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 [1].

On the Complexity of Differentially Private Best-Arm Identification with Fixed Confidence

Authors: Achraf Azize, Marc Jourdan, Aymen Al Marjani, Debabrota Basu

NeurIPS 2023 | Venue PDF | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Finally, we provide an experimental analysis of Ada P-TT that validates our theoretical results.
Researcher Affiliation Academia Achraf Azize Équipe Scool, Univ. Lille, Inria, CNRS, Centrale Lille, UMR 9189CRISt AL F-59000 Lille, France EMAIL Marc Jourdan Équipe Scool, Univ. Lille, Inria, CNRS, Centrale Lille, UMR 9189CRISt AL F-59000 Lille, France EMAIL Aymen Al Marjani UMPA, ENS Lyon Lyon, France EMAIL Debabrota Basu Équipe Scool, Univ. Lille, Inria, CNRS, Centrale Lille, UMR 9189CRISt AL F-59000 Lille, France EMAIL
Pseudocode Yes Algorithm 1 Sequential interaction between a BAI strategy and users. ... Algorithm 2 Ada P-TT. Private statistics are in red. Changes due to privacy are in blue.
Open Source Code No The paper states 'We implement all the algorithms in Python (version 3.8)...' but does not include any statement about releasing the code for public access, nor does it provide a link to a code repository.
Open Datasets No The paper defines the 'Bernoulli instances' (e.g., 'µ1 = (0.95, 0.9, 0.9, 0.9, 0.5)') by their parameters and cites a previous paper ([SS19]) that defines these instances. However, these are mathematical definitions of distributions for simulation, not references to a downloadable or formally cited 'dataset' in the conventional sense (e.g., a collection of data files).
Dataset Splits No The paper does not mention training, validation, or test dataset splits. The experimental setup involves simulating performance on predefined Bernoulli distributions rather than splitting a static dataset.
Hardware Specification Yes We implement all the algorithms in Python (version 3.8) and on an 8-core 64-bits Intel i5@1.6 GHz CPU.
Software Dependencies Yes We implement all the algorithms in Python (version 3.8)...
Experiment Setup Yes We set the risk δ = 10 2. We implement all the algorithms in Python (version 3.8) and on an 8-core 64-bits Intel i5@1.6 GHz CPU. We run each algorithm 100 times.