| { | |
| "overview": { | |
| "where": { | |
| "has-leaderboard": "yes", | |
| "leaderboard-url": "https://eval.ai/challenge/1663/overview", | |
| "leaderboard-description": "It evaluates the models based on the automatic metrics defined in the task paper for the three tasks of detection, selection and generation.", | |
| "website": "https://github.com/alexa/alexa-with-dstc10-track2-dataset", | |
| "data-url": "https://github.com/alexa/alexa-with-dstc10-track2-dataset", | |
| "paper-url": "https://assets.amazon.science/54/a1/5282d47044179737b4289622c824/how-robust-are-you-evaluating-task-oriented-dialogue-systems-on-spoken-conversations.pdf", | |
| "paper-bibtext": "@inproceedings{kim2021robust,\n title={\" How Robust ru?\": Evaluating Task-Oriented Dialogue Systems on Spoken Conversations},\n author={Kim, Seokhwan and Liu, Yang and Jin, Di and Papangelis, Alexandros and Gopalakrishnan, Karthik and Hedayatnia, Behnam and Hakkani-Tur, Dilek},\n journal={IEEE Automatic Speech Recognition and Understanding Workshop},\n year={2021}\n}", | |
| "contact-name": "Seokhwan Kim", | |
| "contact-email": "[email protected]" | |
| }, | |
| "languages": { | |
| "is-multilingual": "no", | |
| "license": "apache-2.0: Apache License 2.0", | |
| "task-other": "N/A", | |
| "language-names": [ | |
| "En" | |
| ], | |
| "intended-use": "To conduct research on dialogue state tracking and knowledge-grounded response generation.", | |
| "license-other": "N/A", | |
| "task": "Dialog Response Generation", | |
| "communicative": "This dataset aims to explore the robustness of conversational models when trained on spoken data. It has two aspects, multi-domain dialogue state tracking and conversation modeling with access to unstructured knowledge." | |
| }, | |
| "credit": { | |
| "organization-type": [ | |
| "industry" | |
| ], | |
| "organization-names": "Amazon", | |
| "creators": "Seokhwan Kim, Yang Liu, Di Jin, Alexandros Papangelis, Karthik Gopalakrishnan, Behnam Hedayatnia, Dilek Hakkani-Tur (Amazon Alexa AI)", | |
| "funding": "Amazon", | |
| "gem-added-by": "Alexandros Papangelis (Amazon Alexa AI), Di Jin (Amazon Alexa AI), Nico Daheim (RWTH Aachen University)" | |
| }, | |
| "structure": { | |
| "data-fields": " features = datasets.Features(\n {\n \"id\": datasets.Value(\"string\"),\n \"gem_id\": datasets.Value(\"string\"),\n \"turns\": [\n {\n \"speaker\": datasets.Value(\"string\"),\n \"text\": datasets.Value(\"string\"),\n \"nbest\": [\n {\n \"hyp\": datasets.Value(\"string\"),\n \"score\": datasets.Value(\"float\"),\n }\n ],\n }\n ],\n \"knowledge\": {\n \"domain\": datasets.Value(\"string\"),\n \"entity_name\": datasets.Value(\"string\"),\n \"title\": datasets.Value(\"string\"),\n \"body\": datasets.Value(\"string\"),\n },\n \"response\": datasets.Value(\"string\"),\n \"source\": datasets.Value(\"string\"),\n \"linearized_input\": datasets.Value(\"string\"),\n \"target\": datasets.Value(\"string\"),\n \"references\": [datasets.Value(\"string\")],\n }\n )\n\nnbest contains an nbest list of outputs generated by an ASR system along with their scores.\n\nknowledge defines the annotated grounding as well as its metadata", | |
| "structure-description": "It was kept compatible with MultiWox 2.X data.", | |
| "structure-example": "{'id': '0',\n 'gem_id': 'GEM-dstc10_track2_task2-test-0',\n 'turns': [{'speaker': 'U',\n 'text': \"hi uh i'm looking for restaurant in lower ha\",\n 'nbest': [{'hyp': \"hi uh i'm looking for restaurant in lower ha\",\n 'score': -25.625450134277344},\n {'hyp': \"hi uh i'm looking for restaurant in lower hai\",\n 'score': -25.969446182250977},\n {'hyp': \"hi uh i'm looking for restaurant in lower haig\",\n 'score': -32.816890716552734},\n {'hyp': \"hi uh i'm looking for restaurant in lower haigh\",\n 'score': -32.84316635131836},\n {'hyp': \"hi uh i'm looking for restaurant in