States and load
Cognitive State Model and Cognitive Load Model: classifying cognitive states and estimating mental workload.
SPARK · Cognitive AI · Cognitive AI models
SPARK does not rely on one generic model. Each question gets its own specialised model, trained on synchronised cognitive data and validated on people it has never seen.
Cognitive State Model and Cognitive Load Model: classifying cognitive states and estimating mental workload.
Fatigue Prediction Model and Performance Correlation Model: anticipating fatigue and relating cognition to performance.
Cognitive Digital Twin and Cognitive Optimization Model: modelling each person and recommending training.
A specialised model answers one question and can be checked on it. The cognitive features it relies on stay readable.
Correlation is not causation: each model is tested on what it really predicts, on sessions and people kept apart from training.
One question, one model, one validation.
The models run on the sessions kept in the cloud, from the raw signal and the extracted features, and their results appear in the debrief next to the measured facts.
A team wants to know when a driver's performance is about to degrade.
A fatigue model and a performance correlation model are applied to the synchronised sessions.
The result is checked on sessions and people the model has never seen before anyone relies on it.
Because each question (states, load, fatigue, performance, training) needs its own model, trained and validated for it.
On people and sessions they have never seen, so that a result does not only reflect the training data.
No: they find relations in the data. Correlation is not causation, and every result stays traceable to the measured signal.
Cognitive AI models, Cognitive Digital Twin and Cognitive Foundation Models: modelling how human cognition works.
Tell us what you do. Someone on the team answers you personally.
Talk to us