Training Latent Diffusion Models with Interacting Particle Algorithms
arXiv:2505.12412v3 Announce Type: replace-cross Abstract: We introduce a novel particle-based algorithm for end-to-end training of latent diffusion models. We reformulate the training task as minimizing a free energy functional and obtain a gradient flow that does so. By approximating the latter with a system of interacting particles, we obtain the algorithm, which we underpin theoretically by providing error guarantees. The novel algorithm compares favorably in experiments with previous particle-based methods and variational inference analogues. — Tim Y. J. Wang, Juan Kuntz, O. Deniz Akyildiz
View PDF
Abstract:We introduce a novel particle-based algorithm for end-to-end training of latent diffusion models. We reformulate the training task as minimizing a free energy functional and obtain a gradient flow that does so. By approximating the latter with a system of interacting particles, we obtain the algorithm, which we underpin theoretically by providing error guarantees. The novel algorithm compares favorably in experiments with previous particle-based methods and variational inference analogues.
Comments: Camera Ready version for AISTATS 2026
Subjects:
Machine Learning (stat.ML); Machine Learning (cs.LG)
Cite as: arXiv:2505.12412 [stat.ML]
(or arXiv:2505.12412v3 [stat.ML] for this version)
https://doi.org/10.48550/arXiv.2505.12412
arXiv-issued DOI via DataCite
Submission history
From: Tim Y. J. Wang [view email] [v1] Sun, 18 May 2025 13:29:07 UTC (7,102 KB) [v2] Fri, 23 May 2025 18:19:28 UTC (7,092 KB) [v3] Sun, 29 Mar 2026 12:10:08 UTC (6,840 KB)
Sign in to highlight and annotate this article
Conversation starters
Daily AI Digest
Get the top 5 AI stories delivered to your inbox every morning.
More about
researchpaperarxiv
Google's TurboQuant saves memory, but won't save us from DRAM-pricing hell
<h4>Chocolate Factory’s compression tech clears the way to cheaper AI inference, not more affordable memory</h4> <p>When Google unveiled <a target="_blank" rel="nofollow" href="https://research.google/blog/turboquant-redefining-ai-efficiency-with-extreme-compression/">TurboQuant</a>, an AI data compression technology that promises to slash the amount of memory required to serve models, many hoped it would help with a memory shortage that has seen prices triple since last year. Not so much.…</p>
Illinois Tech computer science researcher honored by IEEE Chicago Section - EurekAlert!
<a href="https://news.google.com/rss/articles/CBMiXEFVX3lxTE13OVpWMEk1Z3hlMkR2bHNBQ2dkazFwb3VqN3hCa29GWGJvSVlPa00zd2xUakRmYXFqQmc5OWU0eGl4a21FMDAwWUN2Q3p0M3FrbXBkNV8zN0cxaG1s?oc=5" target="_blank">Illinois Tech computer science researcher honored by IEEE Chicago Section</a> <font color="#6f6f6f">EurekAlert!</font>
My Journey to becoming a Quantum Engineer
<p>I have procrastinated on documenting this process for the longest time. But I think i am ready now (maybe). <br> Coming from a front end engineering background, I am fascinated by the work being done by the quantum engineers at IBM. I am not that great with maths and statistics but I believe anything can be learned with tons of practice and consistency. I want to use this platform to hold myself accountable (that is if i don't give up half way and delete all my posts. I'll try not to btw). </p> <p>This is an article describing <a href="https://www.ibm.com/think/topics/quantum-computing" rel="noopener noreferrer">what quantum computing is</a> and some of it's use cases.</p> <p>I became an IBM qiskit advocate late last year and I have been exposed to a lot of resources and networked a bun
Knowledge Map
Connected Articles — Knowledge Graph
This article is connected to other articles through shared AI topics and tags.
More in Research Papers
Illinois Tech computer science researcher honored by IEEE Chicago Section - EurekAlert!
<a href="https://news.google.com/rss/articles/CBMiXEFVX3lxTE13OVpWMEk1Z3hlMkR2bHNBQ2dkazFwb3VqN3hCa29GWGJvSVlPa00zd2xUakRmYXFqQmc5OWU0eGl4a21FMDAwWUN2Q3p0M3FrbXBkNV8zN0cxaG1s?oc=5" target="_blank">Illinois Tech computer science researcher honored by IEEE Chicago Section</a> <font color="#6f6f6f">EurekAlert!</font>
AI maps science papers to predict research trends two to three years ahead - Tech Xplore
<a href="https://news.google.com/rss/articles/CBMie0FVX3lxTE5aTkZYTWdaRDZwTXNRMldpMG1WZ1YzWDZTOHN5M183Z3A1ZTFYbnhEWTdPRmpvZnZFU0xodlRsNWxFaGxTcEpwalhJNmJpQWE5VjhaRS1tOXJIeTc5Z0JNblJ3dFd4WjRYZGJOX0NrWGt6ZmZJVTBpRm5wWQ?oc=5" target="_blank">AI maps science papers to predict research trends two to three years ahead</a> <font color="#6f6f6f">Tech Xplore</font>
AI inspires new research topics in materials science - Nanowerk
<a href="https://news.google.com/rss/articles/CBMiZ0FVX3lxTFBPWlJSM2ExeVQ3LVppTm45NHpEMW9YVkxscThCNDd2OVB0c3J1ZmVCbWNSZWZ0TjZwSzlOdEFXN2UtRk5LU1hxdXd4ZklldGxoM0FZSnhCd19PWkNHQ1ZRVDNwSHNUSk0?oc=5" target="_blank">AI inspires new research topics in materials science</a> <font color="#6f6f6f">Nanowerk</font>
Discussion
Sign in to join the discussion
No comments yet — be the first to share your thoughts!