Public Talks

11/2023: "Discrete Information Bottleneck for Human-Like Naming and Abstractions" - Amazon Alexa Science Hub Fellowship presentation.

07/2023: "Complexity-controlled Emergent Communication" - Invited talk for Emmanuel Dupoux's group at École des Hautes Etudes en Sciences Sociales.

05/2023: "Towards Interpretable Deep Reinforcement Learning with Human-Friendly Prototypes" - Notable paper talk at ICLR 2023.

11/2022: "Information-Theoretic Underpinnings of Generalization and Translation in Emergent Communication" - Spotlight talk at NeurIPS workshop on Information-Theoretic Principles in Cognitive Systems

11/2022: "Trading off Utility, Informativeness, and Complexity in Emergent Communication" - NeurIPS 2022.

07/2022: "Prototype Based Classification from Hierarchy to Fairness" - Spotlight talk at ICML 2022.

07/2022: "When Does Syntax Mediate Neural Language Model Performance? Evidence from Dropout Probes" - Oral presentation at NAACL 2022.

07/2022: "Towards Human-Agent Communication via the Information Bottleneck Principle" - Spotlight talk at RSS Workshop on Social Intelligence in Humans and Robots.

03/2022: "Grounded Word Embeddings via Emergent Communication" - Invited talk in the Computational Psycholinguistics Group at MIT

03/2022: "Enabling Semantic Emergent Communication" -Talk for the Multi-Agent Reinforcement Learning Reading Group at Berkeley

12/2021: "Emergent Discrete Communication in Semantic Spaces" - Poster Session at NeurIPS 2021

06/2021: "What if This Modified That? Syntactic Interventions via Counterfactual Embeddings" - ACL-IJCNLP Findings 2020: https://youtu.be/Pn1ca_N1l6g

04/2021: "And for This Next Bit- Voyager 2, Error Correct Codes, and You" - 1st Place in Aero/Astro Teaching Competition

02/2021: "Adversarially Guided Self-Play for Learning Social Conventions" - Poster Session at the 2021 NSF National Robotics Initiative

06/2020: "Efficient Social Convention Learning" - Talk to the Multi-Agent Reinforcement Learning reading group at Berkeley 

01/2019: "Information-Theoretic Disentanglement" - Poster Session at the MIT College of Computing Launch

10/2017: "DCG-UP-UP-Away" - Talk at International Symposium of Robotics Research