Wei ChenMachine Learning Researcher

Machine Learning | Representation | Generation

Wei Chen

I am a PhD candidate in Electrical and Computer Engineering at Purdue University, advised by Prof. Qiang Qiu. I work on parameter-efficient adaptation, generative models, and neural representation learning.

EmailCurriculum vitaeLinkedIn
Portrait of Wei Chen

Selected publications

Complete list
Generated scene from the sparse fine-tuning project

Sparse Fine-Tuning of Transformers for Generative Tasks

Wei Chen, Jingxi Yu, Zichen Miao, and Qiang Qiu

International Conference on Computer Vision (ICCV) | Conference

Sparse adaptation for generative transformers with controllable interpolation, editing, intensity adjustment, and style merging.

PaperProject
Diagram of the filter subspace similarity method

Inner Product-based Neural Network Similarity

Wei Chen, Zichen Miao, and Qiang Qiu

Advances in Neural Information Processing Systems (NeurIPS) | Conference

A filter-subspace representation that efficiently compares large collections of trained neural networks.

PaperProject

Recent news

  • "Sparse Fine-Tuning of Transformers for Generative Tasks" accepted to ICCV.
  • "Coeff-Tuning" accepted to CVPR.
  • "Large Convolutional Model Tuning via Filter Subspace" accepted to ICLR.
  • Joined Amazon as an Applied Scientist intern.