MARA: Flow-Matching-Guided Multi-Agent Resource Allocation for Computational Resource Efficient Learning

• URL: http://arxiv.org/abs/2608.09130v1
• Date: Mon, 10 Aug 2026 05:13:56 GMT
• Title: MARA: Flow-Matching-Guided Multi-Agent Resource Allocation for Computational Resource Efficient Learning
• Abstract summary: We study a practical setting in which tasks arrive over time and computation is provided by discrete nodes.<n>We propose MARA, which predicts future loss trajectories with conditional flow matching and coordinates compute nodes through a cooperative multi-agent autoregressive policy.<n>At the scheduler's training load, MARA completes 63.46% of tasks on average, 8.54 percentage points above strong baseline Learning with Adaptive Resource Allocation (LARA)
• Score: 29.198856168513743
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Abstract: Allocating limited computation among concurrent learning tasks is difficult when each task must reach a target loss before a deadline but its required training effort is unknown. Existing approaches combine online loss prediction with adaptive resource allocation, yet commonly treat computation as continuously divisible throughput. We instead study a practical setting in which tasks arrive over time and computation is provided by discrete nodes. This setting introduces both uncertain demand and constrained sequential decisions. We propose MARA, which predicts future loss trajectories with conditional flow matching and coordinates compute nodes through a cooperative multi-agent autoregressive policy. A potential-based progress reward supplies intermediate training feedback while preserving the undiscounted task-completion objective. Across in-distribution, reinforcement-learning, and vision workloads, flow matching reduces remaining-resource prediction error relative to weighted least squares. At the scheduler's training load, MARA completes 63.46% of tasks on average, 8.54 percentage points above strong baseline Learning with Adaptive Resource Allocation (LARA), and remains ahead under unseen heavier workloads.

Related papers

• Beyond Solvability: Task Learnability as a Static Prior for LLM RL Post-Training [52.78248352851176]
  Reinforcement learning (RL) has become a central post-training paradigm for eliciting reasoning capabilities in large language models.<n>We study this residual axis as task learnability: a regime-conditional measure of expected positive response to continued training.<n>We propose TrajVal, a lightweight probe-based estimator that approximates per-task learnability from a short probe run and two endpoint evaluations.
  arXiv  Detail & Related papers  (2026-08-10T07:43:05Z)
• Edge-AI-Driven Learning-to-Rank for Decentralized Task Allocation in Circular Smart Manufacturing [54.4933359656657]
  This work proposes an Edge-AI-driven decentralized task allocation framework based on ranking-aware negotiation.<n>A ranking-aware formulation finally reshapes the learning objective to align with the ordering-based nature of winner selection.<n>Results show improved delay and deadline adherence under high load, and enhanced energy efficiency under tighter constraints.
  arXiv  Detail & Related papers  (2026-05-14T20:38:25Z)
• Unveiling Downstream Performance Scaling of LLMs: A Clustering-Based Perspective [7.408649506385476]
  The escalating scale and cost of Large Language Models (LLMs) training necessitate accurate pre-training prediction of downstream task performance.<n>Current prediction methods lack accuracy and reliability.<n>We propose a Clustering-On-Difficulty (COD) framework for downstream performance prediction.
  arXiv  Detail & Related papers  (2025-02-24T15:44:57Z)
• To Train or Not to Train: Balancing Efficiency and Training Cost in Deep Reinforcement Learning for Mobile Edge Computing [15.079887992932692]
  We present a new algorithm to dynamically select when to train a Deep Reinforcement Learning (DRL) agent that allocates resources. Our method is highly general, as it can be directly applied to any scenario involving a training overhead.
  arXiv  Detail & Related papers  (2024-11-11T16:02:12Z)
• Learning to Schedule Online Tasks with Bandit Feedback [7.671139712158846]
  Online task scheduling serves an integral role for task-intensive applications in cloud computing and crowdsourcing. We propose a double-optimistic learning based Robbins-Monro (DOL-RM) algorithm. DOL-RM integrates a learning module that incorporates optimistic estimation for reward-to-cost ratio and a decision module.
  arXiv  Detail & Related papers  (2024-02-26T10:11:28Z)
• Network Contention-Aware Cluster Scheduling with Reinforcement Learning [0.10878040851637999]
  We present a new approach to network contention in GPU clusters using reinforcement learning. We show that our approach reduces average job completion time by up to 18.2% and effectively cuts the tail job completion time by up to 20.7%.
  arXiv  Detail & Related papers  (2023-10-31T06:17:23Z)
• A Multi-Head Ensemble Multi-Task Learning Approach for Dynamical Computation Offloading [62.34538208323411]
  We propose a multi-head ensemble multi-task learning (MEMTL) approach with a shared backbone and multiple prediction heads (PHs) MEMTL outperforms benchmark methods in both the inference accuracy and mean square error without requiring additional training data.
  arXiv  Detail & Related papers  (2023-09-02T11:01:16Z)
• Continual Learning with Distributed Optimization: Does CoCoA Forget? [0.0]
  We focus on the continual learning problem where the tasks arrive sequentially. The aim is to perform well on the newly arrived task without performance degradation on the previously seen tasks. We consider the well-established distributed learning algorithm COCOA.
  arXiv  Detail & Related papers  (2022-11-30T13:49:43Z)
• Joint Client Scheduling and Resource Allocation under Channel Uncertainty in Federated Learning [47.97586668316476]
  Federated learning (FL) over wireless networks depends on the reliability of the client-server connectivity and clients' local computation capabilities. In this article, we investigate the problem of client scheduling and resource block (RB) allocation to enhance the performance of model training using FL. A proposed method reduces the gap of the training accuracy loss by up to 40.7% compared to state-of-theart client scheduling and RB allocation methods.
  arXiv  Detail & Related papers  (2021-06-12T15:18:48Z)
• Coordinated Online Learning for Multi-Agent Systems with Coupled Constraints and Perturbed Utility Observations [91.02019381927236]
  We introduce a novel method to steer the agents toward a stable population state, fulfilling the given resource constraints. The proposed method is a decentralized resource pricing method based on the resource loads resulting from the augmentation of the game's Lagrangian.
  arXiv  Detail & Related papers  (2020-10-21T10:11:17Z)
• Dif-MAML: Decentralized Multi-Agent Meta-Learning [54.39661018886268]
  We propose a cooperative multi-agent meta-learning algorithm, referred to as MAML or Dif-MAML. We show that the proposed strategy allows a collection of agents to attain agreement at a linear rate and to converge to a stationary point of the aggregate MAML. Simulation results illustrate the theoretical findings and the superior performance relative to the traditional non-cooperative setting.
  arXiv  Detail & Related papers  (2020-10-06T16:51:09Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.