Amir Feder
Columbia University
I am a postdoctoral fellow at the Columbia Data Science Institute, working with David Blei on causal inference and natural language processing. I am also a visiting faculty researcher at Google Research, where I work on methods that leverage causality for medical NLP.
My research seeks to develop methods that integrate causality into natural language processing, and use them to build linguistically-informed algorithms for predicting and understanding human behavior. Deep language models are powerful prediction machines that seem to be improving on a daily basis. They have reached super-human performance on tasks that seemed impossible only a few years ago. Unfortunately, this improvement has resulted in unimaginably large models that often do not take into account the world they are operating in, and are both brittle and hard to interpret. My goal is to address such concerns using causal inference, to allow us to use these models to better predict and understand human behavior. Through the paradigm of causal machine learning, I aim to build bridges between machine learning and the social sciences.
Before joining Columbia, I received my PhD from the Technion, where I was advised by Roi Reichart and worked closely with Uri Shalit. In a previous (academic) life, I was an economics, statistics and history student at Tel Aviv University, the Hebrew University of Jerusalem and Northwestern University.
email: amir.feder at columbia dot edu
[google scholar] [semantic scholar] [dblp] [github]
Selected Publications (*=equal contribution)
CausaLM: Causal model explanation through counterfactual language models
Amir Feder, Nadav Oved, Uri Shalit, Roi Reichart
Computational Linguistics (CL), 2021 [arxiv] [code] [data] [pdf]
On Calibration and Out-of-domain Generalization
Yoav Wald*, Amir Feder*, Daniel Greenfeld, Uri Shalit
Neural information processing systems (NeurIPS) 2021 [arxiv] [pdf]
In the Eye of the Beholder: Robust Prediction with Causal User Modeling
Amir Feder, Guy Horowitz, Yoav Wald, Roi Reichart, and Nir Rosenfeld
Neural information processing systems (NeurIPS) 2022 [arxiv]
Causal Inference in Natural Language Processing: Estimation, Prediction, Interpretation and Beyond
Amir Feder*, Katherine A Keith*, Emaad Manzoor*, Reid Pryzant*, Dhanya Sridhar*, Zach Wood-Doughty*, Jacob Eisenstein*, Justin Grimmer*, Roi Reichart*, Margaret E Roberts*, Brandon M Stewart*, Victor Veitch*, Diyi Yang*
Transactions of the Association for Computational Linguistics (TACL), 2022 [arxiv] [reading list]
Tutorial on Causal Inference for NLP
Amir Feder*, Zhijing Jin*, Kun Zhang*
Empirical Methods in Natural Language Processing (EMNLP) 2022 [youtube] [slides (part 1), slides (part 2)]
All Publications (*=equal contribution)
An Invariant Learning Characterization of Controlled Text Generation
Claudia Shi*, Carolina Zheng*, Keyon Vafa, Amir Feder, David Blei
Association for Computational Linguistics (ACL) 2023 [Spotlight at NeurIPS 2022 Workshop on Robustness in Sequence Modeling]
In the Eye of the Beholder: Robust Prediction with Causal User Modeling
Amir Feder, Guy Horowitz, Yoav Wald, Roi Reichart, and Nir Rosenfeld
Neural information processing systems (NeurIPS) 2022 [arxiv]
CEBaB: Estimating the Causal Effects of Real-World Concepts on NLP Model Behavior
