ST PETERSBURG, FLORIDA - Artificial intelligence (AI) has come far since its inception, constantly evolving and expanding its capabilities. A new approach, neurosymbolic AI, has emerged in recent years, whose aim is to bridge the gap between symbolic reasoning and deep learning (DL). Despite flying under the radar, this fusion of symbolic and connectionist models bids fair to unlock new levels of intelligence and understanding in AI systems.

Grasping symbolic reasoning, DL

To truly appreciate the significance of neurosymbolic AI, one must first understand the two paradigms it seeks to integrate - symbolic reasoning and DL.

Rooted in classical AI, symbolic reasoning entails manipulating abstract symbols and rules to perform logical reasoning and problem-solving. This school relies on explicit representations of knowledge and logic-based inference mechanisms, thus suiting it for tasks requiring symbolic manipulation and logical reasoning.

DL, by contrast, is inspired by the structure and function of the human brain and involves training artificial neural networks to learn patterns and representations from huge volumes of data. This approach has achieved remarkable success in image recognition, natural language processing, and speech recognition - achievements largely due to DL’s ability to automatically extract features and learn complex mappings from raw data.

Promise of neurosymbolic AI

While symbolic reasoning and DL have each demonstrat