Batch List-Decodable Linear Regression via Higher Moments

Authors: Ilias Diakonikolas, Daniel Kane, Sushrut Karmalkar, Sihan Liu, Thanasis Pittas

ICML 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical This work is predominantly theoretical and does not present any immediate societal or ethical considerations requiring special attention. Its contribution lies in advancing foundational understanding rather than influencing direct real-world applications or policy.
Researcher Affiliation Collaboration 1Department of Computer Science, University of Wisconsin Madison, Madison, United States 2Department of Compuetr Science and Engineering, University of California San Diego, San Diego, United States 3Microsoft Research, Cambridge, England. Correspondence to: Thanasis Pittas <EMAIL>.
Pseudocode Yes Algorithm 1 Batch-List-Decode-LRegression (Informal) Require: Batch sample access to the linear regression instance, σ, R as specified in Theorem 1.3. ... Algorithm 2 Batch-List-Decode-LRegression Require: Batch sample access to the linear regression instance, and α, σ, R, k as specified in Theorem 1.3.
Open Source Code No The paper does not contain an unambiguous statement of code release or a direct link to a source-code repository for the methodology described.
Open Datasets No The paper does not explicitly state that the dataset used in the experiments is publicly available or an open dataset. It defines a theoretical distribution Dβ from which samples are drawn, rather than using a pre-existing named dataset.
Dataset Splits No The paper describes a theoretical framework and does not conduct experiments on specific datasets requiring explicit training/test/validation splits.
Hardware Specification No The paper is theoretical and focuses on algorithmic guarantees and proofs, thus it does not mention any specific hardware used for experiments.
Software Dependencies No The paper is theoretical and discusses algorithms and mathematical frameworks; it does not provide details of software dependencies with specific version numbers.
Experiment Setup No The paper is theoretical and focuses on algorithmic design and proofs, not on empirical experimental setups or hyperparameter tuning.