Caveats for information bottleneck in deterministic scenarios

Authors: Artemy Kolchinsky, Brendan D. Tracey, Steven Van Kuyk

ICLR 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We demonstrate the three caveats on the MNIST dataset.
Researcher Affiliation Academia Artemy Kolchinsky & Brendan D. Tracey Santa Fe Institute Santa Fe, NM 87501, USA {artemyk,tracey.brendan}@gmail.com Steven Van Kuyk School of Engineering and Computer Science Victoria University of Wellington, New Zealand steven.jvk@gmail.com Dept of Aeronautics & Astronautics, Massachusetts Institute of Technology, Cambridge, MA 02139, USA
Pseudocode No The paper describes the neural network architecture and training process in text, but it does not include any structured pseudocode or algorithm blocks.
Open Source Code Yes Tensor Flow code can be found at https://github.com/artemyk/ibcurve .
Open Datasets Yes We demonstrate the three caveats using the MNIST dataset of hand-written digits. ... This dataset contains a training set of 60,000 images and a test set of 10,000 images, each labeled according to digit.
Dataset Splits Yes This dataset contains a training set of 60,000 images and a test set of 10,000 images, each labeled according to digit.
Hardware Specification No The paper does not provide specific details about the hardware (e.g., CPU, GPU models, memory) used for running the experiments.
Software Dependencies No The paper mentions "Tensor Flow code" and the "Adam algorithm", but does not provide specific version numbers for TensorFlow or any other software libraries or dependencies.
Experiment Setup Yes The neural network was trained using the Adam algorithm (Kingma & Ba, 2014) with a mini-batch size of 128 and a learning rate of 10 4. ... Training was run for 200 epochs.