the conditioning data and also does not require an extensive inference procedure for training or adaptation. Our experiments on the Omniglot dataset demonstrate that Generative Matching Networks significantly improve predictive performance on the fly as more additional data is available and outperform existing state of the art conditional generative models.

Learning to Generate TimeLapse Videos Using MultiStage Dynamic Generative Adversarial Networks ; Taking a photo outside, can we predict the immediate future, e.g., how would the cloud move in the sky We address this problem by presenting a generative adversarial network GAN based twostage approach to generating realistic timelapse videos of high resolution. Given the first frame, our model learns to generate longterm future frames. The first stage generates videos of realistic contents for each frame. The second stage refines the generated video from the first stage by enforcing it to be closer to real videos with regard to motion dynamics. To further encourage vivid motion in the final generated video, Gram matrix is employed to model the motion more precisely. We build a large scale timelapse dataset, and test our approach on this new dataset. Using our model, we are able to generate realistic videos of up to 128times 128 resolution for 32 frames. Quantitative and qualitative experiment results have demonstrated the superiority of our model over the stateoftheart models.

Model Selection for Generalized Zeroshot Learning ; In the problem of generalized zeroshot learning, the datapoints from unknown classes are not available during training. The main challenge for generalized zeroshot learning is the unbalanced data distribution which makes it hard for the classifier to distinguish if a given testing sample comes from a seen or unseen class. However, using Generative Adversarial Network GAN to generate auxiliary datapoints by the semantic embeddings of unseen classes alleviates the above problem. Current approaches combine the auxiliary datapoints and original training data to train the generalized zeroshot learning model and obtain stateoftheart results. Inspired by such models, we propose to feed the generated data via a model selection mechanism. Specifically, we leverage two sources of datapoints observed and auxiliary to train some classifier to recognize which test datapoints come from seen and which from unseen classes. This way, generalized zeroshot learning can be divided into two disjoint classification tasks, thus reducing the negative influence of the unbalanced data distribution. Our evaluations on four publicly available datasets for generalized zeroshot learning show that our model obtains stateoftheart results.

Generating Multiple Diverse Responses for ShortText Conversation ; Neural generative models have become popular and achieved promising performance on shorttext conversation tasks. They are generally trained to build a 1to1 mapping from the input post to its output response. However, a given post is often associated with multiple replies simultaneously in real applications. Previous research on this task mainly focuses on improving the relevance and informativeness of the top one generated response for each post. Very few works study generating multiple accurate and diverse responses for the same post. In this paper, we propose a novel response generation model, which considers a set of responses jointly and generates multiple diverse responses simultaneously. A reinforcement learning algorithm is designed to solve our model. Experiments on two shorttext conversation tasks validate that the multiple responses generated by our model obtain higher quality and larger diversity compared with various stateoftheart generative models.

Factorized Deep Generative Models for Trajectory Generation with SpatiotemporalValidity Constraints ; Trajectory data generation is an important domain that characterizes the generative process of mobility data. Traditional methods heavily rely on predefined heuristics and distributions and are weak in learning unknown mechanisms. Inspired by the success of deep generative neural networks for images and texts, a fastdeveloping research topic is deep generative models for trajectory data which can learn expressively explanatory models for sophisticated latent patterns. This is a nascent yet promising domain for many applications. We first propose novel deep generative models factorizing timevariant and timeinvariant latent variables that characterize global and local semantics, respectively. We then develop new inference strategies based on variational inference and constrained optimization to encapsulate the spatiotemporal validity. New deep neural network architectures have been developed to implement the inference and generation models with newlygeneralized latent variable priors. The proposed methods achieved significant improvements in quantitative and qualitative evaluations in extensive experiments.

Language Generation with MultiHop Reasoning on Commonsense Knowledge Graph ; Despite the success of generative pretrained language models on a series of text generation tasks, they still suffer in cases where reasoning over underlying commonsense knowledge is required during generation. Existing approaches that integrate commonsense knowledge into generative pretrained language models simply transfer relational knowledge by posttraining on individual knowledge triples while ignoring rich connections within the knowledge graph. We argue that exploiting both the structural and semantic information of the knowledge graph facilitates commonsenseaware text generation. In this paper, we propose Generation with MultiHop Reasoning Flow GRF that enables pretrained models with dynamic multihop reasoning on multirelational paths extracted from the external commonsense knowledge graph. We empirically show that our model outperforms existing baselines on three text generation tasks that require reasoning over commonsense knowledge. We also demonstrate the effectiveness of the dynamic multihop reasoning module with reasoning paths inferred by the model that provide rationale to the generation.

