Adapter-Augmented Bandits for Online Multi-Constrained Multi-Modal Inference Scheduling
School of Computer Science and Engineering, Sun Yat-sen University, Guangzhou, 510006, China.（中山大学计算机科学与工程学院）；School of Computing, Macquarie University, NSW 2109, Australia（麦考瑞大学计算机学院）
Xianzhi Zhang, Yue Xu, Yinlin Zhu, Di Wu, Yipeng Zhou, Miao Hu, Guocong Quan
Multi-modal large language model (MLLM) inference scheduling enables strong response quality under practical and heterogeneous budgets, beyond what a homogeneous single-backend setting can offer. Yet online MLLM task scheduling is nontrivial, as requests vary sharply in modality composition and latent reasoning difficulty, while execution backends incur distinct, time-varying costs due to system jitter and network variation. These coupled uncertainties pose two core challenges: deriving semantically faithful yet scheduling-relevant multi-modal task representations, and making low-overhead online decisions over irreversible multi-dimensional budgets. Accordingly, we propose \emph{M-CMAB} (\underline{M}ulti-modal \underline{M}ulti-constraint \underline{C}ontextual \underline{M}ulti-\underline{A}rmed \underline{B}andit), a multi-adapter-enhanced MLLM inference scheduling framework with three components: (i) a CLS-attentive, frozen-backbone \emph{Predictor} that extracts compact task representations and updates only lightweight adapters for action-specific estimation; (ii) a primal-dual \emph{Constrainer} that maintains online Lagrange multipliers to enforce long-horizon constraints via per-round objectives; and (iii) a two-phase \emph{Scheduler} that balances exploration and exploitation under irreversible budgets. We establish a regret guarantee under multi-dimensional knapsack constraints. On a composite multimodal benchmark with heterogeneous backends, \emph{M-CMAB} consistently outperforms state-of-the-art baselines across budget regimes, achieving up to 14.18% higher reward and closely tracking an oracle-aided upper bound. Codes are available at https://anonymous.4open.science/r/M2CMAB/.
Institute for Interdisciplinary Information Sciences（交叉信息研究院）；Tsinghua University（清华大学）；Microsoft Research（微软研究院）；The University of Tokyo（东京大学）；RIKEN AIP（理化学研究所人工智能项目中心）；CENTAI Institute（CENTAI研究所）
Contextual Rollout Bandits for Reinforcement Learning with Verifiable Rewards
School of Computer Science and Engineering, Beihang University（北京航空航天大学计算机科学与工程学院）；School of Artificial Intelligence, Beihang University（北京航空航天大学人工智能学院）；Huawei（华为）
Constrained Restless Bandits for Dynamic Scheduling in Cyber-Physical Systems
Department of Electrical Engineering, Indian Institute of Technology Bombay（印度理工学院班加罗尔电子工程系）；Polytechnique Montreal（蒙特利尔理工学院）；IIIT Allahabad（阿哈迈德纳巴德印度理工学院）；University of Ottawa（渥太华大学）；IIT Bombay（印度理工学院班加罗尔）
Reinforcement Learning for Search Tree Size Minimization in Constraint Programming: New Results on Scheduling Benchmarks
Faculty of Electrical Engineering, Czech Technical University in Prague（电气工程系，布拉格捷克技术大学）；Czech Institute of Informatics, Robotics and Cybernetics, Czech Technical University in Prague（捷克信息学、机器人学与自动化研究所，布拉格捷克技术大学）
Learning to Identify Top Elo Ratings: A Dueling Bandits Approach
Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）；School of Artificial Intelligence, University of Chinese Academy of Sciences（中国科学院大学人工智能学院）；University of Oxford（牛津大学）；University College London（伦敦大学学院）；Beijing Key Laboratory of Big Data Management and Analysis Methods, GSAI, Renmin University of China（中国人民大学高瓴人工智能学院大数据管理与分析方法北京市重点实验室）
Decision-Focused On-Policy Learning for Contextual Linear Optimization with Partial Feedback
Department of Industrial Engineering and Operations Research, University of California, Berkeley（工业工程与运筹学系，加州大学伯克利分校）；H. Milton Stewart School of Industrial and Systems Engineering, Georgia Institute of Technology（H.米尔顿·斯图尔特工业与系统工程学院，佐治亚理工学院）
Zuse Institute Berlin（柏林祖斯研究所）；Hochschule für Technik und Wirtschaft Berlin（柏林工程与经济应用技术大学）；Google Research（谷歌研究院）；Technische Universität Berlin（柏林工业大学）
Decentralized Rank Scheduling for Energy-Constrained Multi-Task Federated Fine-Tuning in Edge-Assisted IoV Networks
School of Computer Science and Engineering, Sun Yat-sen University（中山大学计算机科学与工程学院）；School of Cyber Science and Technology, Sun Yat-sen University（中山大学网络安全科学与技术学院）
Making the most of your day: online learning for optimal allocation of time
Centre Borelli（博雷利中心）；ENS Paris-Saclay（巴黎萨克雷高等师范学院）；EDF Lab（法国电力公司实验室）；CREST（法国国家科学研究中心经济学与统计学研究中心）；ENSAE Paris（巴黎高等统计学院）；Criteo AI Lab（Criteo人工智能实验室）；LUISS（路易斯大学）
A Lyapunov-Based Methodology for Constrained Optimization with Bandit Feedback
Coordinated Science Laboratory, University of Illinois at Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校协调科学实验室）；Department of Electrical and Computer Engineering, The Ohio State University（俄亥俄州立大学电气与计算机工程系）
Co-occurrence Patterns of LoRA Adapters in Production Diffusion Model Inference Services
Guizhou University of Traditional Chinese Medicine（贵州中医药大学）；Guizhou University of Finance and Economics（贵州财经大学）；Beijing Technology and Business University（北京工商大学）
High-Probability Nash Regret for Decentralized Learning in Markov $α$-Potential Games: Episodic and Fully Online Asynchronous Algorithms with Applications to Markov Congestion Games
University of Illinois Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）
A resource-constrained stochastic scheduling algorithm for homeless street outreach and gleaning edible food
IBM Research（IBM研究院）；Lawrence Livermore National Laboratory（劳伦斯利弗莫尔国家实验室）；Microsoft New England（微软新英格兰研究院）；University of California, Berkeley（加州大学伯克利分校）；IBM Research – T.J. Watson（IBM T.J.沃森研究院）；Fordham University（福特汉姆大学）；Leket Israel（以色列莱凯特组织）；Breaking Ground（Breaking Ground（美国无家可归者救助机构））；Change Machine（Change Machine（美国社会服务机构））
