Ari Holtzman

Me.

My name is Ari Holtzman.

I am a PhD student at the University of Washington advised by Luke Zettlemoyer.

I work on looking at what Language Models actually do vs. what we think they do. My main finding is that they're probability distributions rather than readers or writers. My long-term goal is to create discursive machines.

Email: ahai at cs.washington.edu

Google Scholar

Publications

Surface Form Competition: Why the Highest Probability Isn't Always Right
Ari Holtzman, Peter West, Vered Shwartz, Yejin Choi, Luke Zettlemoyer
In EMNLP, 2021.
[PDF] [code][project]

Experience Grounds Language
Yonatan Bisk, Ari Holtzman, Jesse Thomason, Jacob Andreas, Yoshua Bengio, Joyce Y. Chai, Mirella Lapata, Angeliki Lazaridou, Jonathan May, Aleksandr Nisnevich, Nicolas Pinto, and Joseph Turian
In EMNLP, 2020.
[PDF]

Reflective Decoding: Beyond Unidirectional Generation with Off-the-Shelf Language Models
Peter West, Ximing Lu, Ari Holtzman, Chandra Bhagavatula, Jena Hwang, Yejin Choi
In ACL, 2021.
[PDF] [code][project]

Evaluating Machines by their Real-World Language Use
Rowan Zellers, Ari Holtzman, Elizabeth Clark, Lianhui (Karen) Qin, Ali Farhadi and Yejin Choi
[PDF] [project]

The Curious Case of Neural Text Degeneration
Ari Holtzman, Jan Buys, Maxwell Forbes, and Yejin Choi
In Proceedings of the International Conference on Learning Representations (ICLR), 2020
[PDF]

Abductive Commonsense Reasoning
Chandra Bhagavatula, Ronan Le Bras, Chaitanya Malaviya, Keisuke Sakaguchi, Ari Holtzman, Hannah Rashkin, Doug Downey, Scott Wenn-tau Yih, and Yejin Choi
In Proceedings of the International Conference on Learning Representations (ICLR), 2020
[PDF]

BottleSum: Unsupervised and Self-supervised Sentence Summarization using the Information Bottleneck Principle
Peter West, Ari Holtzman, Jan Buys and Yejin Choi
In Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP), 2019.
[PDF] [project]

Counterfactual Story Reasoning and Generation
Lianhui (Karen) Qin, Antoine Bosselut, Ari Holtzman, Chandra Bhagavatula, Elizabeth Clark, and Yejin Choi
In Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP), 2019.
[PDF] [code] [data]

Co-opNet: Cooperative Generator–Discriminator Networks for Abstractive Summarization with Narrative Flow
Saadia Gabriel, Antoine Bosselut, Ari Holtzman, Kyle Lo, Asli Celikyilmaz, and Yejin Choi
In submission.
[PDF]

Defending Against Neural Fake News
Rowan Zellers, Ari Holtzman, Hannah Rashkin, Yonatan Bisk, Ali Farhadi, Franziska Roesner, and Yejin Choi
In Conference on Neural Information Processing Systems (NeurIPS), 2019.
[PDF] [project]

Do Neural Language Representations Learn Physical Commonsense?
Maxwell Forbes, Ari Holtzman, and Yejin Choi
In Proceedings of the 41st Annual Conference of the Cognitive Science Society (CogSci), 2019.
[PDF][project] [code] [data]

HellaSwag: Can a Machine Really Finish Your Sentence?
Rowan Zellers, Ari Holtzman, Yonatan Bisk, Ali Farhadi, and Yejin Choi
In Proceedings of the Association for Computational Linguistic (ACL), 2019
[PDF] [project]

Tactical Rewind: Self-Correction via Backtracking in Vision-and-Language Navigation
Liyiming (Kay) Ke, Xiujun Li, Yonatan Bisk, Ari Holtzman, Zhe Gan, Jingjing Liu, Jianfeng Gao, Yejin Choi, and Siddhartha Srinivasa
In Conference on Computer Vision and Pattern Recognition (CVPR), 2019
[PDF] [ video] [code]

Sounding board: A user-centric and content-driven social chatbot
Hao Fang, Hao Cheng, Maarten Sap, Elizabeth Clark, Ari Holtzman, Yejin Choi,Noah A. Smith, and Mari Ostendorf
In Proceedings of the Association for Computational Linguistic (ACL) Demo Paper, 2018
[PDF]

Learning to Write with Cooperative Discriminators
Ari Holtzman, Jan Buys, Maxwell Forbes, Antoine Bosselut, David Golub, and Yejin Choi
In Proceedings of the Association for Computational Linguistic (ACL), 2018
[PDF] [code] [examples]

Simulating action dynamics with neural process networks
Antoine Bosselut, Omer Levy, Ari Holtzman, Corin Ennis, Dieter Fox, and Yejin Choi
In Proceedings of the International Conference on Learning Representations (ICLR), 2018
[PDF]

Connotation frames of power and agency in modern films
Maarten Sap, Marcella Cindy Prasetio, Ari Holtzman, Hannah Rashkin, and Yejin Choi
In Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP), 2017
[PDF] [ demo]