Rose Yu (@yuqirose) 's Twitter Profile
Rose Yu

@yuqirose

Machine Learning Prof @UCSanDiego, Previously @googlecloud, @Northeastern, @Caltech, @USC, #Physics-Guided #AI, MIT TR-35 Innovator.

ID: 1890694861

linkhttp://roseyu.com calendar_today21-09-2013 16:17:56

1,1K Tweet

8,8K Followers

561 Following

UCSD CSE (@ucsd_cse) 's Twitter Profile Photo

Improving Deep Learning With a Little Help From Physics: UC San Diego Assoc Professor Rose Yu makes #AI better, faster & smarter. bit.ly/3YCYF6v @JSOE  @quantamag #PhysicsGuidedAI #DeepLearning

Anshul Kundaje (anshulkundaje@bluesky) (@anshulkundaje) 's Twitter Profile Photo

Secretary Marco Rubio This is the most moronic self goal ever. U do realize those students will simply go elsewhere (including China) & still be awesome & still drive innovation.

National Academy of Sciences (@thenasciences) 's Twitter Profile Photo

How can we integrate physical laws into AI? 🧠⚙️ On June 4, join Rose Yu (Rose Yu) of UC San Diego for a Distinctive Voices talk on Physics-Guided AI—a framework that fuses #physics with machine learning to boost accuracy, efficiency & insight. Register: ow.ly/A1RK50VWAug

How can we integrate physical laws into AI? 🧠⚙️ On June 4, join Rose Yu (<a href="/yuqirose/">Rose Yu</a>) of <a href="/UCSanDiego/">UC San Diego</a> for a Distinctive Voices talk on Physics-Guided AI—a framework that fuses #physics with machine learning to boost accuracy, efficiency &amp; insight. Register: ow.ly/A1RK50VWAug
Bo Zhao (@bozhao__) 's Twitter Profile Photo

When and why are neural network solutions connected by low-loss paths? In our #ICML2025 paper, we show that mode connectivity often arises from symmetries—transformations of parameters that leave the network’s output unchanged. Paper: arxiv.org/abs/2505.23681 (1/6)

When and why are neural network solutions connected by low-loss paths?

In our #ICML2025 paper, we show that mode connectivity often arises from symmetries—transformations of parameters that leave the network’s output unchanged.

Paper: arxiv.org/abs/2505.23681
(1/6)