@SeonghyeonYe Profile picture

Seonghyeon Ye

@SeonghyeonYe

PhD student KAIST AI (@kaist_ai) & Research intern at Microsoft Research

Similar User
Jie Huang photo

@jefffhj

Dongkeun Yoon photo

@dongkeun_yoon

Joel Jang photo

@jang_yoel

Minjoon Seo photo

@seo_minjoon

Jinheon Baek photo

@jinheonbaek

Sohee Yang @ ACL 2024 photo

@soheeyang_

Dongkwan Kim photo

@_dongkwan_kim

Sungdong Kim photo

@SungdongKim4

Hoyeon Chang photo

@hoyeon_chang

Hyeon min Yun photo

@hyeonmin_Lona

Grant Slatton photo

@GrantSlatton

Sangmin Bae photo

@raymin0223

hyunji amy lee photo

@hyunji_amy_lee

Jay Shin photo

@jshin491

Jaehong Yoon photo

@jaeh0ng_yoon

Pinned

🚀 First step to unlocking Generalist Robots! Introducing 🤖LAPA🤖, a new SOTA open-sourced 7B VLA pretrained without using action labels. 💪SOTA VLA trained with Open X (outperforming OpenVLA on cross and multi embodiment) 😯LAPA enables learning from human videos, unlocking…


We won the 🏆Best Paper Award at #Corl2024 LangRob workshop! Also check out our updated codebase: github.com/LatentActionPr…

Excited to share that 𝐋𝐀𝐏𝐀 has won the Best Paper Award at the CoRL 2024 Language and Robot Learning workshop, selected among 75 accepted papers! Both @SeonghyeonYe and I come from NLP backgrounds, where everything is built around tokenization. Drawing inspiration from…

Tweet Image 1


Seonghyeon Ye Reposted

Really excited to share what I've been working on with my colleagues at Physical Intelligence! We've developed a prototype robotic foundation model that can fold laundry, assemble a box, bus a table, and many other things. We've written a paper and blog post about it. 🧵👇


Seonghyeon Ye Reposted

D4RL is a great benchmark, but is saturated. Introducing OGBench, a new benchmark for offline goal-conditioned RL and offline RL! Tasks include HumanoidMaze, Puzzle, Drawing, and more 🙂 Project page: seohong.me/projects/ogben… GitHub: github.com/seohongpark/og… 🧵↓


Seonghyeon Ye Reposted

Mobile AI assistants (like Apple Intelligence) offer useful features using personal information. But how can we ensure they’re safe to use? Introducing MobileSafetyBench—a benchmark to assess the safety of mobile AI assistants. PDF & Code: mobilesafetybench.github.io 1/N 🧵


Seonghyeon Ye Reposted

Tired of endlessly teleoperating your robot in order to train it? Introducing SkillMimicGen, a data generation system that automatically scales robot imitation learning by synthesizing demos through integrating motion planning and demo adaptation. skillgen.github.io 1/


Seonghyeon Ye Reposted

Excited to introduce 𝐋𝐀𝐏𝐀: the first unsupervised pretraining method for Vision-Language-Action models. Outperforms SOTA models trained with ground-truth actions 30x more efficient than conventional VLA pretraining 📝: arxiv.org/abs/2410.11758 🧵 1/9


Seonghyeon Ye Reposted

LAPA: Latent Action Pretraining from Videos - Proposes a method to learn from internet-scale videos w/o action labels - Outperforms the SotA VLA model trained with robotic action labels on real-world manipulation tasks proj: latentactionpretraining.github.io abs: arxiv.org/abs/2410.11758


Seonghyeon Ye Reposted

❓Do LLMs maintain the capability of knowledge acquisition throughout pretraining? If not, what is driving force behind it? ❗Our findings reveal that decreasing knowledge entropy hinders knowledge acquisition and retention as pretraining progresses. 📄arxiv.org/abs/2410.01380

Tweet Image 1

Seonghyeon Ye Reposted

I wrote a little blog post about robotic foundation models (generalist robotic policies): sergeylevine.substack.com/p/the-promise-…


Seonghyeon Ye Reposted

Meet Molmo: a family of open, state-of-the-art multimodal AI models. Our best model outperforms proprietary systems, using 1000x less data. Molmo doesn't just understand multimodal data—it acts on it, enabling rich interactions in both the physical and virtual worlds. Try it…


Seonghyeon Ye Reposted

Evaluation in robot learning papers, or, please stop using only success rate a paper and a 🧵 arxiv.org/abs/2409.09491


Seonghyeon Ye Reposted

Humans learn and improve from failures. Similarly, foundation models adapt based on human feedback. Can we leverage this failure understanding to enhance robotics systems that use foundation models? Introducing AHA—a vision-language model for detecting and reasoning over…


Loading...

Something went wrong.


Something went wrong.