arXiv Machine Learning

Leveraging Inference-Time Compute for Diffusion Models via Global Scheduling of Denoising Trajectories

The paper studies how to allocate a fixed computational budget across the denoising steps of diffusion models to improve sample quality at deployment. It shows that the expected benefit of evaluating multiple candidates at a step can be decomposed into a step‑specific sensitivity and a universal sample‑size factor, and that the optimal allocation follows a water‑filling structure. Experiments demonstrate that this allocation achieves the same quality as a uniform strategy while reducing function evaluations by 20–50%.

arXiv Machine Learning
Sep 21

Schedule optimization for tau-leaping in masked discrete diffusion

The paper studies how to choose sampling schedules for tau‑leaping in masked discrete diffusion models. By deriving an exact integral representation of the factorization error ε_fact in terms of a dependence density ρ, the authors develop estimators and recursive equations that identify the unique optimal schedule under a monotonicity condition. In the large‑scale limit, they provide explicit characterizations of the optimal smooth schedule and show that while optimizing smooth schedules can improve constants, it does not change the N/K scaling unless the dependence density degenerates, in which case asymptotic improvements are possible.

By Cecilia Secchi, Giacomo Zanella
arXiv Machine Learning
Sep 25

Spectral-Guided Diffusion: Accelerating Inference via Static Spectral Layer Scheduling

Spectral-Guided Diffusion introduces a method to accelerate diffusion inference by identifying and reusing residual branches that need not be recomputed during the trajectory. The approach uses a Spectral Concentration Ratio (SCR) combined with Frobenius magnitude to create an offline sensitivity proxy and deterministic lifetime for each scheduled unit, eliminating the need for routers or input-dependent searches. Experiments on models such as LLaDA-8B, DiT-XL/2, U-ViT-L, and SDXL show that this scheduling preserves quality better than several baselines and achieves up to a 3.0× wall‑clock speedup over eager inference.

By Ibne Farabi Shihab, Abu Sa-Adat Mohamed Moon-Im Al Ahsan, Anuj Sharma
arXiv Machine Learning
Oct 3

Online Generalized-Mean Welfare Maximization: Achieving Near-Optimal Regret from Samples

The paper investigates online fair allocation of sequential items to agents with heterogeneous preferences, aiming to maximize generalized-mean welfare. In an i.i.d. arrival setting, a pure greedy algorithm achieves near-optimal “~O(1/T)” average regret without needing distributional knowledge. For nonstationary arrivals, the authors show that a single historical sample per distribution suffices to recover the same regret rate, using re-solving algorithms that remain robust to distribution shifts.

By Zongjun Yang, Rachitesh Kumar, Christian Kroer
arXiv Machine Learning
Oct 22

Learning to Allocate Resources with Censored Feedback

The paper addresses an online resource allocation problem where a budget must be distributed among multiple arms under censored feedback. Each arm yields a reward only if it is activated by an unknown Bernoulli process and the allocated budget exceeds a random threshold from an unknown parametric distribution. The authors establish a worst‑case regret lower bound of Σ(√T), propose the RA‑UCB algorithm achieving near‑optimal regret when the budget is known, and introduce MG‑UCB for unknown, round‑dependent budgets, validating both approaches experimentally.

By Giovanni Montanari, C\^ome Fiegel, Corentin Pla, Aadirupa Saha, Vianney Perchet
arXiv Machine Learning
Oct 5

Adaptive Second-Order Solvers for Fast Stochastic Diffusion Sampling

The paper introduces adaptive second‑order solvers for diffusion models, employing a proportional‑integral (PI) step‑size controller based on a diffusion noise‑normalised error estimator. Unlike existing adaptive methods that react only to the current error, the PI solver incorporates past error information, producing smoother step adaptations. The authors also demonstrate that per‑sample trajectories share structure, allowing aggregation into a fixed schedule that largely preserves adaptive benefits. Experiments on natural‑image and language datasets show that the fixed schedule outperforms the commonly used EDM schedule for image sampling with the stochastic Heun sampler, and that the PI adaptive solver achieves lower FID than most baselines, though it does not surpass the EDM‑churn sampler at low neural‑network‑evaluation counts. In language diffusion, the PI solver improves perplexity over EDM and entropy schedules at low‑to‑medium NFE, albeit with reduced token entropy. The study concludes that per‑sample adaptivity is highly beneficial for simple 1D toy problems but only marginally useful for image and language data, where an average schedule can sometimes outperform the PI‑adaptive approach.

By Ella Kemperman, Luca Ambrogioni
arXiv AI
Nov 5

Uncertainty-Aware Resource Orchestration for Agentic AI Systems

The paper introduces Conformal Reservation for Orchestration (CRO), an online algorithm that uses past execution traces to allocate compute and bandwidth for agentic AI systems at dispatch time. CRO provides finite‑sample, distribution‑free coverage guarantees, ensuring that resource reservations meet specified probabilistic deadlines. Simulations demonstrate that CRO uniquely satisfies target coverage across all risk levels while reducing over‑reservation—and thus resource waste—by about half compared to a Chebyshev‑bound baseline.

By Masoud Shokrnezhad, Tarik Taleb