Eldar David Abraham*, Karel D'Oosterlinck*, Amir Feder*, Yair Ori Gat*, Atticus Geiger*, Christopher Potts*, Roi Reichart*, Zhengxuan Wu*
Neural information processing systems (NeurIPS) 2022 [arxiv] [data]
Building a Clinically-Focused Problem List From Medical Notes
Amir Feder, Itay Laish, Shashank Agarwal, Uri Lerner, Avel Atias, Cathy Cheung, Peter Clardy, Alon Peled-Cohen, Rachana Fellinger, Hengrui Liu, Lan Huong Nguyen, Birju Patel, Natan Potikha, Amir Taubenfeld, Liwen Xu, Seung Doo Yang, Ayelet Benjamini, Avinatan Hassidim
Proceedings of the 13th International Conference on Health Text Mining and Information Analysis (LOUHI), 2022 [pdf]
Section Classification in Clinical Notes with Multi-task Transformers
Fan Zhang, Itay Laish, Ayelet Benjamini, Amir Feder
Proceedings of the 13th International Conference on Health Text Mining and Information Analysis (LOUHI), 2022 [pdf]
Useful Confidence Measures: Beyond the Max Score
Gal Yona, Amir Feder, Itay Laish
NeurIPS 2022 Workshop on Distribution Shifts: Connecting Methods and Applications [arxiv]
Causal Inference in Natural Language Processing: Estimation, Prediction, Interpretation and Beyond
Amir Feder*, Katherine A Keith*, Emaad Manzoor*, Reid Pryzant*, Dhanya Sridhar*, Zach Wood-Doughty*, Jacob Eisenstein*, Justin Grimmer*, Roi Reichart*, Margaret E Roberts*, Brandon M Stewart*, Victor Veitch*, Diyi Yang*
ransactions of the Association for Computational Linguistics (TACL), 2022 [arxiv] [reading list]
DoCoGen: Domain Counterfactual Generation for Low Resource Domain Adaptation
Nitay Calderon, Eyal Ben-David, Amir Feder, Roi Reichart
Association for Computational Linguistics (ACL) 2022 [arxiv] [code] [pdf]
Shared computational principles for language processing in humans and deep language models
Ariel Goldstein, Zaid Zada*, Eliav Buchnik*, Mariano Schain*, Amy Price*, Bobbi Aubrey*, Samuel A Nastase*, Amir Feder*, Dotan Emanuel*, Alon Cohen*, Aren Jansen, Harshvardhan Gazula, Gina Choe, Aditi Rao, Catherine Kim, Colton Casto, Lora Fanda, Werner Doyle, Daniel Friedman, Patricia Dugan, Lucia Melloni, Roi Reichart, Sasha Devore, Adeen Flinker, Liat Hasenfratz, Omer Levy, Avinatan Hassidim, Michael Brenner, Yossi Matias, Kenneth A Norman, Orrin Devinsky, Uri Hasson
Nature Neuroscience 2022 [arxiv] [pdf]
Structured Understanding of Assessment and Plans in Clinical Documentation
Doron Stupp, Ronnie Barequet, I-Ching Lee, Eyal Oren, Amir Feder, Ayelet Benjamini, Avinatan Hassidim, Yossi Matias, Eran Ofek, Alvin Rajkomar
Machine Learning for Healthcare (ML4H) 2022 [arxiv]
On Calibration and Out-of-domain Generalization
Yoav Wald*, Amir Feder*, Daniel Greenfeld, Uri Shalit
Neural information processing systems (NeurIPS) 2021 [arxiv] [pdf]
Model compression for domain adaptation through causal effect estimation
Guy Rotman*, Amir Feder*, Roi Reichart
Transactions of the Association for Computational Linguistics (TACL), 2021 [arxiv] [code] [pdf]
Learning and evaluating a differentially private pre-trained language model
Shlomo Hoory, Amir Feder, Avichai Tendler, Sofia Erell, Alon Peled-Cohen, Itay Laish, Hootan Nakhost, Uri Stemmer, Ayelet Benjamini, Avinatan Hassidim, Yossi Matias
Findings of the Association for Computational Linguistics: EMNLP 2021 [pdf]
