# Bloom Filter Encoding for Machine Learning

arXiv: https://arxiv.org/abs/2512.19991 · HTML: https://mlsanity.com/paper/2512.19991/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 5 of its 27 references that are in the map.
- First posted on arXiv 23 Dec 2025.

## Builds on (most cited first, 5 of 5)

- [XGBoost: A Scalable Tree Boosting System](https://mlsanity.com/paper/1603.02754/) (2016)
- [Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms](https://mlsanity.com/paper/1708.07747/) (2017)
- [Sampling and Reconstruction Using Bloom Filters](https://mlsanity.com/paper/1701.03308/) (2017)
- [RobustBF: A High Accuracy and Memory Efficient 2D Bloom Filter](https://mlsanity.com/paper/2106.04365/) (2021)
- [Invertible Bloom Lookup Tables with Less Memory and Randomness](https://mlsanity.com/paper/2306.07583/) (2023)

## Built on by (most cited first, 2 of 2)

- [Privacy-Preserving Distributed Learning in IoT Systems: A Unified Threat Model and Evaluation Framework](https://mlsanity.com/paper/2605.09232/) (2026)
- [Entropy-Punctured Bloom Filters for Memory-Efficient Machine Learning](https://mlsanity.com/paper/2609.14187/) (2026)

## Abstract (arXiv)

We present a method that uses a Bloom filter transform to preprocess data for machine learning. Each sample is encoded into a compact bit-array representation using hash-based encoding, producing a fixed-length feature space that reduces memory usage and obfuscates original feature values. The encoding does not rely on keyed hashing; however, a key can optionally be used to control the mapping and would be required to reproduce the representation. We evaluate the approach on six datasets spanning text, time-series, tabular, and image domains: SMS Spam Collection, ECG200, Adult 50K, CDC Diabetes, MNIST, and Fashion MNIST. Four classifiers are considered: Extreme Gradient Boosting, Deep Neural Networks, Convolutional Neural Networks, and Logistic Regression. Results show that models trained on Bloom filter encodings achieve performance comparable to models trained on raw data or standard dimensionality reduction techniques across several datasets, while providing consistent memory savings. These findings suggest that Bloom filter encodings can serve as an efficient, general-purpose pre-processing representation that preserves useful similarity structure for learning tasks while providing a degree of data obfuscation.