Extensions to Generalized Annotated Logic and an Equivalent Neural Architecture
This work addresses the problem of improving explainability and modularity in AI systems for researchers and practitioners in neuro-symbolic AI, though it appears incremental as it builds on existing hybrid approaches.
The paper tackles the limitations of deep neural networks, such as lack of explainability and difficulty incorporating prior knowledge, by proposing an extension to generalized annotated logic that creates an equivalent neural architecture as a neuro-symbolic hybrid, using discrete optimization in a binarized neural network instead of continuous optimization.
While deep neural networks have led to major advances in image recognition, language translation, data mining, and game playing, there are well-known limits to the paradigm such as lack of explainability, difficulty of incorporating prior knowledge, and modularity. Neuro symbolic hybrid systems have recently emerged as a straightforward way to extend deep neural networks by incorporating ideas from symbolic reasoning such as computational logic. In this paper, we propose a list desirable criteria for neuro symbolic systems and examine how some of the existing approaches address these criteria. We then propose an extension to generalized annotated logic that allows for the creation of an equivalent neural architecture comprising an alternate neuro symbolic hybrid. However, unlike previous approaches that rely on continuous optimization for the training process, our framework is designed as a binarized neural network that uses discrete optimization. We provide proofs of correctness and discuss several of the challenges that must be overcome to realize this framework in an implemented system.