Lingxiao Li

Research: As a researcher in machine learning, Lingxiao is focused on improving the generalization ability of machine learning models with minimal supervision. His interests span multiple areas, including generative models, unsupervised (self-supervised) learning, representation learning, transfer learning and OOD generalization.

Bio: Lingxiao Li is a research assistant advised by Prof. Xiangyu Yue at MMLab, the Chinese University of Hong Kong. He received his bachelor’s degree with honours in computer science from the University of Liverpool advised by Prof. Xiaowei Huang and his master’s degree in computer science from Columbia University, working with Prof. Junfeng Yang and Prof. Asaf Cidon. He also works closely with Prof. Shuhui Wang at the Institute of Computing Technology, Chinese Academy of Sciences. He is seeking PhD opportunities starting in 2025 Fall.

In his free time, he is a big fan of Chopin and Mahler.

news

Jul 16, 2023	One paper accepted by ICCV 2023!
May 22, 2023	One paper has been submitted to arXiv.
Nov 22, 2022	One paper has been submitted to arXiv.
Jun 15, 2022	I graduated from @Columbia!

selected publications

arXiv

Monitoring and Adapting ML Models on Mobile Devices

Wei Hao, Zixi Wang, Lauren Hong, Lingxiao Li, Nader Karayanni, Chengzhi Mao, Junfeng Yang, and Asaf Cidon

arXiv preprint arXiv:2305.07772, 2023

Abs PDF

ML models are increasingly being pushed to mobile devices, for low-latency inference and offline operation. However, once the models are deployed, it is hard for ML operators to track their accuracy, which can degrade unpredictably (e.g., due to data drift). We design Nazar, the first end-to-end system for continuously monitoring and adapting models on mobile devices without requiring feedback from users. Our key observation is that often model degradation is due to a specific root cause, which may affect a large group of devices. Therefore, once Nazar detects a consistent degradation across a large number of devices, it employs a root cause analysis to determine the origin of the problem and applies a cause-specific adaptation. We evaluate Nazar on two computer vision datasets, and show it consistently boosts accuracy compared to existing approaches. On a dataset containing photos collected from driving cars, Nazar improves the accuracy on average by 15%.
ICCV

The Euclidean Space is Evil: Hyperbolic Attribute Editing for Few-shot Image Generation

Lingxiao Li, Yi Zhang, and Shuhui Wang

In International Conference on Computer Vision (ICCV), 2023

Abs PDF

Few-shot image generation is a challenging task since it aims to generate diverse new images for an unseen category with only a few images. Existing methods suffer from the trade-off between the quality and diversity of generated images. To tackle this problem, we propose Hyperbolic Attribute Editing(HAE), a simple yet effective method. Unlike prior arts that work in Euclidean space, HAE captures the hierarchy among images using data from seen categories in hyperbolic space. Given a well-trained HAE, images of unseen categories can be generated by moving the latent code of a given image toward any meaningful directions in the Poincaré disk with a fixing radius. Most importantly, the hyperbolic space allows us to control the semantic diversity of the generated images by setting different radii in the disk. Extensive experiments and visualizations demonstrate that HAE is capable of not only generating images with promising quality and diversity using limited data but achieving a highly controllable and interpretable editing process.