Faster-LTN: a neuro-symbolic, end-to-end object detection architecture

TL;DR

Faster-LTN introduces an end-to-end neuro-symbolic object detection architecture combining convolutional neural networks with Logic Tensor Networks, enabling the integration of prior knowledge and learning from labeled data.

Contribution

This is the first end-to-end training of a neuro-symbolic object detector combining CNNs and LTNs with logical axioms.

Findings

Competitive performance with Faster R-CNN

Effective integration of prior knowledge and labeled data

First end-to-end neuro-symbolic object detection approach

Abstract

The detection of semantic relationships between objects represented in an image is one of the fundamental challenges in image interpretation. Neural-Symbolic techniques, such as Logic Tensor Networks (LTNs), allow the combination of semantic knowledge representation and reasoning with the ability to efficiently learn from examples typical of neural networks. We here propose Faster-LTN, an object detector composed of a convolutional backbone and an LTN. To the best of our knowledge, this is the first attempt to combine both frameworks in an end-to-end training setting. This architecture is trained by optimizing a grounded theory which combines labelled examples with prior knowledge, in the form of logical axioms. Experimental comparisons show competitive performance with respect to the traditional Faster R-CNN architecture.