Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..

Replicable Online Learning

Authors: Saba Ahmadi, Siddharth Bhandari, Avrim Blum

NeurIPS 2025 | Venue PDF | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical Our paper is theoretical and we provide proofs for all our theorems. No experiments, just theoretical results.
Researcher Affiliation Academia Saba Ahmadi Toyota Technological Institute at Chicago EMAIL & Siddharth Bhandari Toyota Technological Institute at Chicago EMAIL & Avrim Blum Toyota Technological Institute at Chicago EMAIL
Pseudocode Yes Algorithm 1 Follow the Lazy Leader with Block Updates (FLLB(ε, B)) Input: Sequence S = {c1, , c T } arriving one by one over the time. ε > 0 and block size B. 1 Sample p [0, 1 ε)n uniformly at random. 2 Let a1 be a random action picked from A. 3 Let G {p + 1 εz | z Zn}. 4 for t : 1, , T do 5 if t 1 is a multiple of B then 6 gt 1 G c1:t 1 + [0, 1 7 at arg mina A a gt 1. // Stay with the previous action. 8 at at 1. 9 end for
Open Source Code No No experiments only theoretical results. The answer NA means that paper does not include experiments requiring code.
Open Datasets No No experiments, just theoretical results.
Dataset Splits No No experiments, only theoretical results.
Hardware Specification No No experiments, only theoretical results.
Software Dependencies No No experiments, only theoretical results.
Experiment Setup No No experiments, only theoretical results.