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..

Balancing Gradient and Hessian Queries in Non-Convex Optimization

Authors: Deeksha Adil, Brian Bullins, Aaron Sidford, Chenyi Zhang

NeurIPS 2025 | Venue PDF | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical Though our work provides new algorithms and tools for critical point computation, there are several limitations to the result. First, this work is primarily theoretical, no practical implementation or experiments are provided, and in certain cases our bounds incur multiple logarithmic factors.
Researcher Affiliation Academia Institute for Theoretical Studies, ETH Zรผrich EMAIL Department of Computer Science, Purdue University EMAIL Stanford University EMAIL
Pseudocode Yes Algorithm 1: Critical-or-Progress... Algorithm 2: Restarted-Approx-Hessian-AGD... Algorithm 3: Reduction-To-Unbounded-Hessian
Open Source Code No First, this work is primarily theoretical, no practical implementation or experiments are provided, and in certain cases our bounds incur multiple logarithmic factors.
Open Datasets No The paper does not describe the use of any datasets for empirical evaluation, nor does it provide any links or citations to publicly available datasets. The work is theoretical.
Dataset Splits No The paper does not involve empirical studies or dataset usage, therefore, there is no mention of dataset splits.
Hardware Specification No First, this work is primarily theoretical, no practical implementation or experiments are provided, and in certain cases our bounds incur multiple logarithmic factors.
Software Dependencies No The paper is theoretical and does not report on practical implementation or experiments, therefore no specific software dependencies with version numbers are mentioned.
Experiment Setup No First, this work is primarily theoretical, no practical implementation or experiments are provided, and in certain cases our bounds incur multiple logarithmic factors.