Near-Optimal Design of Experiments via Regret Minimization

Authors: Zeyuan Allen-Zhu, Yuanzhi Li, Aarti Singh, Yining Wang

ICML 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Numerical results on synthetic and real-world design problems verify the practical effectiveness of the proposed algorithm.
Researcher Affiliation Collaboration 1Microsoft Research, Redmond, USA 2Princeton University, Princeton, USA 3Carnegie Mellon University, Pittsburgh, USA.
Pseudocode Yes In Algorithm 1 we give a pseducode description of the algorithm, which makes use of a binary search routine (Algorithm 2) that finds the unique constant ct for which tr(At) = tr[(ct I + P i Λt 1 xix i ) 2] = 1.
Open Source Code No The paper states that algorithms are implemented in MATLAB and C, but does not provide any links or explicit statements about the availability of the source code for the described methodology.
Open Datasets No For the
Dataset Splits No The paper does not specify explicit training, validation, or test dataset splits.
Hardware Specification No The paper mentions that algorithms are implemented in MATLAB and C, but does not specify any hardware details such as CPU, GPU models, or memory.
Software Dependencies No The paper states that algorithms are implemented in MATLAB and C, but does not provide specific version numbers for these environments or any other software libraries.
Experiment Setup Yes The maximum number of iterations T for Fedorov s exchange is set at T = 100. We always set k = r in the optimization problem Eq. (6)... In Algorithm 1 we set α = 10; our similuations suggest that the algorithm is not sensitive to α.