SoUnD Framework: Analyzing (So)cial Representation in (Un)structured (D)ata
Not on the map. Why?
Abstract (arXiv)
The unstructured nature of data used in foundation model development is a challenge to systematic analyses for making data use and documentation decisions. From a Responsible AI perspective, these decisions often rely upon understanding how people are represented in data. We propose a framework designed to guide analysis of human representation in unstructured data and identify downstream risks. We apply the framework in two toy examples using the Common Crawl web text corpus (C4) and LAION-400M. We also propose a set of hypothetical action steps in service of dataset use, development, and documentation.
Key facts. Cited by 3 papers in the map (data through 7 Oct 2026). Builds on 52 of its 76 references that are in the map. First posted on arXiv 28 Nov 2023.
Builds on
52 of 76 references.
References · 52
- 2020Language Models are Few-Shot Learnerscited by 32,199
- 2019Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformercited by 14,491
- 2014ImageNet Large Scale Visual Recognition Challengecited by 8,441
- 2021On the Opportunities and Risks of Foundation Modelscited by 3,961
- 2020Extracting Training Data from Large Language Modelscited by 2,144
- 2021LAION-400M: Open Dataset of CLIP-Filtered 400 Million Image-Text Pairscited by 1,544
- 2018Datasheets for Datasetscited by 1,371
- 2016Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddingscited by 1,304
- 2020RealToxicityPrompts: Evaluating Neural Toxic Degeneration in Language Modelscited by 1,235
- 2018Model Cards for Model Reportingcited by 1,068
- 2021Ethical and social risks of harm from Language Modelscited by 957
- 2016Semantics derived automatically from language corpora contain human-like biasescited by 894
38 more
- 2021Deduplicating Training Data Makes Language Models Bettercited by 705
- 2018Gender Bias in Coreference Resolution: Evaluation and Debiasing Methodscited by 657
- 2016Inherent Trade-Offs in the Fair Determination of Risk Scorescited by 582
- 2021Documenting Large Webtext Corpora: A Case Study on the Colossal Clean Crawled Corpuscited by 481
- 2017Preventing Fairness Gerrymandering: Auditing and Learning for Subgroup Fairnesscited by 381
- 2020Closing the AI Accountability Gap: Defining an End-to-End Framework for Internal Algorithmic Auditingcited by 380
- 2021Persistent Anti-Muslim Bias in Large Language Modelscited by 376
- 2021Multimodal datasets: misogyny, pornography, and malignant stereotypescited by 326
- 2021Pre-Trained Models: Past, Present and Futurecited by 319
- 2020Data and its (dis)contents: A survey of dataset development and use in machine learning researchcited by 259
- 2017Word Embeddings Quantify 100 Years of Gender and Ethnic Stereotypescited by 251
- 2020Large image datasets: A pyrrhic win for computer vision?cited by 237
- 2018Women also Snowboard: Overcoming Bias in Captioning Modelscited by 236
- 2018A Simple Method for Commonsense Reasoningcited by 223
- 2018A comparative study of fairness-enhancing interventions in machine learningcited by 208
- 2015The MegaFace Benchmark: 1 Million Faces for Recognition at Scalecited by 199
- 2020Social Biases in NLP Models as Barriers for Persons with Disabilitiescited by 182
- 2022Data Cards: Purposeful and Transparent Dataset Documentation for Responsible AIcited by 169
- 2017No Classification without Representation: Assessing Geodiversity Issues in Open Data Sets for the Developing Worldcited by 152
- 2019Reducing Sentiment Bias in Language Models via Counterfactual Evaluationcited by 148
- 2018Aequitas: A Bias and Fairness Audit Toolkitcited by 145
- 2020Towards Accountability for Machine Learning Datasets: Practices from Software Engineering and Infrastructurecited by 138
- 2023What's In My Big Data?cited by 129
- 2019Towards a Critical Race Methodology in Algorithmic Fairnesscited by 124
- 2018The Adverse Effects of Code Duplication in Machine Learning Models of Codecited by 123
- 2021Re-imagining Algorithmic Fairness in India and Beyondcited by 102
- 2020Saving Face: Investigating the Ethical Concerns of Facial Recognition Auditingcited by 89
- 2019Predictive Inequity in Object Detectioncited by 77
- 2022CrowdWorkSheets: Accounting for Individual and Collective Identities Underlying Crowdsourced Dataset Annotationcited by 59
- 2021Addressing "Documentation Debt" in Machine Learning Research: A Retrospective Datasheet for BookCorpuscited by 46
- 2020OSCaR: Orthogonal Subspace Correction and Rectification of Biases in Word Embeddingscited by 40
- 2022Understanding Machine Learning Practitioners' Data Documentation Perceptions, Needs, Challenges, and Desideratacited by 28
- 2022The Principles of Data-Centric AI (DCAI)cited by 26
- 2020Principles to Practices for Responsible AI: Closing the Gapcited by 25
- 2021Representativeness in Statistics, Politics, and Machine Learningcited by 17
- 2021Understanding the Representation and Representativeness of Age in AI Data Setscited by 15
- 2022The Conflict Between Explainable and Accountable Decision-Making Algorithmscited by 13
- 2021Ground-Truth, Whose Truth? -- Examining the Challenges with Annotating Toxic Text Datasetscited by 11
50 most cited of 52 All 52 (JSON)
Built on by
Citing papers per year: 2024: 2; 2025: 1.
3 papers.