Research Interests
- Forward-looking fundamental research on Large Language Models, focusing on their design, computational properties, applications in real-world contexts, and transformative potential in scientific discovery.
- Development of practical and innovative frameworks to detect, diagnose, and mitigate hallucinations and reasoning limitations in LLMs, applicable across diverse modalities.
- Addressing educational challenges in an AI-driven future by fostering Renaissance-like thinking, achieving exceptional performance with AI tools, and strengthening core abilities without tool dependence—ultimately empowering individuals to surpass pre-AI capabilities. Developed LLM4LLM Project, an innovative curriculum using LLMs to explain their own mechanisms through interactive learning. Recent talk
Selected Publications
Open-source frameworks for scalable machine learning and graph machine learning.
Initiated popular scalable machine learning and graph neural network framework through robust, open-source platforms (~2.9K and ~1.9K citations, for MXNet and DGL, respectively).
- Minjie Wang, Da Zheng, Zihao Ye, Quan Gan, Mufei Li, Xiang Song, Jinjing Zhou, Chao Ma, Lingfan Yu, Yu Gai, Tianjun Xiao, Tong He, George Karypis, Jinyang Li, Zheng Zhang. Deep Graph Library: A Graph-Centric, Highly-Performant Package for Graph Neural Networks. arXiv:1909.01315. Sep 2019
- Chen T, Li M, Li Y, Lin M, Wang N, Wang M, Xiao T, Xu B, Zhang C, Zhang Z. Mxnet: A flexible and efficient machine learning library for heterogeneous distributed systems. arXiv:1512.01274. Dec 2015.
Large Language Models
Rethinking LLMs as computing devices and addressing their key limitations; mechanisms and benchmarks to perform hallucination detection.
- Zhang Z. Comprehension Without Competence: Architectural Limits of LLMs in Symbolic Computation and Reasoning. TMLR, Nov 2025./li>
- Ru D, Qiu L, Hu X, Zhang T, Shi P, Chang S, Jiayang C, Wang C, Sun S, Li H, Zhang Z. Ragchecker: A fine-grained framework for diagnosing retrieval-augmented generation. NeurIPS 2024 D&B Track. 2024 Aug 15.
- Hu X, Ru D, Qiu L, Guo Q, Zhang T, Xu Y, Luo Y, Liu P, Zhang Y, Zhang Z. RefChecker: Reference-based Fine-grained Hallucination Checker and Benchmark for Large Language Models. EMNLP 2024.
Graph Neural Networks, Algorithms and Applications
GNN algorithms and system co-design, application and benchmarking of GNN for relational data, and applications in text and multi-agent trajectory prediction.
- Huang K, Jiang H, Wang M, Xiao G, Wipf D, Song X, Gan Q, Huang Z, Zhai J, Zhang Z. FreshGNN: Reducing Memory Access via Stable Historical Embeddings for Graph Neural Network Training. VLDB Endowment. 2024 Feb 1;17(6):1473-86.
- Wang M, Gan Q, Wipf D, Cai Z, Li N, Tang J, Zhang Z, et al. 4DBInfer: A 4D Benchmarking Toolbox for Graph-Centric Predictive Modeling on Relational DBs. arXiv:2404.18209. 28 Apr 2024.
- Guo Q, Jin Z, Wang Z, Qiu X, Zhang W, Zhu J, Zhang Z, Wipf D. Fork or fail: Cycle-consistent training with many-to-one mappings. AISTATS. 2021 Mar; pp. 1828-36.
- Guo Q, Jin Z, Qiu X, Zhang W, Wipf D, Zhang Z. Cyclegt: Unsupervised graph-to-text and text-to-graph generation via cycle training. arXiv:2006.04702. 2020 Jun 8.
- Li L, Yao J, Wenliang L, et al. Grin: Generative relation and intention network for multi-agent trajectory prediction. NeurIPS. 2021 Dec;34:27107-18.
Transformer Architecture Optimizations
Early solutions for addressing long-range context challenges, influencing subsequent innovations in attention mechanisms.
- Guo Q, Qiu X, Liu P, Shao Y, Xue X, Zhang Z. Star-transformer. NeurIPS 2019. 2019 Feb 25.
- Ye Z, Guo Q, Gan Q, Qiu X, Zhang Z. Bp-transformer: Modelling long-range context via binary partitioning. ICLR 2020. 2019 Nov 11.
Core Computer Vision Problems
Research key challenges in computer vision, object-centric learning and sequential image attention.
- Seitzer M, Horn M, Zadaianchuk A, et al. Bridging the gap to real-world object-centric learning. ICLR 2023.
- Welleck S, Mao J, Cho K, Zhang Z. Saliency-based sequential image attention with multiset prediction. NeurIPS. 2017
- Xiao T, Zhang J, Yang K, Peng Y, Zhang Z. Error-driven incremental learning in deep convolutional neural network for large-scale image classification. ACM Multimedia. 2014 Nov; pp. 177-86.
- Xiao T, Xu Y, Yang K, Zhang J, Peng Y, Zhang Z. The application of two-level attention models in deep convolutional neural network for fine-grained image classification. CVPR. 2015; pp. 842-50.
More details: https://hackmd.io/H2AnjPGmRv-1OOWkV4A8GA?view