Jincheng He

I am currently a PhD student from University of California Santa Cruz Silicon Valley Extension major in Computer Science and Engineering (CSE) following Prof. Ian Lane, focusing on multimodal conversational interaction, turn-taking models, LLM hallucination detection, speech processing, and natural language processing.

Previously I was a graduate student from Duke University major in Electrical and Computer Engineering (ECE), and an undergraduate student from Duke Kunshan University major in Data Science. I joined SMIIP Lab led by my research mentor Prof. Ming Li since the summer of my freshman year.

I have wide research interests in speech analysis, and especially focuses on Target Speech Separation (TSS) task. It is the task which only extracts target speaker’s voice from the mixed utterance.

I also have interests in mathematics, statistics, mathematical modeling.

Selected Publications

UCSC at SemEval-2025 Task 3: Context, Models and Prompt Optimization for Automated Hallucination Detection in LLM Output

Sicong Huang , Jincheng He, Shiyuan Huang , Karthik Raja Anandan , Arkajyoti Chakraborty , and Ian Lane

In Proceedings of the 19th International Workshop on Semantic Evaluation (SemEval-2025) , Jul 2025

Received Best System Description Award

Abs Bib PDF Website

Hallucinations pose a significant challenge for large language models when answering knowledge-intensive queries. As LLMs become more widely adopted, it is crucial not only to detect if hallucinations occur but also to pinpoint where they arise. SemEval 2025 Task 3, Mu-SHROOM: Multilingual Shared-task on Hallucinations and Related Observable Overgeneration Mistakes, is a recent effort in this direction. This paper describes our solution to the shared task. We propose a framework that first retrieves relevant context, next identifies false content from the answer, and finally maps them back to spans. The process is further enhanced by automatically optimizing prompts. Our system achieves the highest overall performance, ranking #1 in average position across all languages.
@inproceedings{huang-etal-2025-ucsc, title = {{UCSC} at {S}em{E}val-2025 Task 3: Context, Models and Prompt Optimization for Automated Hallucination Detection in {LLM} Output}, author = {Huang, Sicong and He, Jincheng and Huang, Shiyuan and Anandan, Karthik Raja and Chakraborty, Arkajyoti and Lane, Ian}, editor = {Rosenthal, Sara and Ros{\'a}, Aiala and Ghosh, Debanjan and Zampieri, Marcos}, booktitle = {Proceedings of the 19th International Workshop on Semantic Evaluation (SemEval-2025)}, month = jul, year = {2025}, address = {Vienna, Austria}, publisher = {Association for Computational Linguistics}, url = {https://aclanthology.org/2025.semeval-1.257/}, pages = {1981--1992}, isbn = {979-8-89176-273-2}, note = {Received Best System Description Award} }
Single-Channel Target Speaker Separation Using Joint Training with Target Speaker’s Pitch Information

Jincheng He, Yuanyuan Bao , Na Xu , Hongfeng Li , Shicong Li , Linzhang Wang , Fei Xiang , and Ming Li

In Proc. The Speaker and Language Recognition Workshop (Odyssey 2022) , Jul 2022

Abs Bib PDF Website

Despite the great progress achieved in the target speaker separation (TSS) task, we are still trying to find other robust ways for performance improvement which are independent of the model architecture and the training loss. Pitch extraction plays an important role in many applications such as speech enhancement and speech separation. It is also a challenging task when there are multiple speakers in the same utterance. In this paper, we explore if the target speaker pitch extraction is possible and how the extracted target pitch could help to improve the TSS performance. A target pitch extraction model is built and incorporated into different TSS models using two different strategies, namely concatenation and joint training. The experimental results on the LibriSpeech dataset show that both training strategies could bring significant improvements to the TSS task, even the precision of the target pitch extraction module is not high enough.
@inproceedings{he22_odyssey, author = {He, Jincheng and Bao, Yuanyuan and Xu, Na and Li, Hongfeng and Li, Shicong and Wang, Linzhang and Xiang, Fei and Li, Ming}, title = {{Single-Channel Target Speaker Separation Using Joint Training with Target Speaker's Pitch Information}}, year = {2022}, booktitle = {Proc. The Speaker and Language Recognition Workshop (Odyssey 2022)}, pages = {301--305}, doi = {10.21437/Odyssey.2022-42}, }