Graphbased Multihop Reasoning for Long Text Generation ; Long text generation is an important but challenging task.The main problem lies in learning sentencelevel semantic dependencies which traditional generative models often suffer from. To address this problem, we propose a Multihop Reasoning Generation MRG approach that incorporates multihop reasoning over a knowledge graph to learn semantic dependencies among sentences. MRG consists of twoparts, a graphbased multihop reasoning module and a pathaware sentence realization module. The reasoning module is responsible for searching skeleton paths from a knowledge graph to imitate the imagination process in the human writing for semantic transfer. Based on the inferred paths, the sentence realization module then generates a complete sentence. Unlike previous blackbox models, MRG explicitly infers the skeleton path, which provides explanatory views tounderstand how the proposed model works. We conduct experiments on three representative tasks, including story generation, review generation, and product description generation. Automatic and manual evaluation show that our proposed method can generate more informative and coherentlong text than strong baselines, such as pretrained modelse.g. GPT2 and knowledgeenhanced models.

Towards Diverse Paraphrase Generation Using MultiClass Wasserstein GAN ; Paraphrase generation is an important and challenging natural language processing NLP task. In this work, we propose a deep generative model to generate paraphrase with diversity. Our model is based on an encoderdecoder architecture. An additional transcoder is used to convert a sentence into its paraphrasing latent code. The transcoder takes an explicit pattern embedding variable as condition, so diverse paraphrase can be generated by sampling on the pattern embedding variable. We use a Wasserstein GAN to align the distributions of the real and generated paraphrase samples. We propose a multiclass extension to the Wasserstein GAN, which allows our generative model to learn from both positive and negative samples. The generated paraphrase distribution is forced to get closer to the positive real distribution, and be pushed away from the negative distribution in Wasserstein distance. We test our model in two datasets with both automatic metrics and human evaluation. Results show that our model can generate fluent and reliable paraphrase samples that outperform the stateofart results, while also provides reasonable variability and diversity.

Multipleobjective Reinforcement Learning for Inverse Design and Identification ; The aim of the inverse chemical design is to develop new molecules with given optimized molecular properties or objectives. Recently, generative deep learning DL networks are considered as the stateoftheart in inverse chemical design and have achieved early success in generating molecular structures with desired properties in the pharmaceutical and material chemistry fields. However, satisfying a large number larger than 10 objectives of molecular objectives is a limitation of current generative models. To improve the model's ability to handle a large number of molecule design objectives, we developed a Reinforcement Learning RL based generative framework to optimize chemical molecule generation. Our use of Curriculum Learning CL to finetune the pretrained generative network allowed the model to satisfy up to 21 objectives and increase the generative network's robustness. The experiments show that the proposed multipleobjective RLbased generative model can correctly identify unknown molecules with an 83 to 100 percent success rate, compared to the baseline approach of 0 percent. Additionally, this proposed generative model is not limited to just chemistry research challenges; we anticipate that problems that utilize RL with multipleobjectives will benefit from this framework.

Adaptive Parameterization for Neural Dialogue Generation ; Neural conversation systems generate responses based on the sequencetosequence SEQ2SEQ paradigm. Typically, the model is equipped with a single set of learned parameters to generate responses for given input contexts. When confronting diverse conversations, its adaptability is rather limited and the model is hence prone to generate generic responses. In this work, we propose an bf Adaptive bf Neural bf Dialogue generation model, textscAdaND, which manages various conversations with conversationspecific parameterization. For each conversation, the model generates parameters of the encoderdecoder by referring to the input context. In particular, we propose two adaptive parameterization mechanisms a contextaware and a topicaware parameterization mechanism. The contextaware parameterization directly generates the parameters by capturing local semantics of the given context. The topicaware parameterization enables parameter sharing among conversations with similar topics by first inferring the latent topics of the given context and then generating the parameters with respect to the distributional topics. Extensive experiments conducted on a largescale realworld conversational dataset show that our model achieves superior performance in terms of both quantitative metrics and human evaluations.

