Deriving Equivalent Symbol-Based Decision Models from Feedforward Neural Networks

2025-04-16

AI Summary

To address the trust bottleneck arising from poor interpretability of feedforward neural networks (FNNs), this paper proposes a neuro-symbolic bridging method that reconstructs black-box FNNs into functionally equivalent symbolic decision trees via layer-wise inversion of neuron activation trajectories. Our approach introduces, for the first time, a distributed representation parsing mechanism enabling iterative subpath refinement in deep networks while strictly preserving functional equivalence. Technically, it integrates Keras/TensorFlow model introspection, neural activation path tracing, and role-filler structure identification, implemented in a cross-platform prototype built on JavaFX. Experimental evaluation confirms the feasibility of exact FNN-to-decision-tree translation, substantially enhancing model transparency, trustworthiness, and auditability. This work establishes a verifiable, interpretable paradigm for high-stakes AI applications.

Technology Category

• Machine Learning: Neuro-Symbolic Learning
• Computer Vision: Visual Reasoning & Symbolic Representations
• Cognitive Modeling & Cognitive Systems: Symbolic Representations

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Artificial intelligence (AI) has emerged as a transformative force across industries, driven by advances in deep learning and natural language processing, and fueled by large-scale data and computing resources. Despite its rapid adoption, the opacity of AI systems poses significant challenges to trust and acceptance. This work explores the intersection of connectionist and symbolic approaches to artificial intelligence, focusing on the derivation of interpretable symbolic models, such as decision trees, from feedforward neural networks (FNNs). Decision trees provide a transparent framework for elucidating the operations of neural networks while preserving their functionality. The derivation is presented in a step-by-step approach and illustrated with several examples. A systematic methodology is proposed to bridge neural and symbolic paradigms by exploiting distributed representations in FNNs to identify symbolic components, including fillers, roles, and their interrelationships. The process traces neuron activation values and input configurations across network layers, mapping activations and their underlying inputs to decision tree edges. The resulting symbolic structures effectively capture FNN decision processes and enable scalability to deeper networks through iterative refinement of subpaths for each hidden layer. To validate the theoretical framework, a prototype was developed using Keras .h5-data and emulating TensorFlow within the Java JDK/JavaFX environment. This prototype demonstrates the feasibility of extracting symbolic representations from neural networks, enhancing trust in AI systems, and promoting accountability.

Problem

• Deriving interpretable symbolic models from neural networks
• Bridging neural and symbolic AI paradigms for transparency
• Enhancing trust in AI via symbolic decision tree extraction

Innovation

• Derive symbolic models from neural networks
• Map neural activations to decision trees
• Use distributed representations for symbolic components