Estimating Mutual Information for Discrete-Continuous Mixtures

Authors: Weihao Gao, Sreeram Kannan, Sewoong Oh, Pramod Viswanath

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

Reproducibility Variable Result LLM Response
Research Type Experimental We prove the consistency of this estimator theoretically as well as demonstrate its excellent empirical performance. This problem is relevant in a wide-array of applications, where some variables are discrete, some continuous, and others are a mixture between continuous and discrete components. ... Section 5 contains the results of our detailed synthetic and real-world experiments testing the efficacy of the proposed estimator.
Researcher Affiliation Academia Weihao Gao Department of ECE Coordinated Science Laboratory University of Illinois at Urbana-Champaign wgao9@illinois.edu Sreeram Kannan Department of Electrical Engineering University of Washington ksreeram@uw.edu Sewoong Oh Department of IESE Coordinated Science Laboratory University of Illinois at Urbana-Champaign swoh@illinois.edu Pramod Viswanath Department of ECE Coordinated Science Laboratory University of Illinois at Urbana-Champaign pramodv@illinois.edu
Pseudocode Yes Algorithm 1 Mixed Random Variable Mutual Information Estimator
Open Source Code No The paper does not provide any statements about open-sourcing the code for the described methodology, nor does it include any links to a code repository.
Open Datasets Yes Gene regulatory network inference. ... Instead we resorted to a challenge dataset for reconstructing regulatory networks, called the DREAM5 challenge [30]. The simulated (insilico) version of this dataset contains gene expression for 20 genes with 660 data point containing various perturbations.
Dataset Splits No The paper describes synthetic data generation and real-world datasets but does not specify training, validation, or test splits for any of its experiments. It focuses on mean squared error versus sample size.
Hardware Specification No The paper does not provide any specific hardware details (e.g., CPU, GPU models, memory, or cloud instance types) used for running the experiments.
Software Dependencies No The paper does not provide specific software dependencies with version numbers (e.g., Python, PyTorch, or other libraries).
Experiment Setup No The paper describes the characteristics of the synthetic data generated for experiments (e.g., 'X is uniformly distributed over integers {0, 1, . . . , m 1} and Y is uniformly distributed over the range [X, X + 2] for a given X'). However, it does not provide specific hyperparameter values or system-level training settings used for its estimator in these experiments.