SmartTensors
AI/ML Framework • 2021 • Active
SmartTensors is a framework for science-informed ML/AI using methods such as NMFk and NTFk for feature extraction, model diagnostics, and robust analyses of complex datasets.
SmartTensors is a general framework for unsupervised, supervised, and physics-informed machine learning using nonnegative matrix and tensor decomposition algorithms.
The framework includes NMFk and NTFk methods for extracting hidden features, spatial footprints, and temporal signatures from complex datasets with minimal prior assumptions.
SmartTensors algorithms are written in Julia and can be integrated with Flux.jl, TensorFlow, PyTorch, MXNet, and MATLAB.
SmartTensors has been supported for commercial deployment through the DOE Technology Commercialization Fund.
Capabilities
- NMFk and NTFk for nonnegative matrix and tensor factorization
- Feature extraction, anomaly detection, image recognition, and model diagnostics
- Designed for science-informed AI and reduced-order modeling workflows
Images
