Pre-Training Meta-Rule Selection Policy for Visual Generative Abductive Learning

2025-03-09

AI Summary

In visual generative abductive learning, large logical symbol sets and complex rules severely hinder logical inference efficiency. To address this, we propose a meta-rule selection-based pretraining framework. Our key contributions are: (1) the first introduction of a pure symbolic data pretraining strategy for meta-rule selection, decoupling symbolic reasoning from visual grounding and substantially reducing computational overhead; (2) modeling logical structure via symbolic embeddings and attention mechanisms, leveraging their memorization capacity to automatically correct grounding errors for unseen symbols; and (3) designing a neuro-symbolic collaborative architecture that drastically compresses the meta-rule search space without requiring image data. Evaluated on the AbdGen benchmark, our method achieves significant improvements in inference efficiency while maintaining accuracy. The source code is publicly released to ensure reproducibility.

Technology Category

• Computer Vision: Visual Reasoning & Symbolic Representations
• Machine Learning: Neuro-Symbolic Learning
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Visual generative abductive learning studies jointly training symbol-grounded neural visual generator and inducing logic rules from data, such that after learning, the visual generation process is guided by the induced logic rules. A major challenge for this task is to reduce the time cost of logic abduction during learning, an essential step when the logic symbol set is large and the logic rule to induce is complicated. To address this challenge, we propose a pre-training method for obtaining meta-rule selection policy for the recently proposed visual generative learning approach AbdGen [Peng et al., 2023], aiming at significantly reducing the candidate meta-rule set and pruning the search space. The selection model is built based on the embedding representation of both symbol grounding of cases and meta-rules, which can be effectively integrated with both neural model and logic reasoning system. The pre-training process is done on pure symbol data, not involving symbol grounding learning of raw visual inputs, making the entire learning process low-cost. An additional interesting observation is that the selection policy can rectify symbol grounding errors unseen during pre-training, which is resulted from the memorization ability of attention mechanism and the relative stability of symbolic patterns. Experimental results show that our method is able to effectively address the meta-rule selection problem for visual abduction, boosting the efficiency of visual generative abductive learning. Code is available at https://github.com/future-item/metarule-select.

Problem

• Reduces time cost of logic abduction in visual generative learning.
• Proposes pre-training method to prune meta-rule search space.
• Enhances efficiency of symbol-grounded neural visual generation.

Innovation

• Pre-training meta-rule selection policy for efficiency
• Embedding-based symbol and meta-rule representation
• Low-cost learning with pure symbol data