The paper introduces TRACK, a training‑free trajectory routing method that accelerates video diffusion by selectively switching between large and small models during denoising steps. A calibration process generates a disagreement score map, guiding the selection of the appropriate model at each step to maintain quality while reducing computational cost. Experiments on Wan 2.1, Cosmos 3, TurboDiffusion, and FastVideo show speedups ranging from 1.95× to 2.73× with comparable quality and diversity.
By Mustafa Munir, Huy Vu, Shreyas Misra, Rohit Jena, Sajad Norouzi, Ali Taghibakhshi, Anis Ahmad, Anjul Patney, Pavlo Molchanov, Nima Tajbakhsh
Accelerating Video Diffusion via Training-Free Trajectory Routing (TRACK) introduces a heterogeneous denoising strategy that switches between large and small diffusion models at selected steps, determined by a calibration process that measures disagreement between model predictions. By routing quality-sensitive steps to the large model and low-disagreement steps to the small model, TRACK achieves significant speedups—up to 2.73×—across several video diffusion benchmarks while maintaining comparable quality and diversity. The method requires no retraining, architectural changes, or online dual-model evaluation, making it a practical acceleration paradigm for video diffusion.
Double‑Diffusion is a probabilistic forecasting framework for urban sensor networks that blends a closed‑form graph‑heat prior with a denoising diffusion model. The prior provides a low‑pass, parameter‑free forecast that both serves as the residual target and conditions the denoiser, allowing the reverse diffusion process to start near the prior and refine it over a short chain. Experiments on four real‑world air‑quality and traffic datasets show that Double‑Diffusion achieves the lowest CRPS among probabilistic methods while remaining competitive in point accuracy and requiring far fewer sampling steps than traditional from‑noise diffusion models.
By Hanlin Dong, Arian Prabowo, Hao Xue, Shuang Ao, Tianyi Zhou, Yuxuan Liang, Flora D. Salim
arXiv:2606. 05688v1 Announce Type: cross Abstract: Mixture-of-Experts (MoE) models scale foundation models efficiently by activating only a subset of experts for each token, but their large number of expert parameters still makes quantization essential for practical deployment.
By Hancheol Park, Geonho Lee, Tairen Piao, Tae-Ho Kim
arXiv:2607. 26618v1 Announce Type: new Abstract: Federated PEFT enables LLMs to collaboratively adapt to decentralized private data without sharing raw examples.
By Donghang Duan, Xu Zheng, Lizong Zhang, Chong Mu, Meng Han
The paper examines how Denoising Diffusion Probabilistic Models (DDPMs) perform on globally constrained discrete tasks such as Sudoku and N-queens. It shows that standard diffusion sampling, which keeps updates close to the noisy state, often preserves early mistakes, whereas sampling directly from the model’s clean predictions dramatically improves validity (e.g., Sudoku from 31% to 95%). The authors further introduce self‑correction training, exposing the model to its own predictions to reduce inference errors, which enhances the performance of standard samplers across tasks.
By Mariia Drozdova, St\'ephane Liem Nguyen, Fran\c{c}ois Fleuret
The paper presents a statistical framework for Mixture-of-Experts (MoE) models, treating them as localized aggregation systems. It derives oracle risk bounds that separate approximation, expert‑learning, and router‑estimation errors for both dense and sparse routing with evolving experts. The authors also analyze how sparse Top‑K routing balances computational cost with performance, interpret gating geometrically, and explain how shared experts can capture common predictive structure while allowing routed experts to focus on local residuals.
By Siyuan He, Bokai Yang, Jie Hu, Ziwen Gao, Yuhong Yang
arXiv:2609.09241v1 Announce Type: cross
Abstract: Mixture-of-Experts (MoE) architectures have emerged as a powerful paradigm for scaling model capacity while preserving efficient inference in large f...
By Dohyeon Kim, Bedionita Soro, Sung Ju Hwang
The paper studies how the design of Mixture-of-Experts (MoE) routers affects inference speed when combined with Speculative Decoding (SD). It shows that routers promoting high expert coactivation reduce memory transfer costs and improve runtime. By integrating a global load‑balancing loss, shared experts, a consistency loss, and an autoregressive expert selection mechanism, the authors achieve a 21% throughput gain over baseline MoEs while preserving accuracy.
By Kumari Nishu, Han-Byul Kim, Santosh Chilkunda, Maxwell Horton, Arnav Kundu, Mohammad Samragh, Lauren Hannah, Mohammad Sekhavat, Nikhil Bhendawade, Manuel Ciosici, Iman Mirzadeh, Keivan Alizadeh Vahid, David Harrison, Irina Belousova, Mehrdad Farajtabar, Minsik Cho
Federated PEFT enables LLMs to collaboratively adapt to decentralized private data without sharing raw examples. However, task heterogeneity across clients can cause cross-task interference and gradient conflicts during aggregation.
arXiv:2607. 24665v1 Announce Type: cross Abstract: Modern large language models scale successfully by pairing capacity growth with efficiency, keeping per-token and deployment costs under control as capacity grows.
By Yanhao Jia, Jiepeng Wang, Haibin Huang, Chi Zhang, Erik Cambria, Xuelong Li
The paper investigates why diffusion models, despite their practical success, can fully memorize training data when optimal. It shows that as training proceeds, the denoising objective causes overfitting at intermediate noise levels, creating a gap between validation and training performance. In a 2D toy model, this gap arises from the interplay between model error and the data distribution’s support density, leading to a smooth, generalizing flow field rather than exact recall of training points. The study also reveals that predictions from intermediate inference states become increasingly similar to training data in feature space, even as overall similarity to validation data remains unchanged.
By Tim Kaiser, Markus Kollmann