Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Logical-Rule Autoencoder for Interpretable Recommendations

Jinhao Pan, Bowen Wei, Ziwei Zhu

Abstract

Most deep learning recommendation models operate as black boxes, relying on latent representations that obscure their decision process. This lack of intrinsic interpretability raises concerns in applications that require transparency and accountability. In this work, we propose a Logical-rule Interpretable Autoencoder (LIA) for collaborative filtering that is interpretable by design. LIA introduces a learnable logical rule layer in which each rule neuron is equipped with a gate parameter that automatically selects between AND and OR operators during training, enabling the model to discover diverse logical patterns directly from data. To support functional completeness without doubling the input dimensionality, LIA encodes negation through the sign of connection weights, providing a parameter-efficient mechanism for expressing both positive and negated item conditions within each rule. By learning explicit, human-readable reconstruction rules, LIA allows users to directly trace the decision process behind each recommendation. Extensive experiments show that our method achieves improved recommendation performance over traditional baselines while remaining fully interpretable. Code and data are available at https://github.com/weibowen555/LIA.

A Logical-Rule Autoencoder for Interpretable Recommendations

Abstract

Most deep learning recommendation models operate as black boxes, relying on latent representations that obscure their decision process. This lack of intrinsic interpretability raises concerns in applications that require transparency and accountability. In this work, we propose a Logical-rule Interpretable Autoencoder (LIA) for collaborative filtering that is interpretable by design. LIA introduces a learnable logical rule layer in which each rule neuron is equipped with a gate parameter that automatically selects between AND and OR operators during training, enabling the model to discover diverse logical patterns directly from data. To support functional completeness without doubling the input dimensionality, LIA encodes negation through the sign of connection weights, providing a parameter-efficient mechanism for expressing both positive and negated item conditions within each rule. By learning explicit, human-readable reconstruction rules, LIA allows users to directly trace the decision process behind each recommendation. Extensive experiments show that our method achieves improved recommendation performance over traditional baselines while remaining fully interpretable. Code and data are available at https://github.com/weibowen555/LIA.

Paper Structure

This paper contains 23 sections, 4 equations, 2 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Methodology
  3. Problem Formulation
  4. Overall Structure
  5. Learnable Logical Rule Layer
  6. Negation via Signed Weights
  7. Logical Activation Functions
  8. Operator Selection
  9. Disjunctive Aggregation Layer
  10. Training
  11. Rule Extraction and Interpretability.
  12. Experiment
  13. Setup
  14. Data and Metric
  15. Baselines
  16. ...and 8 more sections

Figures (2)

  • Figure 1: Overview of LIA. Left: Architecture with signed-weight negation, learnable logical rules with operator selection, and disjunctive aggregation. Right: Interpretation for recommending Item #1042, with rules visualized as trees showing learned operators, selected items, and negations (red). Natural language explanations are directly derived from the rule structure.
  • Figure 2: Effect of the number of rules $K$ on ML1M.