@maxence_noble

Postdoctoral Researcher @ CFM ML Lab | ML PhD @Polytechnique

 Paris, France
maxencenoble.github.io/
Joined June 2024
  • Tweets71
  • Following160
  • Followers275
14Media
In case you're wondering what my thesis is titled :)
Maxence Noble@maxence_nobleAug 17
🎓 Update: I defended my PhD ! 2 months ago actually, with some vacation in between 😅 Huge thanks to @AlainDurmus for this amazing journey 🙏 🚀 Next: from September, I’ll join CFM ML Lab as a Postdoc with Eric Vanden-Eijnden, still working on generative models & sampling!
4
130
16,894
Maxence Noble retweeted
Sitan Chen@sitanchOct 9
🚨 Thrilled to be co-organizing a Simons Institute semester-long program on diffusions and flows with @joanbruna, Sinho Chewi, @KonstDaskalakis, @ArnaudDoucet1 in Fall '27! The call for Simons postdoctoral research fellowships for the program is out now, more info below ⬇️ 1/3
3
32
2
179
18,879
Maxence Noble retweeted
Dario Shariatian
@dario_shaOct 1
🔥 Accelerate your diffusion LMs with Distribution Matching Distillation Introducing two methods for continuous diffusion LMs with gains at complementary sampling budgets: 🔺 Simplex-DMD at very few-step budgets 🎲 Reinforce-DMD at larger budgets 🔺 At 4 NFEs Simplex-DMD halves the strongest baseline’s generative perplexity at matched entropy (1/n)
2
25
4
101
6,909
Maxence Noble retweeted
Jiequn Han
@JiequnHSep 29
📣 Our workshop on Emerging Directions in Probabilistic Modeling is next week, Oct 5–7 in NYC! Full schedule: edpmworkshop.github.io/#sche… Can't join in person? Talks will be livestreamed on Zoom. Register here: simonsfoundation.zoom.us/web…
Jiequn Han
@JiequnHJul 27
📣 We are excited to announce the workshop Emerging Directions in Probabilistic Modeling: Methods and Scientific Applications 📅 Oct. 5–7, 2026 📍 Flatiron Institute & IBM Research, NYC Applications are welcome by Aug. 22: edpmworkshop.github.io
19
4
51
7,863
TSBM is accepted at NeurIPS 26 !! Looking forward to presenting it and meeting great people at Paris 🇫🇷🇫🇷
Maxence Noble@maxence_nobleJul 21
New preprint alert (my final PhD paper ♥️) Twisted Schrödinger Bridge Matching (TSBM) We “twist” classic diffusion losses to tackle trajectory inference under physical constraints, encoded via a differentiable state cost Paper: arxiv.org/abs/2607.16987 with Github link (1/5)
4
48
4,018
Maxence Noble retweeted
Alex Tong@AlexanderTong7Sep 1
We're recruiting PhD students at #AITHYRA in Vienna! Working on ML methods at the intersection of AI and biology. Details & apply: phd.aithyra.at Feel free to reach out if you have questions!
Michael Bronstein@mmbronsteinSep 1
phd.aithyra.at
22
5
90
14,211
🎓 Update: I defended my PhD ! 2 months ago actually, with some vacation in between 😅 Huge thanks to @AlainDurmus for this amazing journey 🙏 🚀 Next: from September, I’ll join CFM ML Lab as a Postdoc with Eric Vanden-Eijnden, still working on generative models & sampling!
8
5
1
65
23,940
Maxence Noble retweeted
Stat.ML Papers@StatMLPapersJul 23
Diffusion-based Annealed Boltzmann Generators : benefits, pitfalls and hopes ift.tt/UAxsPm3

Diffusion-based Annealed Boltzmann Generators : benefits, pitfalls...

