ExtensityAI Announces SymbolicAI Framework: A New Paradigm for Logic-Based AI

February 03, 2024

We are excited to announce the publication of our research paper introducing SymbolicAI, a versatile and modular framework employing a logic-based approach to concept learning and flow management in generative AI models. This framework represents a significant step forward in our mission to create AI systems that can truly understand, think, and reason.

The SymbolicAI Framework

Our paper, "SymbolicAI: A framework for logic-based approaches combining generative models and solvers", outlines a comprehensive approach to addressing one of the most significant challenges in modern AI: combining the pattern-recognition capabilities of deep learning with the logical reasoning strengths of symbolic AI.

The SymbolicAI framework enables:

• Logic-based concept learning that helps AI systems develop a deeper understanding of relationships between entities

• Flow management in generative models for more coherent and contextually appropriate outputs

• Integration of solvers that can tackle complex reasoning tasks that pure neural approaches struggle with

• A modular architecture that allows for flexible application across various domains and problems

Research Collaboration

This framework is the result of collaborative work led by Marius-Constantin Dinu, alongside co-authors including Markus Holzleitner, Daniel Wetzell, and Sepp Hochreiter. We extend our gratitude to colleagues and friends who provided valuable support and discussions, including Sergei Pereverzyev, Eric Squendor, Gary Marcus, Fabian Paischer, Martin Hofmann, Kevin Schweighofer, Clemens Wasner, and Andreas Windisch.

Community Recognition

We're honored by the positive reception our work has received from the AI research community. Our paper has been shared by several notable researchers and AI-focused accounts, highlighting the interest in neuro-symbolic approaches to artificial intelligence.

Next Steps and Vision

With the release of the SymbolicAI framework, we are working toward establishing an industry standard similar to PyTorch for AI-centric workflows. Our goal is to provide developers and researchers with tools that make it easier to build AI systems capable of human-like reasoning while maintaining the strengths of modern deep learning approaches.

We believe that symbolic AI represents the future of artificial intelligence—not as a replacement for neural approaches, but as a complementary paradigm that addresses fundamental limitations in today's AI systems.

Get Involved

We are actively seeking investors to provide long-term support for our work and help grow the SymbolicAI framework into an industry standard. If you're interested in collaborating with us or learning more about our technology:

• Visit our website at extensity.ai

• Follow us on Twitter/X @ExtensityAI

• Contact us directly through our website

We're excited about the potential of SymbolicAI to transform how AI systems reason and understand the world, and we look forward to sharing more developments in the near future.

The SymbolicAI paper is available on arXiv and HuggingFace.