Online Learning with Feedback Graphs Without the Graphs

Authors: Alon Cohen, Tamir Hazan, Tomer Koren

ICML 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical We study an online learning framework introduced by Mannor and Shamir (2011) in which the feedback is specified by a graph... We prove that even for dense feedback graphs, the learner cannot improve upon a trivial regret bound... we give an algorithm that achieves eΘ(αT) regret... We also extend our results to a more general feedback model... For our algorithm in the stochastic case, we also prove a distribution-dependent regret bound...
Researcher Affiliation Academia Alon Cohen ALON.COHEN@TECHNION.AC.IL Tamir Hazan TAMIR.HAZAN@TECHNION.AC.IL Tomer Koren TOMERK@TECHNION.AC.IL Technion Israel Institute of Technology, Haifa, Israel
Pseudocode Yes Algorithm 1 input Set V of K actions, number of rounds T initialize r 1, V1 = V while |Vr| > 1 and T rounds have not elapsed do... Algorithm 2 ALPHASAMPLE input Set of actions U V initialize S while |U| > 0 do...
Open Source Code No The paper does not provide any links to open-source code or explicitly state that code for the described methodology is being released.
Open Datasets No The paper is theoretical, presenting algorithms and proving bounds; it does not involve experimental evaluation on datasets. Therefore, no information about training datasets or their public availability is provided.
Dataset Splits No The paper is theoretical and does not report on empirical experiments with dataset splits. Thus, there is no mention of training/validation/test splits.
Hardware Specification No The paper is theoretical and does not involve empirical experiments requiring specific hardware. Therefore, no hardware specifications are mentioned.
Software Dependencies No The paper is theoretical and focuses on algorithms and proofs; it does not specify any software dependencies with version numbers.
Experiment Setup No The paper is theoretical and does not involve empirical experiments. Thus, there are no details provided regarding experimental setup, hyperparameters, or training configurations.