Our paper, FindingDory: A Benchmark to Evaluate Memory in Embodied Agents, was accepted to ECCV 2026.
Our paper, Let’s Think in Two Steps: Mitigating Agreement Bias in MLLMs with Self-Grounded Verification, was accepted to ICLR 2026.
Our paper, Memo: Training Memory-Efficient Embodied Agents with Reinforcement Learning, was accepted to NeurIPS 2025 as a Spotlight.
Our paper, Pre-trained Text-to-Image Diffusion Models Are Versatile Representation Learners for Control, was accepted to NeurIPS 2024 as a Spotlight.
Our paper, Look-Ahead Meta Learning for Continual Learning was accepted for Oral Presentation (top 1% of submitted works) at NeurIPS 2020