Zichen Wen

2 papers in our library · 2025–2026

The profile is the library's match of the name; two people with one name are not always told apart.

Affiliations

From the bylines of their papers (OpenAlex, 2026-10-14), ORCID and Wikidata; institutions on papers are matched by machine.

Of its works in our library, with the number of works on each.

The most cited first (OpenAlex and our library).

Stop Looking for “Important Tokens” in Multimodal Language Models: Duplication Matters More (2025 · 44 citations)
Vision tokens make up most of the input to multimodal LLMs, which makes inference expensive. Most training-free pruning methods score each token's "importance" from attention maps and drop the low scorers. The authors argue this approach has four weaknesses. It treats scores as fixed even after other tokens are…
StreamMeCo: Long-Term Agent Memory Compression for Efficient Streaming Video Understanding (2026 · 3 citations)
Agent-based streaming video systems like M3-Agent store what they observe as a memory graph. Text nodes describe visual content, and face and voice nodes represent entities. As videos get longer, the number of text nodes grows sharply, and retrieval latency becomes too high for real-time question answering.

7 co-authors with a page, the most papers together first.

Where their works here were published, the most works first, with the editions of those that have a page.

Papers of our library cite their works 127 times (of the 149,439 papers whose bibliographies we read); 17 citing papers with a page here →.

Updated 31 October 2026