Victor Wang (original) (raw)
Research
My current interests are in natural language understanding and generation. I'm interested in studying the training objectives, model design, and inference procedures that allow for models to be controllable and versatile. I believe in building models that act as a dependable, human-interfaceable model of people and the world.
| Improving LLM-as-a-Judge Inference with the Judgment Distribution Victor Wang, Michael J.Q. Zhang, Eunsol Choi arXiv, 2025 |
|---|
| Designing LLM-Based Support for Homelessness Caseworkers Whitney Nelson,Victor Wang, Eunsol Choi, Min Kyung Lee AAAI PubLLM, 2024 |
| Syllable-PBWT for space-efficient haplotype long-match query Victor Wang, Ardalan Naseri, Shaojie Zhang, Degui Zhi Bioinformatics, 2023 |
| P-smoother: efficient PBWT smoothing of large haplotype panels William Yue, Ardalan Naseri,Victor Wang, Pramesh Shakya, Shaojie Zhang, Degui Zhi Bioinformatics Advances, 2022 |
| MorphSet: Improving Renal Histopathology Case Assessment Through Learned Prognostic Vectors Pietro Antonio Cicalese, Syed Asad Rizvi,Victor Wang, Sai Patibandla, Pengyu Yuan, Samira Zare, Katharina Moos, Ibrahim Batal, Marian Clahsen-van Groningen, Candice Roufosse, Jan Becker, Chandra Mohan, Hien Van Nguyen MICCAI, 2021 |
| Polaratio: A magnitude-contingent monotonic correlation metric and its improvements to scRNA-seq clustering Victor Wang, Pietro Antonio Cicalese, Anto Sam Crosslee Louis Sam Titus, Chandra Mohan bioRxiv, 2020 |
Industry
Schlumberger
Menlo Park, CA, Summer 2022
- Created an application to allow a robot to locate a sound source and approach it
- Worked with sensors such as a lidar, microphone array, and event-based camera
Projects
| [ Anticipating Future Tokens to Help Transformer LLMs Plan ](data/Anticipating Future Tokens to Help Transformer LLMs Plan.pdf) CS 395T Topics in NLP, Fall 2023 Project proposal: We train a transformer LM to predict and condition on a quasi-bag-of-words of future tokens, forming a planning signal separate from the model's textual output. |
|---|
| [ Multimodal Contrastive Decoding for Image Captioning ](data/Multimodal Contrastive Decoding for Image Captioning.pdf) CS 395T Grounded NLP, Spring 2024 We apply multimodal contrastive decoding to multilingual image captioning, with a multimodal expert model and single-modal amateur models. |
| [ Theory of Mind Objectives for Hanabi ](data/Theory of Mind Objectives for Hanabi.pdf) CS 394R Reinforcement Learning, Spring 2024 In the cooperative card game of Hanabi, we train an agent to predict its own cards, and use these predictions to bootstrap training for higher order beliefs. |
| [ Paragraph Segmentation ](data/Paragraph Segmentation.pdf) CS 378 Natural Language Processing, Spring 2023 We add coreference-based attention to a transformer for the task of paragraph segmentation, which has applications in writing assistance. |
| [ Robot for Multistory Traversal ](data/Robot for Multistory Traversal.pdf) CS 378 Autonomous Driving, Fall 2023 We implement a robot that interacts with human passerby to travel on an elevator. |
Personal
I enjoy learning languages. I have studied Spanish and Japanese and am now learning Mandarin. In December 2023, I passed the JLPT N1 with a score of 172/180. Check out my post on recommendations from my journey!
Last updated: 3/5/2025