Results 51 to 60 of about 2,925,819 (300)

Thompson Sampling with Approximate Inference

CoRR, 2019
We study the effects of approximate inference on the performance of Thompson sampling in the $k$-armed bandit problems. Thompson sampling is a successful algorithm for online decision-making but requires posterior inference, which often must be approximated in practice.
My Phan, Yasin Abbasi-Yadkori, Justin Domke   +2 more
openaire   +4 more sources

Distilled Thompson Sampling: Practical and Efficient Thompson Sampling via Imitation Learning

Trans. Mach. Learn. Res., 2020
Thompson sampling (TS) has emerged as a robust technique for contextual bandit problems. However, TS requires posterior inference and optimization for action generation, prohibiting its use in many online platforms where latency and ease of deployment are of concern.
Hongseok Namkoong, Samuel Daulton, Eytan Bakshy   +2 more
openaire   +4 more sources

Thompson Sampling with Diffusion Generative Prior

CoRR, 2023
In this work, we initiate the idea of using denoising diffusion models to learn priors for online decision making problems. Our special focus is on the meta-learning for bandit framework, with the goal of learning a strategy that performs well across bandit tasks of a same class.
Yu-Guan Hsieh, Shiva Prasad Kasiviswanathan, Branislav Kveton, Patrick Blöbaum   +3 more
openaire   +4 more sources

Anticoagulation Treatment Response Among Phenotypic ESUS Clusters: Evidence From CASPR

Annals of Clinical and Translational Neurology, EarlyView.
ABSTRACT Objective To evaluate treatment effect heterogeneity across phenotypic clusters of patients with cryptogenic embolic stroke using causal inference methods. Methods From a retrospective, multicenter cohort of patients with cryptogenic embolic stroke, we estimated the treatment effect of anticoagulation versus antiplatelet therapy in patients ...
Elena Badillo Goicoechea, Patrick Ellsworth, James E. Siegler, James R. Brorson, Christina M. Lineback, Frank G. Conyers, Deborah Kerrigan, Kelsey Eklund, Shadi Yaghi, Farid Khasiyev, Emiliya Melkumova, Collin J. Culbertson, Marissa D'Souza, Dinesh V. Jillella, Patrick Glover, Hassan Aboul‐Nour, Fadi Nahab, Mahan Shahrivari, Simona Nedelcu, David S. Liebeskind, Matthew M. Smith, Aaron Rothstein, Anna Bowman, Alicia Zha, Narendra Kala, Ahmad Abu Qdais, Sami Al Kasab, Mudassir Farooqui, Diana Rojas‐Soto, Yasmin Aziz, Franziska Herpich, Jean‐Philippe Daniel, Sean L. Thompson, Vivek Patel, Thanh N. Nguyen, Siddharth Sehgal, Wenzheng Yu, Wayneho Kam, Russel Van Coevering, Oana M. Dumitrascu, Muhammad Alvi, Balaji Krishnaiah, Setareh Salehi Omran, Guillermo Linares, Eesha Singh, Hamid Ali, Mary Penckofer, Jesse M. Thon, Shani Mannion, Dalia Chahien, Lucia Chen, Fawad Talat, Mehdi Abbasi, Richa Sharma   +53 more
wiley   +1 more source

Trend-aware Thompson sampling algorithm for solving multi-armed bandit problems in e-commerce environment [PDF]

ITM Web of Conferences
Multi-armed bandit algorithms (MAB) have been widely used to make sequential decisions in e-commerce environment such as recommendation systems, online advertising and dynamic pricing strategy adjustments.
Xu Ke
doaj   +1 more source

Lone‐Pair‐Induced Lattice Distortion in Lead‐Free Hybrid Perovskites for High‐Performance Nanogenerators

Advanced Functional Materials, EarlyView.
A lead‐free tin‐based hybrid perovskite is designed in which stereochemically active lone‐pair electrons on tin distort the crystal lattice and break its symmetry. Blended into a flexible polymer, the material converts pressure into electricity at milliwatt‐scale power density and continues working far below freezing, powering a commercial wristwatch ...
Majid Haji Bagheri, Asif Abdullah Khan, Md Fahim Al Fattah, Mahmoud Almadhoun, Shamaila Shahzadi, Aixi Pan, Md Soyaeb Hasan, Henry Wu, Arash Aghigh, Bhawna Bhawna, François Légaré, Vivek Maheshwari, Dayan Ban   +12 more
wiley   +1 more source

Freshness-Aware Thompson Sampling [PDF]

, 2014
To follow the dynamicity of the user's content, researchers have recently started to model interactions between users and the Context-Aware Recommender Systems (CARS) as a bandit problem where the system needs to deal with exploration and exploitation dilemma.
openaire   +3 more sources

Role of Structural Disorder on Phonon Transport and Thermal Conductivity in Li6PS5Br

Advanced Functional Materials, EarlyView.
Solid electrolytes are fast ion and slow heat conductors. Although both transport properties are governed by lattice dynamics and atomic structure, their interplay remains poorly understood. Here, we study this interrelation in anion‐ordered and disordered Li6PS5Br.
Lukas Ketter, Ryoma Sasaki, Kyra Strotmann, Wolfgang G. Zeier, Yoshitaka Tateyama   +4 more
wiley   +1 more source

Collecting diverse near-optimal samples via nested Thompson sampling

npj Computational Materials
Self-driving laboratories (SDLs) that combine automated experiments with machine learning have accelerated data-driven discovery. Although Bayesian optimization (BO) is widely used in SDLs to autonomously propose experimental conditions, many real ...
Ryosuke Shibukawa, Shoichi Matsuda, Kazuha Nakamura, Ryo Tamura, Koji Tsuda   +4 more
doaj   +1 more source

Thompson sampling for improved exploration in GFlowNets

CoRR, 2023
Generative flow networks (GFlowNets) are amortized variational inference algorithms that treat sampling from a distribution over compositional objects as a sequential decision-making problem with a learnable action policy. Unlike other algorithms for hierarchical sampling that optimize a variational bound, GFlowNet algorithms can stably run off-policy,
Jarrid Rector-Brooks, Kanika Madan, Moksh Jain, Maksym Korablyov, Cheng-Hao Liu, Sarath Chandar, Nikolay Malkin, Yoshua Bengio   +7 more
openaire   +3 more sources