Yuxin Li

Yuxin Li

Ph.D. student

Publications

You can also find my articles on my Google Scholar profile.

Vague Prototype-Oriented Diffusion Model for Multi-Class Anomaly Detection

Published in ICML, 2024

We introduce a Vague Prototype-Oriented Diffusion Model (VPDM) designed for multi-class unsupervised anomaly detection. It introduces vague prototypes to avoid the “identical shortcut” problem by starting with minimal information and gradually adding details through a diffusion model. This approach improves the detection of anomalies across diverse datasets by minimizing the influence of abnormal information in the initial conditions.

Recommended citation: Li Y, Feng Y, Chen B, et al. Vague Prototype-Oriented Diffusion Model for Multi-Class Anomaly Detection[C]//Forty-first International Conference on Machine Learning. https://openreview.net/pdf?id=FvLd8Gr7xq

Transformer-Modulated Diffusion Models for Probabilistic Multivariate Time Series Forecasting

Published in ICLR, 2024

We introduce a Transformer-Modulated Diffusion Model (TMDM), uniting conditional diffusion generative process with transformers into a unified framework to enable precise distribution forecasting for MTS.

Recommended citation: Li, Y., Chen, W., Hu, X., Chen, B., & Zhou, M. (2023, October). Transformer-Modulated Diffusion Models for Probabilistic Multivariate Time Series Forecasting. In The Twelfth International Conference on Learning Representations. https://openreview.net/pdf?id=qae04YACHs

Prototype-oriented unsupervised anomaly detection for multivariate time series

Published in ICML, 2023

We propose a prototype-oriented UAD (PUAD) method under a probabilistic framework. Specifically, instead of learning the mappings for each MTS, the proposed PUAD views multiple MTSs as the distribution over a group of prototypes, which are extracted to represent a diverse set of normal patterns.

Recommended citation: Li, Y., Chen, W., Chen, B., Wang, D., Tian, L., & Zhou, M. (2023, July). Prototype-oriented unsupervised anomaly detection for multivariate time series. In International Conference on Machine Learning (pp. 19407-19424). PMLR. https://openreview.net/pdf?id=3vO4lS6PuF

A transfer learning method using speech data as the source domain for micro-Doppler classification tasks.

Published in KBS, 2020

We proposes a transfer learning method using speech data as the source domain for micro-Doppler classification tasks.

Recommended citation: Li, Y., He, K., Xu, D., & Luo, D. (2020). A transfer learning method using speech data as the source domain for micro-Doppler classification tasks. Knowledge-Based Systems, 209, 106449. https://www.sciencedirect.com/science/article/pii/S0950705120305785