Where-What Networks (WWNs)
WWNs are general purpose neuromorphic networks to model the brain's dorsal ("where") and ventral ("what") pathways, aiming at engineering solutions to interactive visual recognition and attention (e.g., WWN-1, WWN-2, WWN-3).

Language Acquisition Model
The model takes inspiration from neuromorphic networks like MILN, WWN-1, WWN-2 and WWN-3 to create a language learner that is able to use generalization to create new sentences. Read more in the paper here. 

Multilayer In-place Learning Network (MILN)
MINL is a biologically inspired general-purpose regression network (See Dr. Luciw's work for details). A description of the code and a tutorial is available here. A key feature in MILN is called Topographic Class Grouping (TCG), indicating that top-down connections from a higher cortical area to an earlier area enable development of discriminative features. Read more details about Topographic Class Grouping via Discriminative Sparse Coding.

Lobe Component Analysis (LCA)
LCA is a biologically inspired model for a layer of neurons to develop into feature detectors. A description of the code and a brief tutorial is available here.

Incremental Hierarchical Discriminant Regression (IHDR)
IHDR is a general-purpose regression network using a tree structure. Read more about IHDR here. Detailed information about this classification and regression method is available here (pdf file).

Candid Covariance-free Principal Components Analysis (CCIPCA)
CCIPCA is a network to derive principal components incrementally, very fast, and without using a covariance matrix. Read more in the paper here.  A proof of convergence is here.


Download the code to the above models here.