02-04-2011, 10:59 AM
[attachment=11540]
Architecture
(US-JDL/UK-TFDF)
Feature Space
(Data representations, Task-specific, feedback)
Dimensionality
(Communication bandwidth constraints,High [ Low, increase SnR
Uncertainty
Data, sensor, communication noise, high level ignorance, model uncertainty
`soft’ decisions – Bayesian inference framework… but ….
Incorrect use of independence between models [ Veto Effect
Inaccurate estimation of probabilities can lead to severe distortion of decisions (product rule dominated by low probability errors)
Simpler decision methods more robust
Dynamics
Fusion is an iterative dynamical process - Continually refining estimates, representations ..
How do constraints on communication bandwidth and processing limit architectures for fusion?
How does the brain create and modify its data representation?
How does the brain encode time, dynamics and use feedback?
How does the brain encode and process probabilities and uncertain knowledge?
Apart from very low level (cellular/subcellular) and very high level binding, the brain appears to leave data sources fragmented. Why?
(interesting clinical exception in synaesthesia! – do we learn ICA?)