Alert button
Picture for Sainbayar Sukhbaatar

Sainbayar Sukhbaatar

Alert button

Learning to Reason and Memorize with Self-Notes

May 01, 2023
Jack Lanchantin, Shubham Toshniwal, Jason Weston, Arthur Szlam, Sainbayar Sukhbaatar

Figure 1 for Learning to Reason and Memorize with Self-Notes
Figure 2 for Learning to Reason and Memorize with Self-Notes
Figure 3 for Learning to Reason and Memorize with Self-Notes
Figure 4 for Learning to Reason and Memorize with Self-Notes
Viaarxiv icon

Think Before You Act: Unified Policy for Interleaving Language Reasoning with Actions

Apr 18, 2023
Lina Mezghani, Piotr Bojanowski, Karteek Alahari, Sainbayar Sukhbaatar

Figure 1 for Think Before You Act: Unified Policy for Interleaving Language Reasoning with Actions
Figure 2 for Think Before You Act: Unified Policy for Interleaving Language Reasoning with Actions
Figure 3 for Think Before You Act: Unified Policy for Interleaving Language Reasoning with Actions
Figure 4 for Think Before You Act: Unified Policy for Interleaving Language Reasoning with Actions
Viaarxiv icon

MINOTAUR: Multi-task Video Grounding From Multimodal Queries

Feb 16, 2023
Raghav Goyal, Effrosyni Mavroudi, Xitong Yang, Sainbayar Sukhbaatar, Leonid Sigal, Matt Feiszli, Lorenzo Torresani, Du Tran

Figure 1 for MINOTAUR: Multi-task Video Grounding From Multimodal Queries
Figure 2 for MINOTAUR: Multi-task Video Grounding From Multimodal Queries
Figure 3 for MINOTAUR: Multi-task Video Grounding From Multimodal Queries
Figure 4 for MINOTAUR: Multi-task Video Grounding From Multimodal Queries
Viaarxiv icon

Learning Goal-Conditioned Policies Offline with Self-Supervised Reward Shaping

Jan 05, 2023
Lina Mezghani, Sainbayar Sukhbaatar, Piotr Bojanowski, Alessandro Lazaric, Karteek Alahari

Figure 1 for Learning Goal-Conditioned Policies Offline with Self-Supervised Reward Shaping
Figure 2 for Learning Goal-Conditioned Policies Offline with Self-Supervised Reward Shaping
Figure 3 for Learning Goal-Conditioned Policies Offline with Self-Supervised Reward Shaping
Figure 4 for Learning Goal-Conditioned Policies Offline with Self-Supervised Reward Shaping
Viaarxiv icon

The CRINGE Loss: Learning what language not to model

Nov 10, 2022
Leonard Adolphs, Tianyu Gao, Jing Xu, Kurt Shuster, Sainbayar Sukhbaatar, Jason Weston

Figure 1 for The CRINGE Loss: Learning what language not to model
Figure 2 for The CRINGE Loss: Learning what language not to model
Figure 3 for The CRINGE Loss: Learning what language not to model
Figure 4 for The CRINGE Loss: Learning what language not to model
Viaarxiv icon

Walk the Random Walk: Learning to Discover and Reach Goals Without Supervision

Jun 23, 2022
Lina Mezghani, Sainbayar Sukhbaatar, Piotr Bojanowski, Karteek Alahari

Figure 1 for Walk the Random Walk: Learning to Discover and Reach Goals Without Supervision
Figure 2 for Walk the Random Walk: Learning to Discover and Reach Goals Without Supervision
Figure 3 for Walk the Random Walk: Learning to Discover and Reach Goals Without Supervision
Figure 4 for Walk the Random Walk: Learning to Discover and Reach Goals Without Supervision
Viaarxiv icon

DIRECTOR: Generator-Classifiers For Supervised Language Modeling

Jun 15, 2022
Kushal Arora, Kurt Shuster, Sainbayar Sukhbaatar, Jason Weston

Figure 1 for DIRECTOR: Generator-Classifiers For Supervised Language Modeling
Figure 2 for DIRECTOR: Generator-Classifiers For Supervised Language Modeling
Figure 3 for DIRECTOR: Generator-Classifiers For Supervised Language Modeling
Figure 4 for DIRECTOR: Generator-Classifiers For Supervised Language Modeling
Viaarxiv icon

Temporal Abstractions-Augmented Temporally Contrastive Learning: An Alternative to the Laplacian in RL

Mar 21, 2022
Akram Erraqabi, Marlos C. Machado, Mingde Zhao, Sainbayar Sukhbaatar, Alessandro Lazaric, Ludovic Denoyer, Yoshua Bengio

Figure 1 for Temporal Abstractions-Augmented Temporally Contrastive Learning: An Alternative to the Laplacian in RL
Figure 2 for Temporal Abstractions-Augmented Temporally Contrastive Learning: An Alternative to the Laplacian in RL
Figure 3 for Temporal Abstractions-Augmented Temporally Contrastive Learning: An Alternative to the Laplacian in RL
Figure 4 for Temporal Abstractions-Augmented Temporally Contrastive Learning: An Alternative to the Laplacian in RL
Viaarxiv icon

Hash Layers For Large Sparse Models

Jun 16, 2021
Stephen Roller, Sainbayar Sukhbaatar, Arthur Szlam, Jason Weston

Figure 1 for Hash Layers For Large Sparse Models
Figure 2 for Hash Layers For Large Sparse Models
Figure 3 for Hash Layers For Large Sparse Models
Figure 4 for Hash Layers For Large Sparse Models
Viaarxiv icon