lower hag\",\n 'score': -32.8637580871582},\n {'hyp': \"hi uh i'm looking for restaurant in lower hah\",\n 'score': -33.1048698425293},\n {'hyp': \"hi uh i'm looking for restaurant in lower hait\",\n 'score': -33.96509552001953},\n {'hyp': \"hi um i'm looking for restaurant in lower hai\",\n 'score': -33.97885513305664},\n {'hyp': \"hi um i'm looking for restaurant in lower haig\",\n 'score': -34.56083679199219},\n {'hyp': \"hi um i'm looking for restaurant in lower haigh\",\n 'score': -34.58711242675781}]},\n {'speaker': 'S',\n 'text': 'yeah definitely i can go ahead and help you with that ummm what kind of option in a restaurant are you looking for',\n 'nbest': []},\n {'speaker': 'U',\n 'text': 'yeah umm am looking for an expensive restaurant',\n 'nbest': [{'hyp': 'yeah umm am looking for an expensive restaurant',\n 'score': -21.272899627685547},\n {'hyp': 'yeah umm m looking for an expensive restaurant',\n 'score': -21.444047927856445},\n {'hyp': 'yeah umm a m looking for an expensive restaurant',\n 'score': -21.565458297729492},\n {'hyp': 'yeah ummm am looking for an expensive restaurant',\n 'score': -21.68832778930664},\n {'hyp': 'yeah ummm m looking for an expensive restaurant',\n 'score': -21.85947608947754},\n {'hyp': 'yeah ummm a m looking for an expensive restaurant',\n 'score': -21.980886459350586},\n {'hyp': \"yeah umm a'm looking for an expensive restaurant\",\n 'score': -22.613924026489258},\n {'hyp': \"yeah ummm a'm looking for an expensive restaurant\",\n 'score': -23.02935218811035},\n {'hyp': 'yeah um am looking for an expensive restaurant',\n 'score': -23.11180305480957},\n {'hyp': 'yeah um m looking for an expensive restaurant',\n 'score': -23.28295135498047}]},\n {'speaker': 'S',\n 'text': \"lemme go ahead and see what i can find for you ok great so i do ummm actually no i'm sorry is there something else i can help you find i don't see anything expensive\",\n 'nbest': []},\n {'speaker': 'U',\n 'text': \"sure ummm maybe if you don't have anything expensive how about something in the moderate price range\",\n 'nbest': [{'hyp': \"sure ummm maybe if you don't have anything expensive how about something in the moderate price range\",\n 'score': -27.492507934570312},\n {'hyp': \"sure umm maybe if you don't have anything expensive how about something in the moderate price range\",\n 'score': -27.75853729248047},\n {'hyp': \"sure ummm maybe if you don't have anything expensive how about something in the moderate price rang\",\n 'score': -29.44410514831543},\n {'hyp': \"sure umm maybe if you don't have anything expensive how about something in the moderate price rang\",\n 'score': -29.710134506225586},\n {'hyp': \"sure um maybe if you don't have anything expensive how about something in the moderate price range\",\n 'score': -31.136560440063477},\n {'hyp': \"sure um maybe if you don't have anything expensive how about something in the moderate price rang\",\n 'score': -33.088157653808594},\n {'hyp': \"sure ummm maybe i you don't have anything expensive how about something in the moderate price range\",\n 'score': -36.127620697021484},\n {'hyp': \"sure umm maybe i you don't have anything expensive how about something in the moderate price range\",\n 'score': -36.39365005493164},\n {'hyp': \"sure ummm maybe if yo don't have anything expensive how about something in the moderate price range\",\n 'score': -36.43605041503906},\n {'hyp': \"sure umm maybe if yo don't have anything expensive how about something in the moderate price range\",\n 'score': -36.70207977294922}]},\n {'speaker': 'S',\n 'text': 'ok moderate lemme go ahead and check to see what i can find for moderate ok great i do have several options coming up how does the view lounge sound',\n 'nbest': []},\n {'speaker': 'U',\n 'text': 'that sounds good ummm do they have any sort of happy hour special',\n 'nbest': [{'hyp': 