Ricky T. Q. Chen (@RickyTQChen) / X

Ricky T. Q. Chen

311 posts

Ricky T. Q. Chen

@RickyTQChen

Research Scientist. Meta. I build simplified abstractions of the world through the lens of dynamics and flows.

Joined May 2013

Ricky T. Q. Chen
@RickyTQChen
Dec 10, 2024
Want to learn continuous & discrete Flow Matching? We've just released: 📙 A guide covering Flow Matching basics & advanced methods arxiv.org/abs/2412.06264. 💻 An open source codebase with image & text examples github.com/facebookresear…. 🗣️ A Flow Matching tutorial #NeurIPS2024.
112K
Ricky T. Q. Chen
@RickyTQChen
Dec 10, 2024
Flow Matching tutorial slides: drive.google.com/file/d/1-QKAT8…
Ricky T. Q. Chen
@RickyTQChen
Dec 10, 2024
Want to learn continuous & discrete Flow Matching? We've just released: 📙 A guide covering Flow Matching basics & advanced methods arxiv.org/abs/2412.06264. 💻 An open source codebase with image & text examples github.com/facebookresear…. 🗣️ A Flow Matching tutorial #NeurIPS2024.
129K
Ricky T. Q. Chen
@RickyTQChen
Jun 11, 2025
Padding in our non-AR sequence models? Yuck. 🙅 👉 Instead of unmasking, our new work *Edit Flows* perform iterative refinements via position-relative inserts and deletes, operations naturally suited for variable-length sequence generation. Easily better than using mask tokens.
GIF
42K
Ricky T. Q. Chen
@RickyTQChen
Dec 11, 2020
We build a probabilistic gradient dynamics model, and explore inference as a sub-routine *for* stochastic optimization! Check out our contributed talk at …cant-believe-its-not-better.github.io (Dec 12) Paper: arxiv.org/abs/2011.04803 w/ @damichoi95 @lukas_balles @DavidDuvenaud @PhilippHennig5
GIF
Ricky T. Q. Chen
@RickyTQChen
Apr 25, 2025
This ICLR is the best conference ever. Attendees are extremely friendly and cuddly. ..What do you mean this is the wrong hall?
28K
Ricky T. Q. Chen
@RickyTQChen
Oct 5, 2022
Introducing Neural Conservation Laws, a pair of density and vector fields that solve the continuity equation simply by construction. This is a key PDE in many research areas: fluid-sim, OT, diffusion, normalizing flows, etc. arxiv.org/abs/2210.01741 w/ @jrichterpowell @lipmanya
Ricky T. Q. Chen
@RickyTQChen
Feb 15, 2023
Excited to share our new work on Riemannian Flow Matching. Unlike diffusion-based approaches, it’s - completely simulation-free on simple manifolds, - trivially applies to higher dimensions, - tractably generalizes to general geometries! arxiv.org/abs/2302.03660 w/ @lipmanya
GIF
86K
Ricky T. Q. Chen
@RickyTQChen
Sep 17, 2024
New paper! We cast reward fine-tuning as stochastic control. 1. We prove that a specific noise schedule *must* be used for fine-tuning. 2. We propose a novel algorithm that is significantly better than the adjoint method*. (*this is an insane claim) arxiv.org/abs/2409.08861
33K
Ricky T. Q. Chen
@RickyTQChen
Oct 10, 2024
I and Yaron @lipmanya are hiring PhD research interns for 2025 in New York City, to work on developing core foundational methods for generative modeling at scale. If you're familiar with some of our works, shoot us an email: {rtqichen,ylipman}@Meta.com metacareers.com/jobs/532549086…
47K
Ricky T. Q. Chen
@RickyTQChen
Apr 27, 2025
Against conventional wisdom, I will be giving a talk with particular focus on the "how" and the various intricacies of applying stochastic control for generative modeling. Mon 9:50am Hall 1 Apex #ICLR2025 Also check out the other talks at delta-workshop.github.io!
20K
Ricky T. Q. Chen
@RickyTQChen
Mar 23, 2021
We differentiate through event handling, and extend Neural ODEs to implicitly defined termination times to allow for discontinuous trajectories! A virtual poster for our ICLR2021 paper on Learning Neural Event Functions. arXiv: arxiv.org/abs/2011.03902 w/ @brandondamos @mnick
Ricky T. Q. Chen
@RickyTQChen
Apr 25, 2025
We are presenting 3 orals and 1 spotlight at #ICLR2025 on two primary topics: On generalizing the data-driven flow matching algorithm to jump processes, arbitrary discrete corruption processes, and beyond. And on highly scalable algorithms for reward-driven learning settings.
GIF
19K
Ricky T. Q. Chen
@RickyTQChen
Aug 19, 2019
We're releasing code *and pretrained models* for Residual Flows, a SOTA invertible generative model, at github.com/rtqichen/resid…. Compared to existing flow models that enforce structured Jacobians, we can use simple ResNets and efficient estimators to get unbiased log-densities.
Ricky T. Q. Chen
@RickyTQChen
May 31, 2025
FUDOKI: A Multimodal Model Purely Based on Discrete Flow Matching Really nice work. Uses embedding distances to define corruption process, and a single unified bidirectional Transformer + Discrete Flow model for both image and text generation. No special mask tokens involved!
00:00
00:00
17K