A
Abhishek Das
Researcher at Facebook
Publications - 61
Citations - 15366
Abhishek Das is an academic researcher from Facebook. The author has contributed to research in topics: Dialog box & Computer science. The author has an hindex of 27, co-authored 52 publications receiving 9447 citations. Previous affiliations of Abhishek Das include Georgia Institute of Technology.
Papers
More filters
ForceNet: A Graph Neural Network for Large-Scale Quantum Chemistry Simulation
Weihua Hu,Muhammed Shuaibi,Abhishek Das,Siddharth Goyal,Anuroop Sriram,Jure Leskovec,Devi Parikh,Larry Zitnick +7 more
TL;DR: In this paper, a graph neural network is used to estimate per-atom forces in a 3D molecular network, which can be applied to accelerate catalyst discovery for renewable energy applications.
Proceedings ArticleDOI
Connecting Language and Vision to Actions
TL;DR: This tutorial will comprehensively review existing state-of-the-art approaches to selected tasks such as image captioning, visual question answering (VQA) and visual dialog, presenting the key architectural building blocks and novel algorithms used to train models for these tasks.
Feel The Music: Automatically Generating A Dance For An Input Song.
TL;DR: In this paper, the authors present a general computational approach that enables a machine to generate a dance for any input music by encoding intuitive, flexible heuristics for what a "good" dance is: the structure of the dance should align with the music.
Proceedings ArticleDOI
IR-VIC: Unsupervised Discovery of Sub-goals for Transfer in RL
Nirbhay Modhe,Prithvijit Chattopadhyay,Mohit Sharma,Abhishek Das,Devi Parikh,Devi Parikh,Dhruv Batra,Dhruv Batra,Ramakrishna Vedantam +8 more
TL;DR: In this paper, a variational intrinsic control framework is proposed to identify sub-goals useful for exploration in sequential decision making tasks under partial observability, which maximizes empowerment.
Journal ArticleDOI
EquiformerV2: Improved Equivariant Transformer for Scaling to Higher-Degree Representations
TL;DR: In this article , the authors propose EquiformerV2, which improves the scalability of the Equivariant Transformers to higher degrees of equivariant representations by introducing attention re-normalization, separable activation, and separable layer normalization.