sumtyme.ai is an AI research startup focused on modelling the directional changes in time-varying complex systems.
We created a new AI architecture called Abstract Generalised Networks (AGNs). AGNs are specifically designed to learn structural abstractions from numerical time series data and model directional movement within time-varying complex systems. They achieve this by leveraging their pre-trained knowledge of how information propagates, transforms, and feeds back through a system, thereby generating observable temporal patterns.
What makes AGNs unique as a learning algorithm is their fundamental understanding of how information moves through time. This knowledge allows them to make inferences through dynamic exploration without hardcoded rules. Critically, this process requires no system-specific training data and eliminates the need for retraining post-deployment. As a result, AGNs can serve as an ideal foundational layer for building adaptive, intelligent systems, offering a significant advantage over traditional approaches that are reliant on extensive training data, system-specific feature engineering and continuous retraining.
Live Demo: Dashboard
Contact us: [email protected]