{{scGetNextMDInteraction.data.NextMDInteraction.interaction_description}}

{{scGetMDFolderInfo.data.MDFolderInfo.folderName}}

{{scGetMDFolderInfo.data.MDFolderInfo.objectsCount}}

Evolution of a 2nd generation hybrid neuro-symbolic AI

1990 - 1994: Early research on hybrid neuro-symbolic AI systems

In the early 1990's, the complementary benefits of combining neural network "soft" capabilities with the "hard" constraints of rule-based expert systems were being explored in both government research and commercial product activities. James M. Mazzu (CEO of Digie Inc.) and Dr. Alper K. Caglayan (Advisor to Digie Inc.) conducted extensive research illuminating the benefits of hybrid neuro-symbolic AI systems, including early applications in smart aircraft structures, nuclear plant monitoring, target recognition, remote sensing, and a neural/expert architecture for task allocation.

1992: NueX neural network/expert system tool

Their work initially resulted in the first commercially available tool for creating hybrid neural/expert AI systems, called NueX. The goal was to create a tool that combined the benefits of neural networks and expert knowledge-based systems to produce intelligent systems modeled after the naturally holistic use of both the right and left hemispheres of the human brain (the NueX graphic is a symbolic representation of the two hemispheres coming together).

1993: Open Sesame! 1st generation hybrid neuro-symbolic AI

Released in 1993, Open Sesame! was the world's first hybrid learning agent for the Apple Macintosh, critically acclaimed and box-shipped to over 35,000 users. It used neural networks to identify user behavior patterns and symbolic expert systems to interpret the patterns and generate personalized rules to automate tasks, making it a first-generation hybrid neuro-symbolic AI system.

1996: Learn Sesame intelligence engine 

Based on extensive user feedback from Open Sesame! 1.0 and deployment challenges from Open Sesame! 2.0, lessons were learned that led to the development of the Learn Sesame intelligence engine, which utilized statistical event clustering over traditional neural network methods.

2017: Digie multi-agent conversational symbolic AI

Independently continuing his doctoral research, James M. Mazzu created a novel self-organizing architecture for symbolic knowledge representation and reasoning (KRR) that captures knowledge by semantic understanding. The resulting multi-agent platform, named Digie, builds user knowledge and insights through natural conversation, applies expert insights/advice based on personal situations, and manages the sharing of knowledge between trusted friends. Digie enables a network of personal AI agents, and was initially applied to the travel domain.

2023: Digie+GPT 2nd generation hybrid neuro-symbolic AI

With the arrival of highly effective LLM systems such as GPT, Digie's symbolic AI now leverages complementary neural network capabilities for personal knowledge capture and multi-agent domain knowledge access, making it a 2nd generation hybrid neuro-symbolic AI system. Digie's unique hybrid multi-agent capabilities provide a truly personal AI that enables specialized AI channels from faith-based dialogues to crafting personal travel experiences.