Our research projects

Additive Parts-Based Data Representation with Non-negative Sparse Autoencoders

Developing autoencoders with receptive fields that show additive contributions as features of 2D and 3D images

Humanizing Black Box Data Predictions with the Power of Explanations

Work with graduate students on recommender systems woth explanation facility

Building Virtual Community in Computational Intelligence and Machine Learning

Developing the CIML Community web portal

Computational Intelligence in Medical Diagnosis

Improving performance of case-based classifiers for breast cancer detection
Case-specific reliability assessment in medical diagnosis

Chaotic Dynamics of Coupled Oscillators

Data Mining Using Computational Intelligence Techniques

Expert systems for testing, diagnostics, and medicine
Fuzzy rule extraction resulting in linguistic descriptions of data relationships

Theory of Neural Networks

Optimization of neural network architectures through perturbation methods, pruning techniques, and improved learning/scaling approaches
Multivalued attractor-type associative memories

Neural Networks for Control, Identification, and Prediction

Inverse neural modeling of dynamic plants
Time series prediction
Prediction of drug dosage levels using neurocomputing
Development of neuro-fuzzy system to predict spine loading as a function of multiple dimensions of risk

Other Applications of Computational Intelligence Techniques

Identification of semiconductor manufacturing processes
Design centering and yield maximization in manufacturing