AffectON Incorporating Affect Into Dialog Generation ; Due to its expressivity, natural language is paramount for explicit and implicit affective state communication among humans. The same linguistic inquiry e.g., How are you might induce responses with different affects depending on the affective state of the conversational partners and the context of the conversation. Yet, most dialog systems do not consider affect as constitutive aspect of response generation. In this paper, we introduce AffectON, an approach for generating affective responses during inference. For generating language in a targeted affect, our approach leverages a probabilistic language model and an affective space. AffectON is language model agnostic, since it can work with probabilities generated by any language model e.g., sequencetosequence models, neural language models, ngrams. Hence, it can be employed for both affective dialog and affective language generation. We experimented with affective dialog generation and evaluated the generated text objectively and subjectively. For the subjective part of the evaluation, we designed a custom user interface for rating and provided recommendations for the design of such interfaces. The results, both subjective and objective demonstrate that our approach is successful in pulling the generated language toward the targeted affect, with little sacrifice in syntactic coherence.

Enhance Convolutional Neural Networks with Noise Incentive Block ; As a generic modeling tool, Convolutional Neural Networks CNNs have been widely employed in image generation and translation tasks. However, when fed with a flat input, current CNN models may fail to generate vivid results due to the spatially shared convolution kernels. We call it the flatness degradation of CNNs. Unfortunately, such degradation is the greatest obstacles to generate a spatiallyvariant output from a flat input, which has been barely discussed in the previous literature. To tackle this problem, we propose a model agnostic solution, i.e. Noise Incentive Block NIB, which serves as a generic plugin for any CNN generation model. The key idea is to break the flat input condition while keeping the intactness of the original information. Specifically, the NIB perturbs the input data symmetrically with a noise map and reassembles them in the feature domain as driven by the objective function. Extensive experiments show that existing CNN models equipped with NIB survive from the flatness degradation and are able to generate visually better results with richer details in some specific image generation tasks given flat inputs, e.g. semantic image synthesis, datahidden image generation, and deep neural dithering.

The Perils of Using Mechanical Turk to Evaluate OpenEnded Text Generation ; Recent text generation research has increasingly focused on openended domains such as story and poetry generation. Because models built for such tasks are difficult to evaluate automatically, most researchers in the space justify their modeling choices by collecting crowdsourced human judgments of text quality e.g., Likert scores of coherence or grammaticality from Amazon Mechanical Turk AMT. In this paper, we first conduct a survey of 45 openended text generation papers and find that the vast majority of them fail to report crucial details about their AMT tasks, hindering reproducibility. We then run a series of story evaluation experiments with both AMT workers and English teachers and discover that even with strict qualification filters, AMT workers unlike teachers fail to distinguish between modelgenerated text and humangenerated references. We show that AMT worker judgments improve when they are shown modelgenerated output alongside humangenerated references, which enables the workers to better calibrate their ratings. Finally, interviews with the English teachers provide deeper insights into the challenges of the evaluation process, particularly when rating modelgenerated text.

Generating Multivariate Load States Using a Conditional Variational Autoencoder ; For planning of power systems and for the calibration of operational tools, it is essential to analyse system performance in a large range of representative scenarios. When the available historical data is limited, generative models are a promising solution, but modelling highdimensional dependencies is challenging. In this paper, a multivariate load state generating model on the basis of a conditional variational autoencoder CVAE neural network is proposed. Going beyond common CVAE implementations, the model includes stochastic variation of output samples under given latent vectors and cooptimizes the parameters for this output variability. It is shown that this improves statistical properties of the generated data. The quality of generated multivariate loads is evaluated using univariate and multivariate performance metrics. A generation adequacy case study on the European network is used to illustrate model's ability to generate realistic tail distributions. The experiments demonstrate that the proposed generator outperforms other data generating mechanisms.

