Haochen Liu

Discrete-WAM enables compositional causal reasoning in autonomous driving, outperforming traditional methods that struggle with complex state-action dynamics.

Even state-of-the-art vision-language models frequently lie and hallucinate when playing social deduction games, raising serious questions about their reliability in real-world applications requiring grounded reasoning.

Predicting driver behavior in response to traffic conditions is now possible with a new world model that causally links external context to internal driver states.

LLMs struggle with conflicting medical evidence, but a clever two-stage agentic approach can reconcile discordant signals while preserving patient privacy.