Bit-Swap: Recursive Bits-Back Coding for Lossless Compression with Hierarchical Latent Variables

Authors: Friso Kingma, Pieter Abbeel, Jonathan Ho

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

Reproducibility Variable Result LLM Response
Research Type Experimental Through experiments we verify that Bit Swap results in lossless compression rates that are empirically superior to existing techniques. Our implementation is available at https:// github.com/fhkingma/bitswap.
Researcher Affiliation Academia 1University of California, Berkeley, California, USA.
Pseudocode Yes Algorithm 1 BB-ANS for lossless compression with hierarchical latent variables. Algorithm 2 Bit-Swap (ours) for lossless compression with hierarchical latent variables.
Open Source Code Yes Our implementation is available at https:// github.com/fhkingma/bitswap.
Open Datasets Yes To compare Bit-Swap against BB-ANS, we use the following image datasets: MNIST, CIFAR-10 and Image Net (32 32).
Dataset Splits No The paper mentions using "test data compression results" and ELBO values from trained models, but it does not explicitly provide specific train/validation/test split percentages, sample counts, or refer to predefined splits with citations for reproducibility beyond general dataset names.
Hardware Specification No The paper does not specify the exact hardware (e.g., GPU models, CPU types, memory) used for running the experiments.
Software Dependencies No The paper mentions implementing models in PyTorch but does not provide specific version numbers for PyTorch or any other software dependencies.
Experiment Setup Yes We train for 100 epochs with Adam Kingma & Ba (2015) optimizer and a batch size of 64.