Ruqi Zhang

Gradient-based Discrete Sampling with Automatic Cyclical Scheduling Discrete distributions, particularly in high-dimensional deep models, are often highly multimodal due to inherent discontinuities. While gradient-based discrete sampling has proven effective, it is su...

Paper: arxiv.org/abs/2402.17699
Code: github.com/patrickpynad...

13.12.2024 18:22 — 👍 1    🔁 0    💬 0    📌 0

This project was led by the amazing PhD student Patrick Pynadath. Hope to see you at our poster on Friday, 11am-2pm, East Hall #4101!

13.12.2024 09:01 — 👍 0    🔁 0    💬 1    📌 0

Cyclical MCMC is a popular method for sampling from complex multimodal distributions. In our #NeurIPS2024 paper, we extend this approach to discrete distributions with novel designs tailored for discrete variables. Applications include graphic models, energy-based models, and language models.

13.12.2024 09:01 — 👍 0    🔁 0    💬 1    📌 0

Excited to be at #NeurIPS2024 all week! If you'd like to chat about:

- Probabilistic & Bayesian ML/DL
- Sampling & generation in discrete data
- LLM & VLM alignment/safety
- ML PhD opportunities at Purdue

…or anything else, feel free to DM or email me!

12.12.2024 08:32 — 👍 2    🔁 0    💬 0    📌 0