Papers by Xiangci Li

10 papers
Contextualizing Generated Citation Texts (2024.lrec-main)

Copied to clipboard

Challenge: Abstractive citation text generation is usually framed as an infilling task . however, examining a recent LED-based citation generation system, we find that many of the generated citations are generic summaries of the reference paper’s main contribution, ignoring the citation context’s focus on a different topic.
Approach: They propose a modification to the citation text generation task by training the generation model to generate a citation given a reference paper and the context window around the target.
Outcome: The proposed model can generate citations based on the entire context window, including the target citation.
CodeScout: Contextual Problem Statement Enhancement for Software Agents (2026.findings-acl)

Copied to clipboard

Challenge: Current AI-powered code assistance tools struggle with ambiguous problem statements . failures on such ambiguously requests are highly correlated with longer trajectories .
Approach: They propose a contextual query refinement approach that transforms ambiguous user requests into comprehensive, actionable problem statements through lightweight pre-exploration of the target codebase.
Outcome: Empirical results show that CodeScout improves resolution rates with 27 additional issues resolved compared to baseline method.
A Knowledge Plug-and-Play Test Bed for Open-domain Dialogue Generation (2024.lrec-main)

Copied to clipboard

Challenge: Knowledge-based open-domain dialogue generation aims to build chit-chat systems that talk to humans using mined support knowledge.
Approach: They propose a benchmark for evaluating multi-source dialogue knowledge selection and response generation using Wikipedia's wizard of Wikipedia.
Outcome: The proposed benchmark is called multi-source Wizard of Wikipedia (Ms.WoW) it contains clean support knowledge, grounded at the utterance level and partitioned into multiple knowledge sources.
Context-aware Stand-alone Neural Spelling Correction (2020.findings-emnlp)

Copied to clipboard

Challenge: Existing natural language processing systems are vulnerable to noisy inputs resulting from misspellings.
Approach: They propose a stand-alone spelling correction problem that corrects the spelling of tokens without additional token insertion or deletion.
Outcome: The proposed solution outperforms the state-of-the-art spelling correction model by 12.8% absolute F0.5 score.
Scientific Discourse Tagging for Evidence Extraction (2021.eacl-main)

Copied to clipboard

Challenge: Primary experimental articles provide the crucial raw material for all subsequent scientific research, but the growing number of scientific literature makes it difficult for domain experts to efficiently utilize them.
Approach: They propose to automatically extract text fragments from primary research papers that describe the evidence presented in that paper's figures and to use them to build models of scientific argument.
Outcome: The proposed method is able to extract text fragments from primary research papers that describe the evidence presented in that paper's figures, and it is transferable to new datasets.
Wizard of Shopping: Target-Oriented E-commerce Dialogue Generation with Decision Tree Branching (2025.acl-long)

Copied to clipboard

Challenge: Prior human-annotated CPS datasets are small in size and lack integration with real-world product search systems.
Approach: They propose a method to generate target-oriented shopping conversations without human annotations by using large language models.
Outcome: The proposed method achieves highly natural and coherent conversations from three shopping domains and significantly improves on human evaluations and downstream tasks.
How Does Knowledge Selection Help Retrieval Augmented Generation? (2025.findings-emnlp)

Copied to clipboard

Challenge: Retrieval-augmented generation (RAG) is a powerful method for enhancing natural language generation by integrating external knowledge into a model’s output.
Approach: They empirically analyze how knowledge selection influences downstream generation performance in RAG systems by simulating different retrieval and selection conditions through a controlled mixture of gold and distractor knowledge.
Outcome: The proposed model is based on a controlled mixture of gold and distractor knowledge and simulated with a gold and distractors.
Related Work and Citation Text Generation: A Survey (2024.emnlp-main)

Copied to clipboard

Challenge: Academic research paper authors must perform literature review to compare work with prior work . authors must compose coherent story that connects prior work and current work based on author's understanding of field .
Approach: They propose to use automatic related work generation (RWG) to generate papers . authors summarize key approaches and define tasks in a zoo of historical works .
Outcome: a new study summarises key approaches and defines the tasks and discusses the challenges of RWG.
CORWA: A Citation-Oriented Related Work Annotation Dataset (2022.naacl-main)

Copied to clipboard

Challenge: Academic research is an exploratory activity to discover new solutions to problems . prior work focused on the sentence as the basic unit of generation, neglecting that related work sections consist of variable length text fragments derived from different information sources.
Approach: They propose a Citation Oriented Related Work Annotation dataset that labels citation text fragments . they propose linguistically-motivated framework for human-in-the-loop, abstractive related work generation .
Outcome: The proposed framework is based on a Citation Oriented Related Work Annotation dataset . it automatically tags unlabeled related work sections on the dataset based upon the proposed model .
Explaining Relationships Among Research Papers (2025.coling-main)

Copied to clipboard

Challenge: Existing literature reviews focus on summarizing individual papers without addressing the need for expository and transition sentences to explain the relationships among multiple papers.
Approach: They propose a feature-based, LLM-prompting approach to generate richer citation texts . they propose to use related work sections of scientific articles as proxy for the kind of short, customized, daily feed summaries .
Outcome: The proposed approach captures complex relationships among multiple papers while generating richer citation texts.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations