fast & light
Boon nano, the fastest and lightest unsupervised ML algorithm available.
enabling solutions previously not possible.
Universal:
Boon Logic has differentiated itself with its patented unsupervised machine learning algorithm available in the Cloud, On-prem, and Far-edge chip deployment. We enable low power and memory requirements operating orders of magnitude faster than K-Means or other approaches found in the world today.
Autonomously Learn:
Boon Logic’s ML technology autonomously learns its environment and performs millions of inferences in seconds to capture anomalous activity and behavior, identify degrees of similarity and dimensionality within unstructured data, and predict future outcomes.
Cloud Independent:
Boon Logic’s unsupervised machine learning technology can operate in real-time without an uplink to cloud resources or prior hands-on training. Since Boon Logic’s ML is unsupervised, its models are constructed autonomously, thereby making it possible to deploy thousands to millions of automatically-generated, individualized ML models without the need for training sets created by domain experts.
why our technology
K-Means vs Boon nano
Boon nano processing time is unaffected by the number of patterns
inference time relative to number of clusters:
far-edge deployments
Deploy the Boon Nano on the ARM Cortex M7 where it generates 1000 inferences per second across thousands of clusters
Core technology features
Pattern matching & searching
Determine whether two n-space vectors are within a certain distance of each other.
Segmentation
Create clusters of similar patterns from unstructured data.
Anomaly Detection
Determine patterns within unstructured data that are significantly different in either occurrence or distance from the rest of the data.
Prediction
If unstructured data is arranged as a time series, then predict likely outcomes based on ML-defined historical precedent.
