Papers by Jiahua Liu

7 papers
Semantically Comprehensive Token Pruning in LVLMs via Maximizing Concept Coverage (2026.acl-long)

Copied to clipboard

Challenge: Existing visual token pruning methods leverage simple metrics derived from human experience, such as attention or similarity, to rank and select tokens within a highly entangled feature space.
Approach: They propose a novel visual token pruning method that uses a concept-driven paradigm to quantify the Marginal Semantic Gain of each token's contribution to uncovered concepts.
Outcome: The proposed method outperforms state-of-the-art methods in a concept-driven model while maintaining semantic completeness.
XQA: A Cross-lingual Open-domain Question Answering Dataset (P19-1)

Copied to clipboard

Challenge: Open-domain question answering aims to answer questions through text retrieval and reading comprehension . but, the success of these models relies on a massive volume of training data, which is not available in other languages . a new dataset aims at investigating cross-lingual OpenQA .
Approach: They propose to use a dataset for cross-lingual OpenQA research to test models . they use XQA dataset to train models with large volumes of labeled data .
Outcome: The proposed model achieves best results in almost all target languages while the performance is lower than that of English.
Semantic Novelty Detection in Natural Language Descriptions (2021.emnlp-main)

Copied to clipboard

Challenge: Existing novelty detection algorithms are coarse-grained, working at the document or topic level.
Approach: They propose to use a fine-grained semantic novelty detection problem to solve a novel novel scene problem.
Outcome: The proposed model outperforms baseline models on the proposed task by large margins.
A Knowledge-Driven Approach to Classifying Object and Attribute Coreferences in Opinion Mining (2020.findings-emnlp)

Copied to clipboard

Challenge: Existing methods to classify and resolve coreferences in opinionated reviews require domain-specific knowledge.
Approach: They propose to automatically mine domain-specific knowledge for opinionated reviews by combining it with commonsense knowledge.
Outcome: The proposed approach extracts domain-specific knowledge from unlabeled review data and trains a knowledgeaware neural coreference classification model to leverage commonsense knowledge for the task.
Parameter-Efficient Prompt Tuning Makes Generalized and Calibrated Neural Text Retrievers (2023.findings-emnlp)

Copied to clipboard

Challenge: Prompt tuning is a technique that updates few parameters in pre-trained models for language understanding and generation tasks.
Approach: They propose to leverage prompt tuning for neural text retrieval to improve generalization and cross-domain generalization.
Outcome: The proposed approach can mitigate the two issues faced by fine-tuning retrieval methods and improve the out-of-domain zero-shot generalization of the retrieval models.
A Multi-answer Multi-task Framework for Real-world Machine Reading Comprehension (D18-1)

Copied to clipboard

Challenge: Existing models of machine reading comprehension (MRC) are based on cloze style questions or crowdworkers given a short passage from well-edited sources.
Approach: They propose a multi-answer multi-task framework that uses multiple reference answers for multiple questions.
Outcome: The proposed model increases the ROUGE-L score on the DuReader dataset from 44.18, the previous state-of-the-art, to 51.09 .
Route to Rome Attack: Directing LLM Routers to Expensive Models via Adversarial Suffix Optimization (2026.acl-long)

Copied to clipboard

Challenge: Existing routing attacks depend on white-box access or heuristic prompts, rendering them ineffective in real-world black-box scenarios.
Approach: They propose a cost-aware routing strategy that routes queries to the least-cost model . they propose heuristic prompts that are ineffective in real-world black-box scenarios .
Outcome: The proposed approach significantly increases the routing rate to expensive models on queries of different distributions.

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