Adaptive Tactics Sapient Human Intelligence Platform (AI-SHIP)
AI-SHIP is a Adaptive Multi-Source Intelligence Fusion solution methods & software framework that support implementation of the Four Machine Intelligence (MI) Waves to provide:
- Massive Data Summarization & Compression without mission relevant information loss
- Normalized Data Ontology Generation & Persistence in Hybrid Cloud with support for disconnected domains (e.g., Navy ships)
- Adaptive Models & Algorithms with Introspection to Adapt to Contextual Change
- Visualization of Hierarchical Summary Data with Explainable Predictions to Earn Operator Trust
- Quantified Accuracy, Statistical Significance, &Confidence with Pedigree
AI-SHIP implements the five main functional areas of the intelligence processing pipeline:
- Adaptive Data Preparation & Analysis (ADPA)
- Multi-source raw observation data aggregation, normalization, alignment, and persistence in an analytics-friendly data store.
- Adaptive Model Generation (AMG)
- Composable abstract model generation and selection in time, space, and frequency to enable adaptation by measurement type, observability, and capability. Includes model introspection to detect changes in context that affect model attribute statistics.
- Adaptive Composable Learning (ACL)
- Composable abstract algorithm generation and selection for learning and introspection that enables algorithms that adapt to contextual change to maximize prediction confidence and accuracy over time.
- Adaptive Course of Action (ACOA) Management
- Rank-ordered source, learning, and mission action options to maximize performance against mission outcome metrics that are constrained by actor possibility and capacity limits.
- Human Feedback & Adaptation (HFA)
- Decision options for human or machine approval to optimize model, algorithm, and mission outcomes by learning from historical predictions against selected outcome labels.