Learning Probabilistic Models from Generator Latent Spaces with Hat EBM ; This work proposes a method for using any generator network as the foundation of an EnergyBased Model EBM. Our formulation posits that observed images are the sum of unobserved latent variables passed through the generator network and a residual random variable that spans the gap between the generator output and the image manifold. One can then define an EBM that includes the generator as part of its forward pass, which we call the Hat EBM. The model can be trained without inferring the latent variables of the observed data or calculating the generator Jacobian determinant. This enables explicit probabilistic modeling of the output distribution of any type of generator network. Experiments show strong performance of the proposed method on 1 unconditional ImageNet synthesis at 128x128 resolution, 2 refining the output of existing generators, and 3 learning EBMs that incorporate nonprobabilistic generators. Code and pretrained models to reproduce our results are available at httpsgithub.compoint0bar1hatebm.

Graph Generation with DestinationPredicting Diffusion Mixture ; Generation of graphs is a major challenge for realworld tasks that require understanding the complex nature of their nonEuclidean structures. Although diffusion models have achieved notable success in graph generation recently, they are illsuited for modeling the structural information of graphs since learning to denoise the noisy samples does not explicitly capture the graph topology. To tackle this limitation, we propose a novel generative framework that models the topology of graphs by predicting the destination of the diffusion process, which is the original graph that has the correct topology information, as a weighted mean of data. Specifically, we design the generative process as a mixture of diffusion processes conditioned on the endpoint in the data distribution, which drives the process toward the predicted destination, resulting in rapid convergence. We introduce new simulationfree training objectives for predicting the destination, and further discuss the advantages of our framework that can explicitly model the graph topology and exploit the inductive bias of the data. Through extensive experimental validation on general graph and 2D3D molecule generation tasks, we show that our method outperforms previous generative models, generating graphs with correct topology with both continuous e.g. 3D coordinates and discrete e.g. atom types features.

Coincidental Generation ; Generative A.I. models have emerged as versatile tools across diverse industries, with applications in privacypreserving data sharing, computational art, personalization of products and services, and immersive entertainment. Here, we introduce a new privacy concern in the adoption and use of generative A.I. models that of coincidental generation, where a generative model's output is similar enough to an existing entity, beyond those represented in the dataset used to train the model, to be mistaken for it. Consider, for example, synthetic portrait generators, which are today deployed in commercial applications such as virtual modeling agencies and synthetic stock photography. Due to the low intrinsic dimensionality of human face perception, every synthetically generated face will coincidentally resemble an actual person. Such examples of coincidental generation all but guarantee the misappropriation of likeness and expose organizations that use generative A.I. to legal and regulatory risk.

Ambigram Generation by A Diffusion Model ; Ambigrams are graphical letter designs that can be read not only from the original direction but also from a rotated direction especially with 180 degrees. Designing ambigrams is difficult even for human experts because keeping their dual readability from both directions is often difficult. This paper proposes an ambigram generation model. As its generation module, we use a diffusion model, which has recently been used to generate highquality photographic images. By specifying a pair of letter classes, such as 'A' and 'B', the proposed model generates various ambigram images which can be read as 'A' from the original direction and 'B' from a direction rotated 180 degrees. Quantitative and qualitative analyses of experimental results show that the proposed model can generate highquality and diverse ambigrams. In addition, we define ambigramability, an objective measure of how easy it is to generate ambigrams for each letter pair. For example, the pair of 'A' and 'V' shows a high ambigramability that is, it is easy to generate their ambigrams, and the pair of 'D' and 'K' shows a lower ambigramability. The ambigramability gives various hints of the ambigram generation not only for computers but also for human experts. The code can be found at httpsgithub.comunivesutyambifusion.

