Subhajit Chaudhury

Research Scientist at IBM Research; Previously University of Tokyo; IIT Bombay; Jadavpur University (views are my own)

[Keerthiram Murugesan @keerthi166]

Our paper on imagination-based generalization work in #ReinforcementLearning for text adventure games is now accepted at #AAAI2022. Check out the paper at https://t.co/rJiTycK4ob

[Subhajit Chaudhury @CosmicSubha]

Can imagining scenes from textual descriptions helps in generalization in text-based games? Check out our recent paper for the answer (spoiler alert: it does!)

[Keerthiram Murugesan @keerthi166]

Is an🍎 by any other color an🍏too? Our recent paper on imagination-based generalization for RL models in text-based Interactive Fiction games (like TextWorld, Zork1, D&D, etc.) addresses this age-old question: https://t.co/rJiTycK4ob
- joint work with Subhajit & Kartik.

[Keerthiram Murugesan @keerthi166]

Is an🍎 by any other color an🍏too? Our recent paper on imagination-based generalization for RL models in text-based Interactive Fiction games (like TextWorld, Zork1, D&D, etc.) addresses this age-old question: https://t.co/rJiTycK4ob
- joint work with Subhajit & Kartik.

[AK @_akhaliq]

Bootstrapped Q-learning with Context Relevant Observation Pruning to Generalize in Text-based Games
pdf: https://t.co/awaqEcgjC9
abs: https://t.co/oL0U5PKM0F

[Subhajit Chaudhury @CosmicSubha]

@srvmshr This confirms my suspicions! Wonder if this is a cause for this studies conclusions: https://t.co/hfJ10GyS2T

[Subhajit Chaudhury @CosmicSubha]

@hardmaru It will be interesting to run GameGAN on Bipedal-Walker Hard/CarRacing and compare it with VAE latent codes.

[Daiki Kimura / 木村 大毅 @daiki_kimura]

Our paper "Adversarial Discriminative Attention for Robust Anomaly Detection" is accepted to #WACV2020. 今年のSSIIで発表し賞を頂いた、GANからのattentionを使った異常検出の研究が #WACV に通りました！Thank you for your efforts @CosmicSubha @chugmagaga @trlryuki.