iiinova — where your digital self keeps evolving

Yinglun Zhu

@yinglunzhu.bsky.social

Assistant Prof @ UC Riverside. Research on Efficient ML, RL, and LLMs. CS PhD @ UW Madison. yinglunz.com

Yinglun Zhu @yinglunzhu.bsky.social · 31/10/2025

Super excited to share Test-Time Matching (TTM), an iterative, self-improving algorithm that unlocks substantial compositional reasoning capabilities in multimodal models. TTM enables SigLIP-B16 (~0.2B params) to outperform GPT-4.1 on MMVP-VLM, establishing a new SOTA.

130

Yinglun Zhu @yinglunzhu.bsky.social · 14/10/2025

🚀Excited to share our new paper: Online Finetuning Decision Transformers with Pure RL Gradients RL drives reasoning in LLMs—but remains underexplored for online finetuning of Decision Transformers (DTs), where most methods still rely mainly on supervised objectives. Why?

140

Yinglun Zhu @yinglunzhu.bsky.social · 10/10/2025

Sharing new paper: Towards Multimodal Active Learning: Efficient Learning with Limited Paired Data We extend classical unimodal active learning to the multimodal AL with unaligned data, allowing data-efficient finetuning and pretraining of vision-language models such as CLIP and SigLIP. 1/3

130

Yinglun Zhu @yinglunzhu.bsky.social · 01/07/2025

🚀Excited to share our new paper: Strategic Scaling of Test-Time Compute: A Bandit Learning Approach. We turn test-time compute allocation into a bandit learning problem, achieving: ✅ +11.10% on MATH-500 ✅ +7.41% on LiveCodeBench Paper: arxiv.org/pdf/2506.12721 (1/3)

110

Reposted by Yinglun Zhu

Robert Nowak @rdnowak.bsky.social · 04/12/2024

There is a ton of interest in the question of whether AI can be be funny: www.bbc.com/future/artic.... Our paper at NeurIPS investigates the humor generation capabilities of the latest and greatest AI models using one of world’s largest humor datasets! arxiv.org/pdf/2406.10522

091

Yinglun Zhu @yinglunzhu.bsky.social · 03/12/2024

I’m recruiting multiple PhD students for Fall 2025 at UCR! If you’re interested in working on efficient ML, RL, and LLMs, please apply to the UCR CS/EE PhD program. Please visit yinglunz.com for detailed information on research directions and contact instructions.

032