JEN1 TextGuided Universal Music Generation with Omnidirectional Diffusion Models ; Music generation has attracted growing interest with the advancement of deep generative models. However, generating music conditioned on textual descriptions, known as texttomusic, remains challenging due to the complexity of musical structures and high sampling rate requirements. Despite the task's significance, prevailing generative models exhibit limitations in music quality, computational efficiency, and generalization. This paper introduces JEN1, a universal highfidelity model for texttomusic generation. JEN1 is a diffusion model incorporating both autoregressive and nonautoregressive training. Through incontext learning, JEN1 performs various generation tasks including textguided music generation, music inpainting, and continuation. Evaluations demonstrate JEN1's superior performance over stateoftheart methods in textmusic alignment and music quality while maintaining computational efficiency. Our demos are available at httpfutureverse.comresearchjendemosjen1

A Model of HotSector Generations ; Possible existence of hotsector generations above the well known 3 generation bound is investigated on the basis of a model of leptons and quarks, which is based on the Harari and Shupe's one. Our model predicts the existence of bf3 1 generations above the ordinary coldsector 3 generations. Majorana neutrinos are introduced to realize the heavy neutrino masses in hotsector generations. Properties of heavy neutrinos are also discussed.

A single model of traversable wormholes supported by generalized phantom energy or Chaplygin gas ; This paper discusses a new variable equation of state parameter leading to exact solutions of the Einstein field equations describing traversable wormholes. In addition to generalizing the notion of phantom energy, the equation of state generates a mathematical model that combines the generalized phantom energy and the generalized Chaplygin gas models.

The Probabilistic Model of Keys Generation of QKD Systems ; The probabilistic model of keys generation of QKD systems is proposed. The model includes all phases of keys generation starting from photons generation to states detection taking characteristics of fiberoptics components into account. The paper describes the tree of events of QKD systems. Equations are found for estimation of the effectiveness of the process of sifted keys generation as well as for biterror probability and for the rate of private keys generation.

Statefinder Description in Generalized Holographic and Ricci Dark Energy Models ; We have considered the generalized holographic and generalized Ricci dark energy models for acceleration of the universe. If the universe filled with only GHDEGRDE the corresponding decel eration parameter, EOS parameter and statefinder parameters have been calculated. Next we have considered that the mixture of GHDEGRDE and dark matter in interacting and noninteracting situations. Also the mixture of GHDEGRDE and generalized Chaplygin gas have been analyzed during evolution of the universe. The natures of above mentioned parameters have been investigated for interacting and noninteracting situations. Finally, it follows that the prescribed models derive the acceleration of the universe.

Statefinder Diagnostic for Dark Energy Models in Bianchi I Universe ; In this paper, we investigate the statefinder, the deceleration and equation of state parameters when universe is composed of generalized holographic dark energy or generalized Ricci dark energy for Bianchi I universe model. These parameters are found for both interacting as well as noninteracting scenarios of generalized holographic or generalized Ricci dark energy with dark matter and generalized Chaplygin gas. We explore these parameters graphically for different situations. It is concluded that these models represent accelerated expansion of the universe.

Cross Domain Image Generation through Latent Space Exploration with Adversarial Loss ; Conditional domain generation is a good way to interactively control sample generation process of deep generative models. However, once a conditional generative model has been created, it is often expensive to allow it to adapt to new conditional controls, especially the network structure is relatively deep. We propose a conditioned latent domain transfer framework across latent spaces of unconditional variational autoencodersVAE. With this framework, we can allow unconditionally trained VAEs to generate images in its domain with conditionals provided by a latent representation of another domain. This framework does not assume commonalities between two domains. We demonstrate effectiveness and robustness of our model under widely used image datasets.

SO32 heterotic standard model vacua in general CalabiYau compactifications ; We study a direct flux breaking scenario in SO32 heterotic string theory on general CalabiYau threefolds. The direct flux breaking, corresponding to hypercharge flux breaking in the Ftheory context, allows us to derive the Standard Model in general CalabiYau compactifications. We present a general formula leading to the three generations of quarks and leptons and no chiral exotics in a backgroundindependent way. As a concrete example, we show the threegeneration model on a complete intersection CalabiYau threefold.

Unpriortized Autoencoder For Image Generation ; In this paper, we treat the image generation task using an autoencoder, a representative latent model. Unlike many studies regularizing the latent variable's distribution by assuming a manually specified prior, we approach the image generation task using an autoencoder by directly estimating the latent distribution. To this end, we introduce 'latent density estimator' which captures latent distribution explicitly and propose its structure. Through experiments, we show that our generative model generates images with the improved visual quality compared to previous autoencoderbased generative models.

