Papers by Siqiang Luo
StructAM: Enhancing Address Matching through Semantic Understanding of Structure-aware Information (2024.lrec-main)
Copied to clipboard
| Challenge: | Existing approaches to address matching rely on string-based similarity matching or manually-designed rules. |
| Approach: | They propose a method to match unstructured addresses to standard ones in a database using pre-trained language models and graph neural networks. |
| Outcome: | The proposed method outperforms state-of-the-art methods on real-world addresses . it incorporates spatial coordinates and contextual information from the surrounding area as auxiliary guidance. |
LLM as Prompter: Low-resource Inductive Reasoning on Arbitrary Knowledge Graphs (2024.findings-acl)
Copied to clipboard
| Challenge: | Knowledge Graph (KG) inductive reasoning is widely adopted in various applications. |
| Approach: | They propose a framework for low-resource inductive reasoning using Large Language Models to generate a graph-structural prompt for pre-trained KGs. |
| Outcome: | The proposed framework outperforms previous methods in three-shot, one-shot and zero-shot reasoning tasks. |
TopoRAG: Graph-based RAG via Topology-aware Approximate Nearest Neighbor Search (2026.findings-acl)
Copied to clipboard
| Challenge: | Recent studies extend RAG with graph-structured knowledge, enhancing retrieval to capture relational context beyond isolated text chunks. |
| Approach: | They propose a retrieval framework that integrates structural constraints into ANN search . they propose heuristic neighbor expansion which augments the retrieved set by traversing immediate neighbors . |
| Outcome: | The proposed framework improves precision and reduces context redundancy compared to existing methods. |
Permitted Knowledge Boundary: Evaluating the Knowledge-Constrained Responsiveness of Large Language Models (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Recent research has raised concerns about the controllability of large language models. |
| Approach: | They propose to define a "boundary bias" to depict KCR in large language models . they propose to quantify the boundary bias of LLMs and assess the KCR . |
| Outcome: | The proposed model is based on two new datasets to assess its performance. |