'that sounds good ummm do they have any sort of happy hour special',\n 'score': -30.316478729248047},\n {'hyp': 'that sounds good umm do they have any sort of happy hour special',\n 'score': -30.958009719848633},\n {'hyp': 'that sounds good um do they have any sort of happy hour special',\n 'score': -34.463165283203125},\n {'hyp': 'that sounds good ummm do they have any sirt of happy hour special',\n 'score': -34.48350143432617},\n {'hyp': 'that sounds good umm do they have any sirt of happy hour special',\n 'score': -35.12503433227539},\n {'hyp': 'that sounds good ummm do they have any sord of happy hour special',\n 'score': -35.61939239501953},\n {'hyp': 'that sounds good umm do they have any sord of happy hour special',\n 'score': -36.26092529296875},\n {'hyp': 'that sounds good ummm do they have any sont of happy hour special',\n 'score': -37.697105407714844},\n {'hyp': 'that sounds good umm do they have any sont of happy hour special',\n 'score': -38.33863830566406},\n {'hyp': 'that sounds good um do they have any sirt of happy hour special',\n 'score': -38.630191802978516}]}],\n 'knowledge': {'domain': 'restaurant',\n 'entity_name': 'The View Lounge',\n 'title': 'Does The View Lounge offer happy hour?',\n 'body': 'The View Lounge offers happy hour.'},\n 'response': 'uhhh great question lemme go ahead and check that out for you ok fantastic so it looks like they do offer happy hour',\n 'source': 'sf_spoken',\n 'linearized_input': \"<U> hi uh i'm looking for restaurant in lower ha <S> yeah definitely i can go ahead and help you with that ummm what kind of option in a restaurant are you looking for <U> yeah umm am looking for an expensive restaurant <S> lemme go ahead and see what i can find for you ok great so i do ummm actually no i'm sorry is there something else i can help you find i don't see anything expensive <U> sure ummm maybe if you don't have anything expensive how about something in the moderate price range <S> ok moderate lemme go ahead and check to see what i can find for moderate ok great i do have several options coming up how does the view lounge sound <U> that sounds good ummm do they have any sort of happy hour special || knowledge domain: restaurant, entity: The View Lounge, title: Does The View Lounge offer happy hour?, information: The View Lounge offers happy hour.\",\n 'target': 'uhhh great question lemme go ahead and check that out for you ok fantastic so it looks like they do offer happy hour',\n 'references': ['uhhh great question lemme go ahead and check that out for you ok fantastic so it looks like they do offer happy hour']}", | |
| "structure-splits": "train: training set, val: validation set, test: test set", | |
| "structure-splits-criteria": "The track dataset originally only consists of a validation and test set in the spoken domain with noisy ASR transcripts.\nThe training set is taken from the predecessor task DSTC9 Track 1 and contains written conversations." | |
| }, | |
| "what": { | |
| "dataset": "The DSTC10 Track2 Task 2 follows the DSTC9 Track1 task, where participants have to implement knowledge-grounded dialog systems.\nThe training dataset is inherited from the DSTC9 challenge and is in the written domain, while the test set is newly collected and consists of noisy ASR transcripts.\nHence, the dataset facilitates building models for grounded dialog response generation." | |
| } | |
| }, | |
| "curation": { | |
| "original": { | |
| "is-aggregated": "no", | |
| "aggregated-sources": "N/A", | |
| "rationale": "We want to explore how conversational models perform on spoken data. ", | |
| "communicative": "This dataset aims to explore the robustness of conversational models when evaluated on spoken data. It has two aspects, multi-domain dialogue state tracking and conversation modeling with access to unstructured knowledge." | |
