Machine Learning TRIZ Cases
- Anomaly Detection Event Tree for Computational Efficiency
Deterministic n-way splitting in event trees reduces computational complexity while maintaining detection coverage.
- Reinforcement Learning Agent Intent Encoding via CMDP and Automata
A reinforcement learning agent uses a companion Markov decision process to encode environmental states for action selection.
- Time Series Prediction Model Using Derivative Segmentation
Segmenting time-series data intervals prevents error accumulation in long-term physical phenomenon predictions while managing variance-covariance interactions.
- Vehicle Interaction Machine Learning Model for Collision Avoidance
A vehicle interaction machine learning model predicts future interactions and transmits alerts to prevent mid-air collisions.
- Session Continuity via User Engagement Detection
Detecting user engagement through sensor data prevents premature session expiration while maintaining authentication security.
- ROC Curve Area Minimization for Stable Two-Class Classification
Learning apparatus minimizes area over ROC curve to avoid local optima and ensure stable high accuracy in two-class classification.
- Iterative Learning Control for LiDAR Memory and Convergence
Piecewise functions approximate error tables in iterative learning controllers, reducing memory requirements and accelerating convergence for LiDAR systems.