📰 BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP 2020 — accepted papers

BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP 2020 is the 2020 edition of the BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP, held in Online, November 2020 (publisher: Association for Computational Linguistics). The ACL Anthology lists 32 papers of it; our library holds 3 with our summaries. Below: the most cited papers, every accepted paper, and ours with their topics, authors and institutions.

Most cited papers

By their OpenAlex citations; a paper with our summary links to its page here, the others to the original.

1. The elephant in the interpretability room: Why use attention as explanation when we have saliency methods?Jasmijn Bastings, Katja Filippova150 citations · our summary
2. What Happens To BERT Embeddings During Fine-tuning?Amil Merchant, Elahe Rahimtoroghi, Ellie Pavlick et al.139 citations · our summary
3. BERTs of a feather do not generalize together: Large variability in generalization across models with similar test set performanceR. Thomas McCoy, Junghyun Min, Tal Linzen98 citations
4. How does BERT capture semantics? A closer look at polysemous wordsDavid Yenicelik, Florian Schmidt, Yannic Kilcher50 citations
5. Neural Natural Language Inference Models Partially Embed Theories of Lexical Entailment and NegationAtticus Geiger, Kyle Richardson, Christopher Potts44 citations · our summary
6. Searching for a Search Method: Benchmarking Search Algorithms for Generating NLP Adversarial ExamplesJin Yong Yoo, John Morris, Eli Lifland et al.43 citations
7. It's not Greek to mBERT: Inducing Word-Level Translations from Multilingual BERTHila Gonen, Shauli Ravfogel, Yanai Elazar et al.40 citations
8. BERTnesia: Investigating the capture and forgetting of knowledge in BERTJonas Wallat, Jaspreet Singh, Avishek Anand30 citations
9. Discovering the Compositional Structure of Vector Representations with Role Learning NetworksPaul Soulos, R. Thomas McCoy, Tal Linzen et al.26 citations
10. Linguistically-Informed Transformations (LIT): A Method for Automatically Generating Contrast SetsChuanrong Li, Lin Shengshuo, Zeyu Liu et al.25 citations

All accepted papers (32)

The most cited first; 3 with our summary link to their page here, the others to the original.

1. The elephant in the interpretability room: Why use attention as explanation when we have saliency methods?Jasmijn Bastings, Katja Filippova150 citations · our summary
2. What Happens To BERT Embeddings During Fine-tuning?Amil Merchant, Elahe Rahimtoroghi, Ellie Pavlick et al.139 citations · our summary
3. BERTs of a feather do not generalize together: Large variability in generalization across models with similar test set performanceR. Thomas McCoy, Junghyun Min, Tal Linzen98 citations
4. How does BERT capture semantics? A closer look at polysemous wordsDavid Yenicelik, Florian Schmidt, Yannic Kilcher50 citations
5. Neural Natural Language Inference Models Partially Embed Theories of Lexical Entailment and NegationAtticus Geiger, Kyle Richardson, Christopher Potts44 citations · our summary
6. Searching for a Search Method: Benchmarking Search Algorithms for Generating NLP Adversarial ExamplesJin Yong Yoo, John Morris, Eli Lifland et al.43 citations
7. It's not Greek to mBERT: Inducing Word-Level Translations from Multilingual BERTHila Gonen, Shauli Ravfogel, Yanai Elazar et al.40 citations
8. BERTnesia: Investigating the capture and forgetting of knowledge in BERTJonas Wallat, Jaspreet Singh, Avishek Anand30 citations
9. Discovering the Compositional Structure of Vector Representations with Role Learning NetworksPaul Soulos, R. Thomas McCoy, Tal Linzen et al.26 citations
10. Linguistically-Informed Transformations (LIT): A Method for Automatically Generating Contrast SetsChuanrong Li, Lin Shengshuo, Zeyu Liu et al.25 citations
11. The Explanation Game: Towards Prediction Explainability through Sparse CommunicationMarcos Treviso, André F. T. Martins25 citations
12. Do Language Embeddings capture Scales?Xikun Zhang, Deepak Ramachandran, Ian Tenney et al.22 citations
13. The EOS Decision and Length ExtrapolationBenjamin Newman, John Hewitt, Percy Liang et al.22 citations
14. On the Interplay Between Fine-tuning and Sentence-Level Probing for Linguistic Knowledge in Pre-Trained TransformersMarius Mosbach, Anna Khokhlova, Michael A. Hedderich et al.17 citations
15. Emergent Language Generalization and Acquisition Speed are not tied to CompositionalityEugene Kharitonov, Marco Baroni16 citations
16. Structured Self-Attention Weights Encode Semantics in Sentiment AnalysisZhengxuan Wu, Thanh-Son Nguyen, Desmond C. Ong16 citations
17. Investigating Novel Verb Learning in BERT: Selectional Preference Classes and Alternation-Based Syntactic GeneralizationTristan Thrush, Ethan Wilcox, Roger Levy13 citations
18. Probing for Multilingual Numerical Understanding in Transformer-Based Language ModelsDevin Johnson, Denise Mak, Andrew Barker et al.13 citations
19. This is a BERT. Now there are several of them. Can they generalize to novel words?Coleman Haley11 citations
20. Defining Explanation in an AI ContextTejaswani Verma, Christoph Lingenfelder, Dietrich Klakow9 citations
21. Dissecting Lottery Ticket Transformers: Structural and Behavioral Study of Sparse Neural Machine TranslationRajiv Movva, Jason Zhao9 citations
22. Examining the rhetorical capacities of neural language modelsZining Zhu, Chuer Pan, Mohamed Abdalla et al.8 citations
23. Unsupervised Distillation of Syntactic Information from Contextualized Word RepresentationsShauli Ravfogel, Yanai Elazar, Jacob Goldberger et al.8 citations
24. Evaluating Attribution Methods using White-Box LSTMsYiding Hao7 citations
25. diagNNose: A Library for Neural Activation AnalysisJaap Jumelet7 citations

See all 32 papers in BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP 2020 — accepted papers →

📄 Papers with our summary in our library (3)

The most cited first (OpenAlex citations plus our library's).

1. The elephant in the interpretability room: Why use attention as explanation when we have saliency methods?BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLPJasmijn Bastings, Katja Filippova2020 · 219 citations
2. What Happens To BERT Embeddings During Fine-tuning?BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLPAmil Merchant, Elahe Rahimtoroghi, Ellie Pavlick, et al.2020 · 218 citations
3. Neural Natural Language Inference Models Partially Embed Theories of Lexical Entailment and NegationAtticus Geiger, Kyle Richardson, Christopher Potts2020 · 112 citations

🧭 Top topics

Topics of its works in our library, with the number of works on each.

🧮Reasoning (1)🔬Interpretability (1)English (1)

📚 Fields

🖥️Computer Science (3)🎭Arts and Humanities (2)

🧑‍🔬 Top authors

Authors with a page who published here most often: their works here, and the citations of those works.

🏛️ Top institutions

Institutions on the bylines of its works here, by the number of works.

🏢GoogleMountain View · United States2 works here🎓University of California, Santa CruzSanta Cruz · United States1 work here🎓Brown UniversityProvidence · United States1 work here🔬Allen Institute for Artificial IntelligenceSeattle · United States1 work here🎓Stanford UniversityStanford · United States1 work here

Place and dates from the event's site or the ACL Anthology; works and citations from OpenAlex and our library.

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/venues/blackboxnlp/2020/