| }, | |
| "language": { | |
| "found": [], | |
| "crowdsourced": [], | |
| "created": "N/A", | |
| "machine-generated": "N/A", | |
| "validated": "not validated", | |
| "is-filtered": "not filtered", | |
| "filtered-criteria": "N/A", | |
| "obtained": [ | |
| "Other" | |
| ], | |
| "topics": "The conversations revolve around 5 domains (or topics): hotels, restaurants, attractions, taxi, train." | |
| }, | |
| "annotations": { | |
| "origin": "none", | |
| "rater-number": "N/A", | |
| "rater-qualifications": "N/A", | |
| "rater-training-num": "N/A", | |
| "rater-test-num": "N/A", | |
| "rater-annotation-service-bool": "no", | |
| "rater-annotation-service": [], | |
| "values": "N/A", | |
| "quality-control": [], | |
| "quality-control-details": "N/A" | |
| }, | |
| "consent": { | |
| "has-consent": "yes", | |
| "consent-policy": "N/A", | |
| "consent-other": "N/A", | |
| "no-consent-justification": "N/A" | |
| }, | |
| "pii": { | |
| "has-pii": "no PII", | |
| "no-pii-justification": "The subjects were instructed to conduct fictional conversations about booking restaurants or requesting fictional information.", | |
| "is-pii-identified": "N/A", | |
| "pii-identified-method": "N/A", | |
| "is-pii-replaced": "N/A", | |
| "pii-replaced-method": "N/A", | |
| "pii-categories": [] | |
| }, | |
| "maintenance": { | |
| "has-maintenance": "no", | |
| "description": "N/A", | |
| "contact": "N/A", | |
| "contestation-mechanism": "N/A", | |
| "contestation-link": "N/A", | |
| "contestation-description": "N/A" | |
| } | |
| }, | |
| "gem": { | |
| "rationale": { | |
| "sole-task-dataset": "yes", | |
| "distinction-description": "This dataset contains transcribed spoken interactions.", | |
| "contribution": "This dataset can be used to evaluate conversational models on spoken inputs (using ASR hypotheses). In particular, we can evaluate the models\u2019 ability to understand language by tracking the dialogue state, and their ability to generate knowledge-grounded responses.", | |
| "sole-language-task-dataset": "no", | |
| "model-ability": "We can measure the model\u2019s ability to understand language and to generate knowledge-grounded responses." | |
| }, | |
| "curation": { | |
| "has-additional-curation": "no", | |
| "modification-types": [], | |
| "modification-description": "N/A", | |
| "has-additional-splits": "no", | |
| "additional-splits-description": "N/A", | |
| "additional-splits-capacicites": "N/A" | |
| }, | |
| "starting": {} | |
| }, | |
| "results": { | |
| "results": { | |
| "other-metrics-definitions": "BLEU-1, BLEU-2, BLEU-3, BLEU-4, METEOR, ROGUE-1, ROGUE-2, ROGUE-L", | |
| "has-previous-results": "no", | |
| "current-evaluation": "N/A", | |
| "previous-results": "N/A", | |
| "model-abilities": "This dataset can be used to evaluate conversational models on spoken inputs (using ASR hypotheses). In particular, we can evaluate the models\u2019 ability to generate knowledge-grounded responses.", | |
| "metrics": [ | |
| "Other: Other Metrics" | |
| ] | |
| } | |
| }, | |
| "considerations": { | |
| "pii": { | |
| "risks-description": "There should be no risk related to PII as the subjects conduct fictional conversations." | |
| }, | |
| "licenses": { | |
| "dataset-restrictions-other": "N/A", | |
| "data-copyright-other": "N/A", | |
| "dataset-restrictions": [ | |
| "open license - commercial use allowed" | |
| ], | |
| "data-copyright": [ | |
| "open license - commercial use allowed" | |
| ] | |
| }, | |
| "limitations": {} | |
| }, | |
| "context": { | |
| "previous": { | |
| "is-deployed": "no", | |
| "described-risks": "N/A", | |
| "changes-from-observation": "N/A" | |
| }, | |
| "underserved": { | |
| "helps-underserved": "no", | |
| "underserved-description": "N/A" | |
| }, | |
| "biases": { | |
| "has-biases": "unsure", | |
| "bias-analyses": "N/A" | |
| } | |
| } | |
| } |