Tal Schuster

PhD Student

CSAIL, MIT

Biography

I'm a Ph.D. student at MIT CSAIL and a member of the NLP and Learn To Cure groups. I'm fortunate to be advised by Prof. Regina Barzilay.

Currently, I focus on developing ML and NLP methods to improve the robustness of models, prevent undesired outcomes, and increase their performance in challenging and realistic scenarios.
I have also worked on developing deep learning screening-based methods to improve cancer detection and risk assessment.

Before coming to MIT, I was an MSc student in the Computer Science Dep. at Tel-Aviv University, advised by Prof. Lior Wolf.

Interests

Machine Learning
Natural Language Processing
Computer Vision

Education

PhD in Computer Science, in progress

Massachusetts Institute of Technology
MSc in Computer Science, 2017

Tel Aviv University
BSc in Mathematics and Computer Science

Ben-Gurion University

Projects

Conformal Prediction

Predicting with confidence estimates

Word Embeddings

Dense representations of words

Fake News

Misinformation detection

Medical

Deep learning for cancer risk prediction

Metric Learning

Deep metric learning

Recent Publications

Click here to see all

Programming Puzzles

We introduce a new type of programming challenge called programming puzzles, as an objective and comprehensive evaluation of program …

Tal Schuster, Ashwin Kalyan, Oleksandr Polozov, Adam Tauman Kalai

June 2021

PDF Code Dataset

Consistent Accelerated Inference via Confident Adaptive Transformers

We develop a novel approach for confidently accelerating inference in the large and expensive multilayer Transformers that are now …

Tal Schuster*, Adam Fisch*, Tommi Jaakkola, Regina Barzilay

April 2021

PDF Code Project Project

Get Your Vitamin C! Robust Fact Verification with Contrastive Evidence

Typical fact verification models use retrieved written evidence to verify claims. Evidence sources, however, often change over time as …

Tal Schuster, Adam Fisch, Regina Barzilay

March 2021 In North American Chapter of the Association for Computational Linguistics (NAACL)

PDF Code Dataset Project News article

Efficient Conformal Prediction via Cascaded Inference with Expanded Admission

In this paper, we present a novel approach for conformal prediction (CP), in which we aim to identify a set of promising prediction …

Adam Fisch*, Tal Schuster*, Tommi Jaakkola, Regina Barzilay

January 2021 In International Conference on Learning Representations (ICLR)

PDF Code Project Project

Cross-Lingual Alignment of Contextual Word Embeddings, with Applications to Zero-shot Dependency Parsing

We introduce a novel method for multilingual transfer that utilizes deep contextual embeddings, pretrained in an unsupervised fashion. …

Tal Schuster*, Ori Ram*, Regina Barzilay, Amir Globerson

June 2019 In North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL)

PDF Code Project Slides Video

More Publications

Few-shot Conformal Prediction with Auxiliary Tasks
Adam Fisch, Tal Schuster, Tommi Jaakkola, Regina Barzilay.
2021.

PDF Project

Exploiting Rules to Enhance Machine Learning in Extracting Information From Multi-Institutional Prostate Pathology Reports
Enrico Santus, Tal Schuster, Amir M Tahmasebi, Clara Li, Adam Yala, Conor R Lanahan, Peter Prinsen, Scott F Thompson, Samuel Coons, Lance Mynderse, Regina Barzilay, Kevin Hughes.
In American Society of Clinical Oncology Journal (JCO), 2020.

PDF Project

Distilling the Evidence to Augment Fact Verification Models
Beatrice Portelli, Jason Zhao, Tal Schuster, Giuseppe Serra, Enrico Santus.
In Workshop on Fact Extraction and VERification (FEVER workshop at ACL), 2020.

PDF Project

The Limitations of Stylometry for Detecting Machine-Generated Fake News
Tal Schuster, Roei Schuster, Darsh J Shah, Regina Barzilay.
In Computational Linguistics, 2020.

PDF Dataset Project News article

Humpty Dumpty: Controlling Word Meanings via Corpus Poisoning
Roei Schuster, Tal Schuster, Yoav Meri, Vitaly Shmatikov.
In IEEE Symposium on Security and Privacy (IEEE S&P), 2020.

PDF Project Video

Automatic Fact-guided Sentence Modification
Darsh J Shah*, Tal Schuster*, Regina Barzilay.
In AAAI Conference on Artificial Intelligence (AAAI), 2020.

PDF Code Project News article

See all publications