Topological Clustering


Kinect Box:

Original GNG

GNG-CIM (Num. of Nodes)

GNG-CIM (Node Distribution)

Kernel Bayesian ART /ARTMAP /Associative Memory


  • Unsupervised / supervised / semi-supervised / Associative Memory
  • Dimension reduction
  • Incremental and online Classification
  • Robust to noise & Outliers
  • Information Theoretic approach – Correntropy Induced Metric
  • Solve stability & plasticity dilemma

Genetic Ensemble Fuzzy Extreme Learning Machine


  • Fast classification & regression
  • Fuzzy rules extraction
  • Genetic ensemble approach
  • Hybrid knowledge-based & data-driven approach
  • Automatic fuzzy membership generation, selection.
  • Automatic fuzzy rules selection
  • Automatic feature selection
  • Random projection based on hypernetworks

Topological Kernel Bayesian ART/ARTMAP


  • Unsupervised / supervised / semi-supervised / Associative Memory
  • Incremental and online Classification
  • Robust to noise & Outliers
  • Information Theoretic approach – Correntropy Induced Metric
  • Solve stability & plasticity dilemma

Online Recurrent Kernel - based Extreme Learning / Reservoir Machine

  • Incremental & deterministic hidden neuron insertion
  • Self-determined number of kernel/neuron
  • Multiple-time steps ahead prediction
  • Output recurrent feedback
  • Multivariate regression analysis
  • Online confidence interval estimation and outlier detection
  • Lateral recurrent loop to expand the time horizon

Biologically Inspired Vision


Quantum Inspired Computational Intelligence


  • Dynamic Quantum Clustering
  • Information theoretic Quantum Particle Swarm Optimization
  • Quantum Inspired Associative Memory