Results 81 to 90 of about 3,478,919 (201)

Thompson Sampling for Non-Stationary Bandit Problems

Entropy
Non-stationary multi-armed bandit (MAB) problems have recently attracted extensive attention. We focus on the abruptly changing scenario where reward distributions remain constant for a certain period and change at unknown time steps.
Han Qi, Fei Guo, Li Zhu
doaj   +1 more source

Multi-armed Bandits in Action

, 2020
Making decisions is part and parcel of being human. Among a set of actions, we want to choose the one that has the highest reward. But the uncertainty of the outcome prevents us from always making the right decision. Making decisions under uncertainty can be studied in a principled way by the exploitation-exploration framework.
openaire   +2 more sources

Gaussian Process with Vine Copula-Based Context Modeling for Contextual Multi-Armed Bandits

Mathematics
We propose a novel contextual multi-armed bandit (CMAB) framework that integrates copula-based context generation with Gaussian Process (GP) regression for reward modeling, addressing complex dependency structures and uncertainty in sequential decision ...
Jong-Min Kim
doaj   +1 more source

A Survey for Deep Reinforcement Learning Based Network Intrusion Detection

Applied AI Letters, Volume 7, Issue 2, June 2026.
This paper surveys deep reinforcement learning (DRL) for network intrusion detection, evaluating model efficiency, minority attack detection, and dataset imbalance. Findings show DRL achieves state‐of‐the‐art results on public datasets, sometimes surpassing traditional deep learning.
Wanrong Yang, Alberto Acuto, Yihang Zhou, Dominik Wojtczak   +3 more
wiley   +1 more source

Lifelong Learning in Multi-Armed Bandits

CoRR, 2020
Continuously learning and leveraging the knowledge accumulated from prior tasks in order to improve future performance is a long standing machine learning problem. In this paper, we study the problem in the multi-armed bandit framework with the objective to minimize the total regret incurred over a series of tasks.
Matthieu Jedor, Jonathan Louëdec, Vianney Perchet   +2 more
openaire   +2 more sources

Strategic Adaptation in Shifting Landscapes: A Paradox‐Theoretic and Computational Perspective

Canadian Journal of Administrative Sciences / Revue Canadienne des Sciences de l'Administration, Volume 43, Issue 2, June 2026.
ABSTRACT Organizations must simultaneously exploit current competencies and explore new opportunities—yet theory offers limited guidance on when each strategic logic dominates and under what combination of environmental conditions ambidexterity is warranted.
Javad Feiz Abadi
wiley   +1 more source

RAAS: Runtime Adaptive Approximation System

Concurrency and Computation: Practice and Experience, Volume 38, Issue 12, June 2026.
ABSTRACT Software‐level approximation techniques, such as loop perforation and functional replacement, can improve energy and performance in error‐tolerant applications, but typically require extensive programmer intervention even when compiler support is available.
Lucas Reis, Sandro Rigo, Lucas Wanner
wiley   +1 more source

Modified Index Policies for Multi-Armed Bandits with Network-like Markovian Dependencies

Network
Sequential decision-making in dynamic and interconnected environments is a cornerstone of numerous applications, ranging from communication networks and finance to distributed blockchain systems and IoT frameworks. The multi-armed bandit (MAB) problem is
Abdalaziz Sawwan, Jie Wu
doaj   +1 more source

Dynamic Pipeline Scheduling for Hybrid Edge Collaboration in Multi‐Modal LLM Inference

Artificial Intelligence for Engineering, Volume 2, Issue 2, Page 180-194, June 2026.
A hybrid edge‐collaboration paradigm is proposed that combines in‐device NPU/DLA–GPU co‐inference with across‐device model‐parallel pipelines to accelerate Multi‐modal Large Language Model inference. A UCB‐based dynamic scheduling algorithm maximizes system throughput while meeting latency constraints, achieving significant speedups over single‐device ...
He Li
wiley   +1 more source

A multi-armed bandit approach for exploring partially observed networks

Applied Network Science, 2019
Background real-world networks such as social and communication networks are too large to be observed entirely. Such networks are often partially observed such that network size, network topology, and nodes of the original network are unknown.
Kaushalya Madhawa, Tsuyoshi Murata
doaj   +1 more source