Personal index · Rev B

DWG NO. PL-2026-001

01

About

SHT 1 · GEN

I am a Ph.D. student at Yale CS, co-advised by Professors Rex Ying and David van Dijk in the Graph & Geometric Learning Lab and van Dijk Lab.

My research interests lie at the intersection of LLM agents, multi-agent systems, and efficient LLM adaptation and inference (MoE, LoRA, KV cache), together with memory management and retrieval-augmented generation. Previously I worked on AutoML, OOD generalization, causal discovery, and graph learning, with applications in biochemistry and biomedicine.

Before Yale, I received my Master's degree from Tsinghua University, advised by Prof. Wenwu Zhu, and was fortunate to intern with Menghua Wu in the Regina Barzilay Group at MIT CSAIL, and also at Alibaba Cloud under the Alibaba Innovative Research Program.

DWG 001 · REV B PLAN ACT CRITIC NOTES delegate verify remember 96 UNITS
Fig. 1 — a small society of agents.Plate 01 · Multi-agent memo
02

News

SHT 2 · LOG
  1. Two papers accepted to KDD 2026: NumCache and Fin-RATE.

  2. MoRSE is now on arXiv — our task-oriented multi-agent system with a mixture of role-subtask experts, specializing agents at both the task-structure and parameter levels for complex, long-horizon coding tasks.

  3. Started my Ph.D. at Yale, joining the Graph & Geometric Learning Lab and the van Dijk Lab.

Earlier entries · pre-Ph.D.
  1. Received my Master's degree from Tsinghua University.

  2. CARNAS was accepted to KDD 2025.

  3. One paper accepted to the NeurIPS 2024 AIDrugX workshop, from my internship at MIT CSAIL.

03

Publications

SHT 3 · IDX 07
Sort by Full index · Google Scholar ↗
F.02

MoRSE: Task-Oriented Multi-Agent System with Mixture of Role-Subtask Experts

Peiwen Li, Shiyang Zhang, Yangtian Zhang, Sizhuang He, David van Dijk, Rex Ying

arXiv 2026Under reviewPDF·Code
F.03

NumCache: Number-Preserving KV Cache Compression with Contrastive Retrieval for Financial Document QA

Eftychia Makri, Peiwen Li, Yidong Jiang, Junrong Chen, Jialin Chen, Ali Maatouk, Leandros Tassiulas, Eliot Brenner, Bing Xiang, Rex Ying

KDD 2026PDF·Code
F.04

Fin-RATE: A Real-world Financial Analytics and Tracking Evaluation Benchmark for LLMs on SEC Filings

Yidong Jiang, Junrong Chen, Eftychia Makri, Jialin Chen, Peiwen Li, Ali Maatouk, Leandros Tassiulas, Eliot Brenner, Bing Xiang, Rex Ying

KDD 2026PDF·Code·Dataset
F.05

Causal-aware Graph Neural Architecture Search under Distribution Shifts

Peiwen Li, Xin Wang, Zeyang Zhang, Ziwei Zhang, Fang Shen, Jialong Wang, Yang Li, Wenwu Zhu

KDD 2025PDF·Code
F.06

Learning to Refine Domain Knowledge for Biological Network Inference

Peiwen Li, Menghua Wu

NeurIPS 2024 · AIDrugX WorkshopPDF
F.07

RealTCD: Temporal Causal Discovery from Interventional Data with Large Language Model

Peiwen Li, Xin Wang, Zeyang Zhang, Yuan Meng, Fang Shen, Yue Li, Jialong Wang, Yang Li, Wenwu Zhu

CIKM 2024PDF·Code
F.08

Causal Discovery in Temporal Domain from Interventional Data

Peiwen Li, Yuan Meng, Xin Wang, Fang Shen, Yue Li, Jialong Wang, Wenwu Zhu

CIKM 2023PDF·Code
04

Experience

SHT 4 · REC

Yale University · Graduate Research Assistant

Graph & Geometric Learning Lab and van Dijk Lab, with Prof. Rex Ying and Prof. David van Dijk.

MIT CSAIL · Research Intern

Regina Barzilay group, with Menghua Wu — amortized inference for biological network inference.

Tsinghua University · Graduate Research Assistant

Multimedia and Network Big Data Lab, with Prof. Wenwu Zhu and Prof. Xin Wang.

Alibaba Cloud · Research Intern

Alibaba Innovative Research Program — causal-discovery-based anomaly detection for AIOps.

05

Awards & service

SHT 5 · APP
Comprehensive Excellence Scholarship, Tsinghua University
Outstanding Graduate Award of Beijing (top 4%)
Merit Student of Beijing (top 1%)
Meritorious Winner, Interdisciplinary Contest in Modeling
Second Prize, National Mathematics Competition for Undergraduates

Conference reviewer — NeurIPS, ACM SIGKDD, ACM MM

Journal reviewer — IEEE TPAMI, Pattern Recognition

projectPEIWEN LI — PERSONAL INDEX
drawn byP. LI
sheet01 / 01
scale1 : 1
revision2026.08 · REV B
doc