CausaLM: Causal model explanation through counterfactual language models
Amir Feder, Nadav Oved, Uri Shalit, Roi Reichart
Computational Linguistics (CL), 2021 [arxiv] [code] [data] [pdf]
Are VQA systems RAD? measuring robustness to augmented data with focused interventions
Daniel Rosenberg, Itai Gat, Amir Feder, Roi Reichart
Association for Computational Linguistics (ACL) 2021 [arxiv] [code] [pdf]
Predicting In-game Actions from Interviews of NBA Players
Amir Feder*, Nadav Oved*, Roi Reichart
Computational Linguistics (CL), 2020 [arxiv] [code] [pdf]
Active deep learning to detect demographic traits in free-form clinical notes
Amir Feder, Danny Vainstein, Roni Rosenfeld, Tzvika Hartman, Avinatan Hassidim, Yossi Matias
Journal of Biomedical Informatics (JBI), 2020 [data] [pdf]
An examination of the cryptocurrency pump-and-dump ecosystem
JT Hamrick, Farhang Rouhi, Arghya Mukherjee, Amir Feder, Neil Gandal, Tyler Moore, Marie Vasek
Information Processing & Management, 2019 [pdf]
The rise and fall of cryptocurrencies
Amir Feder, Neil Gandal, JT Hamrick, Tyler Moore, Marie Vasek
Workshop on the economics of information security 2018 [pdf]
The impact of DDoS and other security shocks on Bitcoin currency exchanges: Evidence from Mt. Gox
Amir Feder, Neil Gandal, JT Hamrick, Tyler Moore
Journal of Cybersecurity, 2018 [pdf]
Preprints
Measuring Causal Effects of Data Statistics on Language Model's 'Factual' Predictions
Yanai Elazar, Nora Kassner, Shauli Ravfogel, Amir Feder, Abhilasha Ravichander, Marius Mosbach, Yonatan Belinkov, Hinrich Schütze, Yoav Goldberg
[arxiv]
Correspondence between the layered structure of deep language models and temporal structure of natural language processing in the human brain
Ariel Goldstein, Eric Ham, Samuel A Nastase, Zaid Zada, Avigail Dabush, Bobbi Bobbi Aubrey, Mariano Schain, Harshvardhan Gazula, Amir Feder, Werner Doyle, Sasha Devore, Patricia Dugan, Daniel Friedman, Michael Brenner, Avinatan Hassidim, Orrin Devinsky, Adeen Flinker, Omer Levy, Uri Hasson
[arxiv]
Brain embeddings with shared geometry to artificial contextual embeddings, as a code for representing language in the human brain
Ariel Goldstein, Avigail Dabush, Bobbi Aubrey, Mariano Schain, Samuel A Nastase, Zaid Zada, Eric Ham, Zhuoqiao Hong, Amir Feder, Harshvardhan Gazula, Eliav Buchnik, Werner Doyle, Sasha Devore, Patricia Dugan, Daniel Friedman, Michael Brenner, Avinatan Hassidim, Orrin Devinsky, Adeen Flinker, Uri Hasson
[arxiv]
Explaining Classifiers with Causal Concept Effect (CaCE)
Yash Goyal, Amir Feder, Uri Shalit, Been Kim
2019 [arxiv]
Community Activities
Area Chair:
EACL: 2022-
Program committee member:
NeurIPS: 2019-
ICML: 2020-
AISTATS: 2022-
ACL Rolling Review: 2021-
ACL: 2021-
NAACL: 2021-
EMNLP: 2021-
CoNLL: 2020-
ICLR: 2022
AAAI: 2021
IJCAI: 2020-2021
TMLR: 2021-
FAccT: 2023-
Workshop organizer:
First Workshop on Causal Inference & NLP (CI+NLP), EMNLP 2021
With Katherine Keith, Emaad Manzoor, Reid Pryzant, Dhanya Sridhar, Zach Wood-Doughty,
Jacob Eisenstein, Justin Grimmer, Roi Reichart, Molly Roberts, Uri Shalit, Brandon Stewart, Victor Veitch, Diyi Yang
[url]
Tutorial organizer:
Tutorial on Causal Inference for NLP
With Kun Zhang, Zhijing Jin
Empirical Methods in Natural Language Processing (EMNLP) 2022 [youtube] [slides (part 1), slides (part 2)]