📝 Publications

Assembling the Mind’s Mosaic: Towards EEG Semantic Intent Decoding
Jiahe Li, Junru Chen, Fanqi Shen, Jialan Yang, Jada Li, Zhizhang Yuan, Baowen Cheng, Meng Li, Yang Yang
BrainMosaic introduces Semantic Intent Decoding (SID), a novel framework that translates EEG/SEEG signals into natural language by modeling meaning as flexible, compositional semantic units. By aligning neural representations with a continuous semantic space and leveraging LLMs for reconstruction, it enables interpretable and open-vocabulary BCI communication.

Deep Learning-Powered Electrical Brain Signals Analysis: Advancing Neurological Diagnostics
Jiahe Li, Xin Chen, Fanqi Shen, Junru Chen, Yuxin Liu, Daoze Zhang, Zhizhang Yuan, Fang Zhao, Meng Li, Yang Yang
This review provides a comprehensive overview of recent deep learning advances in EEG and iEEG analysis for neurological disorder diagnosis, summarizing trends across 46 datasets and 7 conditions while proposing a standardized benchmark for future research.

Con4m: Context-aware Consistency Learning Framework for Segmented Time Series Classification
Junru Chen, Tianyu Cao, Jing Xu, Jiahe Li, Zhilong Chen, Tao Xiao, Yang Yang
Con4m is a consistency learning framework, which effectively utilizes contextual information more conducive to discriminating consecutive segments in segmented TSC tasks, while harmonizing inconsistent boundary labels for training.

Fast Updating Truncated SVD for Representation Learning with Sparse Matrices
Haoran Deng, Yang Yang, Jiahe Li, Cheng Chen, Weihao Jiang, and Shiliang Pu
code
The goal is to develop a more efficient method for dynamically updating truncated Singular Value Decomposition (SVD) of sparse and evolving matrices while maintaining high precision.

Accelerating Dynamic Network Embedding with Billions of Parameter Updates to Milliseconds
Haoran Deng, Yang Yang, Jiahe Li, Haoyang Cai, Shiliang Pu, and Weihao Jiang
The purpose of the DAMF algorithm is to achieve efficient and accurate dynamic network embedding by updating billion-edge graphs in under 10 milliseconds while capturing higher-order neighbor information.