# Rule Extraction in Machine Learning: Chat Incremental Pattern Constructor

arXiv: https://arxiv.org/abs/2208.00335 · HTML: https://mlsanity.com/paper/2208.00335/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 6 of its 26 references that are in the map.
- First posted on arXiv 31 Jul 2022.

## Builds on (most cited first, 6 of 6)

- [Towards A Rigorous Science of Interpretable Machine Learning](https://mlsanity.com/paper/1702.08608/) (2017)
- [Efficient Decompositional Rule Extraction for Deep Neural Networks](https://mlsanity.com/paper/2111.12628/) (2021)
- [MAIRE -- A Model-Agnostic Interpretable Rule Extraction Procedure for Explaining Classifiers](https://mlsanity.com/paper/2011.01506/) (2020)
- [Trepan Reloaded: A Knowledge-driven Approach to Explaining Artificial Neural Networks](https://mlsanity.com/paper/1906.08362/) (2019)
- [X-TREPAN: a multi class regression and adapted extraction of comprehensible decision tree in artificial neural networks](https://mlsanity.com/paper/1508.07551/) (2015)
- [PBRE: A Rule Extraction Method from Trained Neural Networks Designed for Smart Home Services](https://mlsanity.com/paper/2207.08814/) (2022)

## Abstract (arXiv)

Rule extraction is a central problem in interpretable machine learning because it seeks to convert opaque predictive behavior into human-readable symbolic structure. This paper presents Chat Incremental Pattern Constructor (ChatIPC), a lightweight incremental symbolic learning system that extracts ordered token-transition rules from text, enriches them with definition-based expansion, and constructs responses by similarity-guided candidate selection. The system may be viewed as a rule extractor operating over a token graph rather than a conventional classifier. I formalize the knowledge base, definition expansion, candidate scoring, repetition control, English-rule heuristics, and response construction mechanisms used by ChatIPC. I further situate the method within the literature on rule extraction, decision tree induction, association rules, interpretable machine learning, and sequence construction. The updated C++ code implementation of ChatIPC is also reviewed in detail: it parses an embedded dictionary, normalizes lexical keys, caches definition tokens and part-of-speech tags, computes Jaccard scores on bitsets, applies heuristic linguistic bonuses, and persists the knowledge base with a versioned binary format. The paper emphasizes mathematical formulation and algorithmic clarity, and it provides pseudocode for the learning, scoring, and construction algorithms.