Accelerating ERM for data-driven algorithm design using output-sensitive techniques

Authors: Maria-Florina F. Balcan, Christopher Seiler, Dravyansh Sharma

NeurIPS 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical We design novel approaches that use tools from computational geometry and lead to output-sensitive algorithms for learning good parameters by implementing the ERM (Empirical Risk Minimization) for several distinct data-driven design problems. The resulting learning algorithms scale polynomially with the number of sum dual class function pieces RΣ in the worst case (See Table 1) and are efficient for small constant d.
Researcher Affiliation Academia Carnegie Mellon University, ninamf@cs.cmu.edu. Work done by Christopher Seiler while he was at CMU. Corresponding author: dravy@ttic.edu. Work done by Dravyansh Sharma while he was at CMU.
Pseudocode Yes Algorithm 1: OUTPUTSENSITIVEPARTITIONSEARCH
Open Source Code No The paper does not include an unambiguous statement about releasing its source code or a direct link to a code repository.
Open Datasets No The paper discusses 'problem instances' and 'problem samples' but does not provide concrete access information (link, DOI, repository, or formal citation) for a specific publicly available or open dataset used in its analysis.
Dataset Splits No The paper does not specify exact train/validation/test split percentages, absolute sample counts, or reference predefined splits for reproducibility.
Hardware Specification No The paper does not provide specific hardware details (exact GPU/CPU models, processor types, or memory amounts) used for running experiments.
Software Dependencies No The paper does not list specific software components with version numbers (e.g., Python 3.8, PyTorch 1.9) or self-contained solvers (e.g., CPLEX 12.4) needed to replicate the work.
Experiment Setup No The paper does not provide specific experimental setup details such as concrete hyperparameter values, training configurations, or system-level settings.