Kun Zhao

Kun Zhao

PhD Candidate | LLM & NLP | Medical AI

Posts by Collection

portfolio

publications

Evaluating Open-Domain Dialogues in Latent Space with Next Sentence Prediction and Mutual Information

Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)

CMN evaluates open-domain dialogue responses in latent space using next-sentence prediction and mutual information.

Recommended citation: Kun Zhao, Bohao Yang, Chenghua Lin, Wenge Rong, Aline Villavicencio, Xiaohui Cui, "Evaluating Open-Domain Dialogues in Latent Space with Next Sentence Prediction and Mutual Information." Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), 2023.
Download Paper

Effective distillation of table-based reasoning ability from llms

Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)

Use Google Scholar for full citation

Recommended citation: Bohao Yang, Chen Tang, Kun Zhao, Chenghao Xiao, Chenghua Lin, "Effective distillation of table-based reasoning ability from llms." Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024), 2024.

SLIDE: A Framework Integrating Small and Large Language Models for Open-Domain Dialogues Evaluation

Findings of the Association for Computational Linguistics ACL 2024

SLIDE combines a specialized small language model with an LLM evaluator for open-domain dialogue evaluation.

Recommended citation: Kun Zhao, Bohao Yang, Chen Tang, Chenghua Lin, Liang Zhan, "SLIDE: A Framework Integrating Small and Large Language Models for Open-Domain Dialogues Evaluation." Findings of the Association for Computational Linguistics ACL 2024, 2024.
Download Paper

Interpretable Spatio-Temporal Embedding for Brain Structural-Effective Network with Ordinary Differential Equation

MICCAI 2024

STE-ODE models the interplay between structural and effective brain networks using directed graph embeddings and ordinary differential equations. Access the paper.

Recommended citation: Haoteng Tang, Guodong Liu, Siyuan Dai, Kai Ye, Kun Zhao, Wenlu Wang, Carl Yang, Lifang He, Alex Leow, Paul Thompson, "Interpretable Spatio-Temporal Embedding for Brain Structural-Effective Network with Ordinary Differential Equation." MICCAI 2024.
Download Paper

Who Matters More in Radiology Report Generation: Vision Encoders or Language Models?

SAIMBio Workshop at IEEE ICDM Workshops 2025

This study uses controlled component swaps to isolate the individual and interaction effects of vision encoders and language backbones. Access the paper.

Recommended citation: Kun Zhao, Yang Du, Rhianna Zhang, Liang Zhan, Dongkuan Xu, Pengfei Gu, Haoteng Tang, "Who Matters More in Radiology Report Generation: Vision Encoders or Language Models?" IEEE ICDM Workshops, 2025.
Download Paper

Overview of the BioLaySumm 2025 Shared Task on Lay Summarization of Biomedical Research Articles and Radiology Reports

Proceedings of the 24th Workshop on Biomedical Language Processing

Access paper here

Recommended citation: Chenghao Xiao, Kun Zhao, Xiao Wang, Siwei Wu, Sixing Yan, Tomas Goldsack, Sophia Ananiadou, Noura Al Moubayed, Liang Zhan, William K. Cheung, Chenghua Lin, "Overview of the BioLaySumm 2025 Shared Task on Lay Summarization of Biomedical Research Articles and Radiology Reports." Proceedings of the 24th Workshop on Biomedical Language Processing, pp. 365–377, 2025.
Download Paper

Emphasising Structured Information: Integrating Abstract Meaning Representation into LLMs for Enhanced Open-Domain Dialogue Evaluation

Findings of the Association for Computational Linguistics: EMNLP 2025

Access paper here

Recommended citation: Bohao Yang, Kun Zhao, Dong Liu, Chen Tang, Liang Zhan, Chenghua Lin, "Emphasising Structured Information: Integrating Abstract Meaning Representation into LLMs for Enhanced Open-Domain Dialogue Evaluation." Findings of the Association for Computational Linguistics: EMNLP 2025, 2025.
Download Paper

Pretraining a Generative AI Model on CBT Sessions to Emulate Human Therapist Strategies: A Study Using the APA PsycTherapy Database

SAIMBio Workshop at IEEE ICDM Workshops 2025, pp. 2011–2014

Publication details are indexed in Google Scholar.

Recommended citation: Fahad Islam, Kun Zhao, Bart P. Knijnenburg, Ming-Hui Chen, Alex Leow, Prasanna Muthukanagaraj, Liang Zhan, Chao Shi, "Pretraining a Generative AI Model on CBT Sessions to Emulate Human Therapist Strategies: A Study Using the APA PsycTherapy Database." IEEE ICDM Workshops, 2025, pp. 2011–2014.

Interpretable Multimodal Learning for Integrating Neuroimaging and Genetic Data in Alzheimer’s Disease

Frontiers in Radiology, 2026 (2025 JIF: 4.0)

R-GenIMA integrates ROI-wise 3D MRI representations with prompted SNP tokens for interpretable four-way NC/SMC/MCI/AD classification. Access the paper.

Recommended citation: Kun Zhao, Siyuan Dai, Yingying Zhang, Guodong Liu, Pengfei Gu, Chenghua Lin, Paul M. Thompson, Alex Leow, Heng Huang, Lifang He, Liang Zhan, Haoteng Tang, "Interpretable Multimodal Learning for Integrating Neuroimaging and Genetic Data in Alzheimer's Disease." Frontiers in Radiology, 2026.
Download Paper

HERO: Hierarchical Evidential Reasoning Optimization for Radiology Report Generation via Reason-then-Summarize

Preprint · Under review at AAAI 2027

HERO develops factorized GRPO credit assignment for evidence-grounded radiology generation. Access the paper.

Recommended citation: Kun Zhao, Guodong Liu, Hui Ji, Siyuan Dai, Pan Wang, Jifeng Song, Chenghua Lin, Liang Zhan, Haoteng Tang, "HERO: Hierarchical Evidential Reasoning Optimization for Radiology Report Generation via Reason-then-Summarize." arXiv:2601.03321, 2026.
Download Paper

X-ray Made Simple: Lay Radiology Report Generation and Robust Evaluation

Findings of the Association for Computational Linguistics: ACL 2026

This work introduces refined lay radiology datasets, semantics-based evaluation, and lay-guided training. Access the paper.

Recommended citation: Kun Zhao, Chenghao Xiao, Sixing Yan, Haoteng Tang, William K. Cheung, Noura Al Moubayed, Liang Zhan, Chenghua Lin, "X-ray Made Simple: Lay Radiology Report Generation and Robust Evaluation." Findings of the Association for Computational Linguistics: ACL 2026, 2026.
Download Paper

DRE: An Effective Dual-Refined Method for Integrating Small and Large Language Models in Open-Domain Dialogue Evaluation

Neural Networks, 2026, 109574 (2025 JIF: 6.3)

DRE combines SLM-guided interior prompt refinement with exterior score refinement for open-domain dialogue evaluation. Access the paper.

Recommended citation: Kun Zhao, Bohao Yang, Chen Tang, Siyuan Dai, Haoteng Tang, Chenghua Lin, Liang Zhan, "DRE: An Effective Dual-Refined Method for Integrating Small and Large Language Models in Open-Domain Dialogue Evaluation." Neural Networks, 2026, 109574.
Download Paper

talks

teaching

Teaching experience 1

Undergraduate course, University 1, Department, 2014

This is a description of a teaching experience. You can use markdown like any other post.

Teaching experience 2

Workshop, University 1, Department, 2015

This is a description of a teaching experience. You can use markdown like any other post.