Liangliang Zhang

Computer Science Ph.D. Candidate

Large language models, knowledge graphs, and graph machine learning

Ph.D. Candidate in Computer Science · Reliable RAG

Liangliang Zhang 张靓靓

My Ph.D. research asks a practical question: when a large language model uses external evidence, how can we make its answer reliable? I study retrieval, planning, knowledge graphs, and evaluation to build RAG systems that are grounded, efficient, and trustworthy.

Professional portrait of Liangliang Zhang

Curious by nature. Serious about reliable AI.

Current role
Ph.D. Candidate, RPI
Research areas
Reliable RAG, knowledge graphs, graph ML

I work to make RAG systems reliable enough to trust.

I target the gap between retrieving information and using it responsibly: helping LLMs plan retrieval, connect evidence across knowledge sources, and demonstrate that their answers are grounded in sufficient, relevant evidence.

Reliable RAG and supporting areas

Reliable RAG

Retrieval planning, grounded reasoning, and evidence-aware generation for more dependable LLM behavior.

Knowledge Graphs

Knowledge grounding, structured retrieval, and benchmark design for robust AI systems.

Graph Machine Learning

Graph neural networks, graph condensation, and scalable representation learning.

Evaluation and Data

Dataset diagnosis, benchmarking, and evaluation methods that make RAG claims more reliable and actionable.

Industry research experience

Returning Summer Research Intern · Summer 2026

Yorktown Heights. With the same IBM Research collaboration team, developed AGENTICTT for adaptive table-text QA.

Output: Adaptive Agentic Retrieval and Reasoning for Table-Text QA, submitted to AAAI 2027; arXiv preprint forthcoming.

Ph.D. Candidate · Ongoing

Rensselaer Polytechnic Institute. Reliable RAG, knowledge graphs, graph learning, and trustworthy evaluation in the DAMI Lab.

Selected publications

Selected first-author work on reliable RAG, data discovery, and graph learning. See the full publication list for all papers and academic service.

arXiv 2025

From Factoid Questions to Data Product Requests: Benchmarking Data Product Discovery over Tables and Text

Liangliang Zhang, Nandana Mihindukulasooriya, Niharika S. D'Souza, Sola Shirai, Sarthak Dash, Yao Ma, Horst Samulowitz

Preprint

NeurIPS 2025

Diagnosing and Addressing Pitfalls in KG-RAG Datasets: Toward More Reliable Benchmarking

Liangliang Zhang, Zhuorui Jiang, Hongliang Chi, Haoyang Chen, Mohammed Elkoumy, Fali Wang, Qiong Wu, Zhengyi Zhou, Shirui Pan, Suhang Wang, Yao Ma

Paper

TMLR 2025

Extending Graph Condensation to Multi-Label Datasets: A Benchmark Study

Liangliang Zhang, Haoran Bao, Yao Ma

Paper

Let's connect

Email me at zhangl41@rpi.edu for research collaborations, research scientist, applied scientist, and machine learning opportunities.

Learn more on the About, Publications, and DAMI Lab pages.