Weixi Feng (冯蔚熙)

I am a fifth-year CS Ph.D. candidate at UCSB NLP group. I am fortunate to be advised by Prof. William Wang and work closely with Prof. Xin Wang at UCSC. I obtained my M.Sc. in ECE from UCLA and my B.S. in ECE from the Hong Kong Polytechnic Unversity. I interned at Adobe Research and Amazon Ads before.

Email: weixifeng at ucsb dot edu

Email  /  CV  /  Google Scholar  /  Github  /  Twitter  /  LinkedIn

profile photo

I am actively looking for full-time scientists positions that focus on image or video generation and the application of LLMs/VLMs on visual generation.

Research

My research interests lie at the intersection of vision and language. Recently, I am specifically interested in compositionality problems in image/video generation and the application of generative models for design. I am trying to build LLM-centered visual generation systems across multiple domains (images, videos, 3D).

Publication
ddm_png VELMA: Verbalization Embodiment of LLM Agents for Vision and Language Navigation in Street View

Raphael Schumann,   Wanrong Zhu,   Weixi Feng,   Tsu-Jui Fu,   Stefan Riezler William Yang Wang,  

AAAI 2024
arxiv / paper / code

layoutgpt_gif LayoutGPT: Compositional Visual Planning and Generation with Large Language Models

Weixi Feng*,   Wanrong Zhu*,   Tsu-Jui Fu,   Varun Jampani,   Arjun Akula,   Xuehai He,   Sugato Basu,   Xin Eric Wang,   William Yang Wang
* equal contribution

NeurIPS 2023
arxiv / project page / code

edis_png EDIS: Entity-Driven Image Search over Multimodal Web Content

Siqi Liu*,   Weixi Feng*,   Tsu-Jui Fu,   Wenhu Chen,   William Yang Wang
* equal contribution

EMNLP 2023 Main
arxiv / code

structure_png Training-Free Structured Diffusion Guidance for Compositional Text-to-Image Synthesis

Weixi Feng,   Xuehai He,   Tsu-Jui Fu,   Varun Jampani,   Arjun Akula,   Pradyumna Narayana,   Sugato Basu,   Xin Eric Wang,   William Yang Wang

ICLR 2023
OpenReview / arxiv / project page / code

plan_png Neuro-Symbolic Procedural Planning with Commonsense Prompting

Yujie Lu,   Weixi Feng,   Wanrong Zhu,   Wenda Xu,   Xin Eric Wang,   Miguel Eckstein,   William Yang Wang

ICLR 2023 (Spotlight)
OpenReview / arxiv / code

uln_png ULN: Towards Underspecified vision-and-Language Navigation

Weixi Feng,   Tsu-Jui Fu,   Yujie Lu,   William Yang Wang

EMNLP 2022 Main
Abstract in 2nd Unimplicit Workshop, NAACL , 2022
Proceedings / arxiv / code

cpl_png CPL: Counterfactual Prompt Learning for Vision and Language Models

Xuehai He,   Diji Yang,   Weixi Feng,   Tsu-Jui Fu,   Arjun Akula,   Varun Jampani,   Pradyumna Narayana,   Sugato Basu,   William Yang Wang,   Xin Eric Wang

EMNLP 2022 Main
Proceedings / arxiv / code

Preprint
ddm_png T2V-Turbo: Breaking the Quality Bottleneck of Video Consistency Model with Mixed Reward Feedback

Jiachen Li,   Weixi Feng,   Tsu-Jui Fu,   Xinyi Wang,   Sugato Basu,   Wenhu Chen,   William Yang Wang

arxiv / Project page / code

ddm_png Reward Guided Latent Consistency Distillation

Jiachen Li,   Weixi Feng,   Wenhu Chen,   William Yang Wang

arxiv / Project page / code

ddm_png Discriminative Diffusion Models as Few-shot Vision and Language Learners

Xuehai He,   Weixi Feng,   Tsu-Jui Fu,   Varun Jampani,   Arjun Akula,   Pradyumna Narayana,   Sugato Basu,   William Yang Wang,   Xin Eric Wang

arxiv / code

davis_png Anticipating the Unseen Discrepancy for Vision and Language Navigation

Yujie Lu,   Huiliang ZHang,   Ping Nie,   Weixi Feng,   Wenda Xu,   Xin Eric Wang,   William Yang Wang

arxiv / code

Service

  • Reviewer: EACL 2023, ACL 2023, NeurIPS 2023, EMNLP 2023, ICLR 2024
  • Teaching

  • CS165B Machine Learning, 2020-2021, Spring 2022
  • ECE239 Deep Learning, Winter 2019

  • Template from Jon Barron.