Scholar
Patrick Yin
Google Scholar ID: AMVmM84AAAAJ
University of Washington
Citations & Impact
Citations
1,196
H-index
7
i10-index
7
Publications
8
Co-authors
0
Publications
Resume (English only)
Academic Achievements
- ICLR 2025: Proposed Simulation-Guided Fine-Tuning (SGFT), a sim2real framework for accelerating real-world RL (first author, equal contribution)
- RSS 2024: Contributed to DROID, a large-scale in-the-wild robot manipulation dataset with 76k trajectories
- ICLR 2024 Oral: Developed ASID, a system that refines simulation models using minimal real-world data for sim-to-real transfer
- ICLR 2024 Spotlight: Investigated architectural choices to stabilize contrastive RL for robotic goal reaching
- ICRA 2024 Best Paper: Contributed to Open X-Embodiment, an open-source dataset with 1M+ real robot trajectories across 22 robot embodiments
- CoRL 2022 Oral: Proposed FLAP, a framework leveraging diverse offline data for rapid fine-tuning to new visuomotor tasks
- IROS 2022: Introduced Planning to Practice (PTP) for efficient online fine-tuning of long-horizon goal-conditioned policies (equal contribution)
- ICML 2022: Contributed to work on bisimulation-based state abstraction for goal-conditioned reinforcement learning
Co-authors