Sampling configurations at thermodynamic equilibrium is a central challenge in statistical physics. Boltzmann Generators (BGs) tackle it by combining a generative model with a Monte Carlo (MC)...

arxiv.org
3
17
1,125
New preprint alert (my final PhD paper ♥️) Twisted Schrödinger Bridge Matching (TSBM) We “twist” classic diffusion losses to tackle trajectory inference under physical constraints, encoded via a differentiable state cost Paper: arxiv.org/abs/2607.16987 with Github link (1/5)
3
19
3
147
12,725
more replies
We run GSBM and TSBM on crowd navigation and single-cell, across increasing dimensions, and observe a clear trade-off: - TSBM better integrates the physical constraint - GSBM better matches the marginal constraints 🫣Looks like there’s a sweet spot waiting to be found... (5/5)
1
1
2
205
Tagging people who might be interested : @ValentinDeBort1 @ArnaudDoucet1 @guanhorng_liu @MoleiTaoMath @cdomingoenrich @YuanqiD @k_neklyudov @AlexanderTong7 @mmbronstein @iNikitaGushchin @PPotaptchik @_sophia_tang_ @DenBless94 @julberner @lorenz_richter @JiequnH @theh2o64
51
Can we improve energy-based diffusion models by solving a classification task? 🤔 🚀 Excited to share DiffCLF, accepted at ICML 2026 🇰🇷! ! We predict a sample’s noise level, and recover densities as a by-product. Joint work with @theh2o64 (co-lead) and @jmhernandez233. [🧵1/n]
1
31
1
156
9,367
New preprint ! We extend SLIPS ideas (arxiv.org/abs/2402.10758) to the Riemannian case, to sample efficiently from multi-modal distributions (high-dimensional and heavy-tail). More xps coming soon !

Stochastic Localization via Iterative Posterior Sampling

Building upon score-based learning, new interest in stochastic localization techniques has recently emerged. In these models, one seeks to noise a sample from the data distribution through a...

arxiv.org
Stat.CO Papers@StatCOupdatesFeb 3
Alain Durmus, Maxence Noble, Thibaut Pellerin. [statML]. Sampling from multi-modal distributions on Riemannian manifolds with training-free stochastic interpolants. arxiv.org/abs/2602.00641

Sampling from multi-modal distributions on Riemannian manifolds...

In this paper, we propose a general methodology for sampling from un-normalized densities defined on Riemannian manifolds, with a particular focus on multi-modal targets that remain challenging...

arxiv.org
2
8
569
Sampling from multi-modal distributions on Riemannian manifolds with training-free stochastic interpolants Alain Durmus, Maxence Noble, Thibaut Pellerin arxiv.org/abs/2602.00641 [𝚜𝚝𝚊𝚝.𝙼𝙻 𝚌𝚜.𝙻𝙶 𝚜𝚝𝚊𝚝.𝙲𝙾]
1
12
537
A Diffusive Classification Loss for Learning Energy-based Generative Models Louis Grenioux, RuiKang OuYang, José Miguel Hernández-Lobato arxiv.org/abs/2601.21025 [𝚜𝚝𝚊𝚝.𝙼𝙻 𝚌𝚜.𝙻𝙶]
2
14
393
Diffusion-based Annealed Boltzmann Generators : benefits, pitfalls and hopes Louis Grenioux, Maxence Noble arxiv.org/abs/2601.21026 [𝚜𝚝𝚊𝚝.𝙼𝙻 𝚌𝚜.𝙻𝙶]
4
13
672
🚀🔍Excited to share my last work on fast diffusion-based image super-resolution done @heyjasperai ! We build on Flow Maps for more photorealistic results, check out the paper for lots of cool visuals : arxiv.org/abs/2601.16660. xitter.cf/maxence_noble/status/1…

Fast, faithful and photorealistic diffusion-based image...

Diffusion-based image super-resolution (SR) has recently attracted significant attention by leveraging the expressive power of large pre-trained text-to-image diffusion models (DMs). A central...

arxiv.org
Maxence Noble@maxence_noble15 Sep 2025
🚀 6-months PhD internship done !🚀 Last day in the Image Research team @heyjasperai, where I developed a novel diffusion model for super resolution that only needs 2 NFE ! ⤵️ With this model, I turned my low-quality profile photo into a 4k photo (zoom in at my left eye) [1/n]
1
11
573
Load more