Ya-Ping Hsieh

[Molei Tao @MoleiTaoMath]

#NeurIPS2025 workshop "DynaFront: Dynamics at the Frontiers of Optimization, Sampling, and Games" will take place on
Sat 12/6

If interested in some theoretical foundation and dynamical processes for ML, plz consider dropping by!

https://t.co/lUl7uKIeGF

[Molei Tao @MoleiTaoMath]

In addition to top notch speakers and awesome posters, we also have amazing panelists:
* Damek Davis @damekdavis (Penn Wharton)
* Jacob Abernethy @Thejakeyboy (Georgia Tech & Google)
* Krishna Balasubramanian @krizna_b (UC Davis & Amazon)
* Yi'an Ma (UCSD)

[Riccardo De Santi @r_de_santi]

How can we perform generative (scientific) discovery beyond the data?

Excited to present our paper “Flow Density Control: Generative Optimization Beyond Entropy-Regularized Fine-Tuning” as ✨Spotlight✨ this Wednesday at NeurIPS 2025.

We employ calculus of variations to strictly generalize RL-based fine-tuning schemes. This unlocks new fundamental capabilities for generative discovery, including generative exploration: the ability to sample from low-probability, yet promising regions, even beyond the original training data.

Feel free to reach out if you’re interested in generative discovery (methods, theory, chem/bio applications) — and check out https://t.co/d8yUOorX47 for updates. We plan to release soon a generative exploration library!

📄 Paper: https://t.co/rT4Q7Ybn3g
⏳Wednesday at 1pm (San Diego time), Exhibit Hall C,D,E #3619

Work done with amazing collaborators @vlastelicap , @yapinghsieh, @ZebangShen, Niao He, and @arkrause .

[Molei Tao @MoleiTaoMath]

Interested in some foundation aspects?
Waiting or unhappy about NeurIPS reviews?

Plz consider NeurIPS workshop
DynaFront: Dynamics at the Frontiers of Optimization, Sampling, and Games
https://t.co/lUl7uKHGR7

@yuejiec @Andrea__M @btreetaiji @T_Chavdarova ++
Sponsor appreciated!

[Riccardo De Santi @r_de_santi]

Generative models are great at mimicking data — but real (scientific) discovery requires going beyond it.

Excited to present our paper “Provable Maximum Entropy Manifold Exploration via Diffusion Models” this Wednesday at ICML 2025!

We propose a scalable, theoretically grounded method to fine-tune a pre-trained diffusion model to become maximally explorative over its learned manifold. This makes it possible to go beyond high-density regions and uncover hidden modes via a novel mechanism for self-guided surprise maximization.

Feel free to reach out if interested — and check out https://t.co/d8yUOosuTF for updates!

📄 Paper: https://t.co/Kt0nD9woLH
⏳Wednesday at 4:30pm (Vancouver time), Hall A-B/E-2011!

Work done with amazing collaborators @vlastelicap, @yapinghsieh, @ZebangShen, Niao He, and @arkrause

[Guan-Horng Liu @guanhorng_liu]

Adjoint-based diffusion samplers have simple & scalable objectives w/o impt weight complication. Like many, though, they solve degenerate Schrödinger bridges, despite all being SB-inspired.

📢 Proudly introduce #Adjoint #Schrödinger #Bridge #Sampler, a full SB-based sampler that is simple to implement, scalable, practically very effective, theoretically sounded, and extends AM beyond memoryless noise schedule.

Great collab w/ @jaemoo51133 @RickyTQChen @bkmi13 @YongxinChen1 🙌🙌

https://t.co/YxTNkx6YgK

[Jeremy Bernstein @jxbz]

👀 @yapinghsieh @CevherLIONS

[Mufan Li @mufan_li]

I’m excited to announce that in July 2025 I will be joining @UWaterloo as an Assistant Professor in the Department of Statistics and Actuarial Science! Until then, I will continue at Princeton as a DataX Postdoc Fellow, working with Boris Hanin.

I have many exciting projects lined up to study the scaling limits of neural networks and their training dynamics, with some important practical implications on the horizon. Stay tuned!

[Amin Karbasi @aminkarbasi]

Now that it is official, my amazing student Chris Harshaw is joining the statistics department @Columbia as an Assistant Professor. Super proud of him. https://t.co/5H3YLFuxWM

[Charlotte Bunne @_bunnech]

Super excited that our work #CellOT was selected among SIB's Remarkable Outputs 2023🎉: Curious about how we can predict single-cell responses to perturbations? Take a look at our @NatureMethods paper (https://t.co/6uRLQPckiN) and the Research Briefing (https://t.co/K5QwdNUqem)!

[Ya-Ping Hsieh @yapinghsieh]

One of my math heroes!

[American Mathematical Society @amermathsoc]

Congratulations to Michel Talagrand of the Centre national de la recherche scientifique (CNRS), Paris, France, who has been awarded the 2024 Abel Prize.

Read more: https://t.co/G93ofexuhU

[Ya-Ping Hsieh @yapinghsieh]

@sp_monte_carlo Not sure if this is what you're looking for, but using Hoffman's inequality, one can prove that a lot of linearly constrained problems satisfy PL but not strong convexity; see e.g. https://t.co/Fy0YhWd9Qe.

[Jia-Jie Zhu @__jzhu__]

https://t.co/5aIO3js4Pm
Anna Korba (ENSAE & CREST, Palaiseau, France)
Sampling through Optimization of Discrepancies
@Korba_Anna

[Jia-Jie Zhu @__jzhu__]

(Slides ✨) Alex Mielke on the Hellinger--Kantorovich (a.k.a. Wasserstein-Fisher-Rao) Spaces and Gradient Flows. HK combines the strength of the Wasserstein and Hellinger distance. Its metric properties and gradient flows are an active area of OT research. https://t.co/5rLkMX9GEq

[Jason Lee @jasondeanlee]

By one of the leaders in data science about predictability, computability and stability!

[Radu Ioan Boț @RaduIoanBot]

At the Faculty of Mathematics of @univienna, we are currently hiring a Tenure-Track Professor in the field of Mathematics of Data Science - Application deadline 15 April 2024.

https://t.co/bJKQnQv7Jt

[Quentin Berthet @qberthet]

🚨 Paper alert: 𝗜𝗺𝗽𝗹𝗶𝗰𝗶𝘁 𝗗𝗶𝗳𝗳𝘂𝘀𝗶𝗼𝗻!

New work on taking gradients through sampling process, e.g. in diffusion models

w @PierreMari0n @Korba_Anna Peter Bartlett @mblondel @ValentinDeBort1 @ArnaudDoucet1 @FelipeLlinares @cypaquette

https://t.co/fPLSyDhUxo

[Ya-Ping Hsieh @yapinghsieh]

A beautiful summarization of "what is time", from a physicist on quantum gravity who thinks deeply psychologically: "Time, for us, is this deeply emotional connection to the world that passes." https://t.co/RHqhH5uaWP

[Joe Benton @JoeJBenton]

Very pleased that our work obtaining the first linear convergence bounds for diffusion models has been awarded a spotlight at ICLR! @ValentinDeBort1 @GeorgeDeligian9 @ArnaudDoucet1

[Ya-Ping Hsieh @yapinghsieh]

@skoularidou You don't look weird. You look confident. 🤘