Funding
Funding not disclosed
Founders
Product
Problem
Traditional AI models often struggle with complex reasoning, knowledge integration across diverse domains, and the ability to adapt to new information, limiting their effectiveness in fields like scientific research, investment banking, and drug discovery. These models typically lack the ability to understand the underlying relationships between concepts, hindering their capacity for abstract thought and problem-solving.
Solution
New York General Group addresses these limitations with its "World System," an AI framework that combines Bidirectional Encoder Representations from Transformers (BERT), Categorical Networks (CN), and Point-Voxel Convolutional Neural Networks (PVCNN). The system uses CNs, based on category theory, as the "brain" of the AI, enabling it to engage in genuine cognitive processes and abstract reasoning. BERT provides advanced natural language understanding, while Point-Voxel CNN allows the AI to perceive and interact with the three-dimensional world. This integration enables a level of cognitive sophistication that is designed to be unprecedented in AI systems, allowing for complex problem-solving that requires both linguistic understanding and spatial reasoning.
Target Audience
The primary target audience includes researchers in artificial intelligence, biotechnology, and pharmaceuticals, as well as professionals in investment banking and policy-making who require advanced analytical and problem-solving capabilities.
Features
- Categorical Networks (CNs) enable the AI to model complex relationships and transformations between data objects with mathematical precision.
- Functorial semantics allow the AI to maintain consistent meaning across different domains and data types, facilitating knowledge transfer between contexts.
- Integration of BERT's contextual language understanding capabilities enables processing of natural language with deeper structural understanding.
- Point-Voxel CNN enables efficient processing of three-dimensional data, creating a unified system capable of reasoning across text, images, and spatial information.
- The system can generate novel propulsion concepts for rocket engineering by leveraging vast datasets encompassing fluid dynamics, materials science, and thermodynamics.
- The AI system can analyze astronomical data, formulate and test hypotheses about the fundamental principles governing our solar system.
- The AI system can design microscopic particles that can carry cancer drugs directly to tumors, increasing treatment effectiveness while reducing side effects on healthy tissues.