Teahose.
SIGN IN
NEW HERE — WHAT TEAHOSE DOES
We read the entire AI & tech firehose — so you don't have to.
PODPodcastsAll-In, No Priors, Acquired…
NEWNewslettersStratechery, Newcomer…
PAPPapersPhysical AI research
PHProduct Huntdaily launches
VCInvestor ScoutSequoia, a16z, Benchmark…
CLAUDE DISTILLS →
7 reads, 30 sec each — free, 6 AM ET.
+ a live graph of the companies, people & themes underneath.
HOME/PEOPLE/YE LI
// PERSON

Ye Li

ROLE RESEARCHERAT TSINGHUA UNIVERSITYMENTIONS 1LAST SEEN JUNE 2, 2026
// BIO

Ye Li is a PhD student at Tsinghua University, where his research focuses on model efficiency, reinforcement learning, and embodied AI. He is known for work on accelerating vision-language-action models and diffusion policies, including the SP-VLA framework for VLA model acceleration through joint model scheduling and token pruning. His publications also include contributions to adaptive ViT inference, KV cache compression, and offline meta reinforcement learning.

// RECENT MENTIONS
// SIGNALS
1 SIGNAL
01
product·arXiv Physical AI·JUNE 2, 2026

ElegantVLA provides a general acceleration framework for modern VLA pipelines with explicit action-generation modules, without modifying or retraining the base model.

Source

AI-extracted from podcast / newsletter / paper summaries. May contain errors.