Yonatan Bisk – יונתן ביסק




CS Demographics

[CV][Google Scholar]

Conferences / Journals

Postdoc Work
Synthetic and Natural Noise Both Break Neural Machine Translation

Yonatan Belinkov and Yonatan Bisk

Learning Interpretable Spatial Operations in a Rich 3D Blocks World

Yonatan Bisk, Kevin Shih, Yejin Choi and Daniel Marcu
[Poster] [bib]

Natural Language Inference from Multiple Premises

Alice Lai, Yonatan Bisk, and Julia Hockenmaier

Natural Language Communication with Robots

Yonatan Bisk, Deniz Yuret, and Daniel Marcu
[Data] [Slides] [Video] [bib]

A dataset & models for isolating difficulties in language grounding
Supertagging with LSTMs

Ashish Vaswani, Yonatan Bisk, Kenji Sagae, and Ryan Musa
[Video] [bib]

Supertagging performance and errors correlate with conditioning context
PhD Work
Evaluating Induced CCG Parsers on Grounded Semantic Parsing

Yonatan Bisk, Siva Reddy, John Blitzer, Julia Hockenmaier and Mark Steedman

A dataset & evaluation for semantic parsing using induced grammars
Probing the Linguistic Strengths and Limitations of Unsupervised Grammar Induction
Yonatan Bisk and Julia Hockenmaier

Unsupervised Labeled dependencies and Analysis
[Video] [bib]

Labeled Grammar Induction with Minimal Supervision
ACL Short

Yonatan Bisk, Christos Christodoulopoulos and Julia Hockenmaier
[Poster] [bib]

Induced clusters get you 70% of the performance you get from gold tags. (*Table 3 corrected)
An HDP Model for Inducing Combinatory Categorial Grammars
TACL Vol 1.

Yonatan Bisk and Julia Hockenmaier
[Slides] [Video] [bib]

Parameter Sharing is essential to unsupervised performance
Simple Robust Grammar Induction with Combinatory Categorial Grammar
Yonatan Bisk and Julia Hockenmaier

A CCG Grammar can be automatically induced from nouns and verbs

Document-Topic Hierarchies from Document Graphs

Tim Weninger, Yonatan Bisk, and Jiawei Han

Normal-form parsing for Combinatory Categorial Grammars with generalized composition and type-raising

Julia Hockenmaier and Yonatan Bisk

Appropriately restricting Type-Raising and Composition greatly reduces ambiguity


Unsupervised Neural Hidden Markov Models
EMNLP Structure Pred Wksp

Ke Tran, Yonatan Bisk, Ashish Vaswani, Daniel Marcu, and Kevin Knight
[Slides] [bib]

EM style training of unsupervised HMM with neural networks
Towards a Dataset for Human Computer Communication via Grounded Language Acquisition
AAAI Wksp on Symbolic Cognitive Systems

Yonatan Bisk, Daniel Marcu and William Wong

Induction of Linguistic Structure with Combinatory Categorial Grammars
Yonatan Bisk and Julia Hockenmaier
NAACL Wksp on Inducing Linguistic Structure

[Corrected PASCAL Overview Paper] [Corrected Challenge Results]


Unsupervised Grammar Induction with Combinatory Categorial Grammars
Advised by Julia Hockenmaier
Committee: Jason Eisner, Dan Roth, Cheng Zhai
The Necessity of Separating Control and Logic When Grounding Language using Neuroevolution
Undergraduate Thesis Advised by Risto Miikkulainen

What is CCG?

Sadly there is no good intro to CCG that is geared at Computer Scientists that I'm aware of, but hopefully a coherent picture can be constructed without too much effort from the following links:
My Thesis: Chapter 3
Mark Steedman's CCG Page and NASSLI Tutorial or for the adventurous: The Syntactic Process