Christopher Agia
@agiachris
PhD in Computer Science candidate @Stanford. My interests span learning for robotic planning, control, vision systems, and their interfacing representations.
What makes data “good” for robot learning? We argue: it’s the data that drives closed-loop policy success! Introducing CUPID 💘, a method that curates demonstrations not by "quality" or appearance, but by how they influence policy behavior, using influence functions. (1/6)
Q: How can we reduce action error in VLA policies to improve performance at deployment? 💡 Scale inference-time compute via VLM-based action verification! 🔍 Check out @jackyk02's work RoboMonkey 👇
✨ Test-Time Scaling for Robotics ✨ Excited to release 🤖 RoboMonkey, which characterizes test-time scaling laws for Vision-Language-Action (VLA) models and introduces a framework that significantly improves the generalization and robustness of VLAs! 🧵(1 / N) 🌐 Website:…
Excited to share CUPID 💘 won the Best Paper Award at the #RSS2025 RoboEval workshop! TL;DR: In 🤖 imitation learning, data quality is crucial. CUPID 💘 directly identifies “good” robot data, i.e., whether a demo will improve success rates. ➡️ big gains via data curation! 🧵👇
What makes data “good” for robot learning? We argue: it’s the data that drives closed-loop policy success! Introducing CUPID 💘, a method that curates demonstrations not by "quality" or appearance, but by how they influence policy behavior, using influence functions. (1/6)
As a researcher, it is immensly satisfying when the community tackles open problems from your previous work! In Gen2Act last year we showed how video generation models can be used zero-shot for manipulation. This paper takes the idea further via richer motion cues.
Research arc: ⏪ 2 yrs ago, we introduced VRB: learning from hours of human videos to cut down teleop (Gibson🙏) ▶️ Today, we explore a wilder path: robots deployed with no teleop, no human demos, no affordances. Just raw video generation magic 🙏 Day 1 of faculty life done! 😉…
“How will my model behave if I change the training data?” Recent(-ish) work w/ @logan_engstrom: we nearly *perfectly* predict ML model behavior as a function of training data, saturating benchmarks for this problem (called “data attribution”).
Presenting DemoDiffusion: An extremely simple approach enabling a pre-trained 'generalist' diffusion policy to follow a human-demonstration for a novel task during inference One-shot human imitation *without* requiring any paired human-robot data or online RL 🙂 1/n
If you are at RSS 2025 and interested to learn more, I will be presenting CUPID at the Robot Evaluation workshop! Details below: - Spotlight talk, 11:00am @ RTH211 - Poster session, 4:00pm @ Epstein Plaza Check out @Ransalu and @hocherie1's workshop! sites.google.com/stanford.edu/r…
What makes data “good” for robot learning? We argue: it’s the data that drives closed-loop policy success! Introducing CUPID 💘, a method that curates demonstrations not by "quality" or appearance, but by how they influence policy behavior, using influence functions. (1/6)
Grateful to share that I have been selected as a @KnightHennessy Scholar and will be joining the 2025 cohort! Deepest thanks to my advisor, @leto__jean, and to @sagan1934 and @FundlaCaixaCAT for empowering my journey from the start.
Meet the 2025 cohort of Knight-Hennessy scholars! These 84 scholars will join a diverse community of changemakers at Stanford to build lifelong friendships, deepen leadership skills, & collaborate with peers to address complex challenges facing the world. knight-hennessy.stanford.edu/news/knight-he…
📢 Only two months until our @ieee_ras_icra workshop on Safely Leveraging VLMs in Robotics! #ICRA2025 🎯 How can we safely leverage vision-language foundation models to expand robot deployment? 📅 Short papers & failure demos due 04/11/23 🌐 tinyurl.com/safe-vlm
📢 Announcing the first @ieee_ras_icra workshop on Safely Leveraging VLMs in Robotics! #ICRA2025 🎯 How can we safely leverage vision-language foundation models to expand robot deployment? 📅 Short papers & failure demos due 04/11/23 🌐 tinyurl.com/safe-vlm 🧵(1/5)
We released PARTNR, the largest benchmark to study human-robot collaboration in households, with +100K natural language tasks! PARTNR tests agents in key capabilities including: 🔍 Perceiving dynamic environments 🎯 Task planning and skill execution 🤝 Coordination with humans
Meta PARTNR is a research framework supporting seamless human-robot collaboration. Building on our research with Habitat, we’re open sourcing a large-scale benchmark, dataset and large planning model that we hope will enable the community to effectively train social robots.
New episode of NVIDIA #DRIVELabs features collaboration between #NVIDIAResearch and @Stanford on real-time anomaly detection and reactive planning with #LLMs. Winner of @RoboticsSciSys Outstanding Paper Award. #AI #automomousdriving #robotics nvda.ws/4hktF2F