# Extracting PAC Decision Trees from Black Box Binary Classifiers: The Gender Bias Case Study on BERT-based Language Models

arXiv: https://arxiv.org/abs/2412.10513 · HTML: https://mlsanity.com/paper/2412.10513/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 9 of its 37 references that are in the map.
- First posted on arXiv 13 Dec 2024.

## Builds on (most cited first, 9 of 9)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://mlsanity.com/paper/1907.11692/) (2019)
- [Distilling a Neural Network Into a Soft Decision Tree](https://mlsanity.com/paper/1711.09784/) (2017)
- [Neural Networks are Decision Trees](https://mlsanity.com/paper/2210.05189/) (2022)
- [Learning Description Logic Ontologies. Five Approaches. Where Do They Stand?](https://mlsanity.com/paper/2104.01193/) (2021)
- [Exact Learning of Lightweight Description Logic Ontologies](https://mlsanity.com/paper/1709.07314/) (2017)
- [Trepan Reloaded: A Knowledge-driven Approach to Explaining Artificial Neural Networks](https://mlsanity.com/paper/1906.08362/) (2019)
- [Rule Extraction from Binary Neural Networks with Convolutional Rules for Model Validation](https://mlsanity.com/paper/2012.08459/) (2020)
- [X-TREPAN: a multi class regression and adapted extraction of comprehensible decision tree in artificial neural networks](https://mlsanity.com/paper/1508.07551/) (2015)

## Abstract (arXiv)

Decision trees are a popular machine learning method, known for their inherent explainability. In Explainable AI, decision trees can be used as surrogate models for complex black box AI models or as approximations of parts of such models. A key challenge of this approach is determining how accurately the extracted decision tree represents the original model and to what extent it can be trusted as an approximation of their behavior. In this work, we investigate the use of the Probably Approximately Correct (PAC) framework to provide a theoretical guarantee of fidelity for decision trees extracted from AI models. Based on theoretical results from the PAC framework, we adapt a decision tree algorithm to ensure a PAC guarantee under certain conditions. We focus on binary classification and conduct experiments where we extract decision trees from BERT-based language models with PAC guarantees. Our results indicate occupational gender bias in these models.