GraphNVP An Invertible Flow Model for Generating Molecular Graphs ; We propose GraphNVP, the first invertible, normalizing flowbased molecular graph generation model. We decompose the generation of a graph into two steps generation of i an adjacency tensor and ii node attributes. This decomposition yields the exact likelihood maximization on graphstructured data, combined with two novel reversible flows. We empirically demonstrate that our model efficiently generates valid molecular graphs with almost no duplicated molecules. In addition, we observe that the learned latent space can be used to generate molecules with desired chemical properties.

SemiImplicit Generative Model ; To combine explicit and implicit generative models, we introduce semiimplicit generator SIG as a flexible hierarchical model that can be trained in the maximum likelihood framework. Both theoretically and experimentally, we demonstrate that SIG can generate high quality samples especially when dealing with multimodality. By introducing SIG as an unbiased regularizer to the generative adversarial network GAN, we show the interplay between maximum likelihood and adversarial learning can stabilize the adversarial training, resist the notorious mode collapsing problem of GANs, and improve the diversity of generated random samples.

Towards an Accurate Mathematical Model of Generic NominallyTyped OOP ; The construction of GNOOP as a domaintheoretic model of generic nominallytyped OOP is currently underway. This extended abstract presents the concepts of nominal intervals' and full generication' that are likely to help in building GNOOP as an accurate mathematical model of generic nominallytyped OOP. The abstract also presents few related categorytheoretic suggestions. The presented concepts and suggestions are particularly geared towards enabling GNOOP to offer a precise and simple view of sofarhardtoanalyze features of generic OOP such as variance annotations e.g., Java wildcard types and erased generics e.g., Java type erasure.

Correlated discrete data generation using adversarial training ; Generative Adversarial Networks GAN have shown great promise in tasks like synthetic image generation, image inpainting, style transfer, and anomaly detection. However, generating discrete data is a challenge. This work presents an adversarial training based correlated discrete data CDD generation model. It also details an approach for conditional CDD generation. The results of our approach are presented over two datasets; jobseeking candidates skill set private dataset and MNIST public dataset. From quantitative and qualitative analysis of these results, we show that our model performs better as it leverages inherent correlation in the data, than an existing model that overlooks correlation.

Measuring Fairness in Generative Models ; Deep generative models have made much progress in improving training stability and quality of generated data. Recently there has been increased interest in the fairness of deepgenerated data. Fairness is important in many applications, e.g. law enforcement, as biases will affect efficacy. Central to fair data generation are the fairness metrics for the assessment and evaluation of different generative models. In this paper, we first review fairness metrics proposed in previous works and highlight potential weaknesses. We then discuss a performance benchmark framework along with the assessment of alternative metrics.

Selective Sampling and Mixture Models in Generative Adversarial Networks ; In this paper, we propose a multigenerator extension to the adversarial training framework, in which the objective of each generator is to represent a unique component of a target mixture distribution. In the training phase, the generators cooperate to represent, as a mixture, the target distribution while maintaining distinct manifolds. As opposed to traditional generative models, inference from a particular generator after training resembles selective sampling from a unique component in the target distribution. We demonstrate the feasibility of the proposed architecture both analytically and with basic MultiLayer Perceptron MLP models trained on the MNIST dataset.

Texygen A Benchmarking Platform for Text Generation Models ; We introduce Texygen, a benchmarking platform to support research on opendomain text generation models. Texygen has not only implemented a majority of text generation models, but also covered a set of metrics that evaluate the diversity, the quality and the consistency of the generated texts. The Texygen platform could help standardize the research on text generation and facilitate the sharing of finetuned opensource implementations among researchers for their work. As a consequence, this would help in improving the reproductivity and reliability of future research work in text generation.

Generative Models for Pose Transfer ; We investigate nearest neighbor and generative models for transferring pose between persons. We take in a video of one person performing a sequence of actions and attempt to generate a video of another person performing the same actions. Our generative model pix2pix outperforms kNN at both generating corresponding frames and generalizing outside the demonstrated action set. Our most salient contribution is determining a pipeline pose detection, face detection, kNN based pairing that is effective at performing the desired task. We also detail several iterative improvements and failure modes.