Ongoing Research Work

• Unsupervised Clustering:
• Developing automatic clustering techniques for the noisy high-dimensional datasets.
• Neural Architecture Search:
• Developing effective optimization techniques to search best architecture of neural network for given application.
• Hyper-parameter Optimization:
• Developing effective optimization techniques to search optimal parameters of neural network or other models for given application.
• Adversarial Attack Strategies:
• Formulating optimization problem to find the most sensitive pixels of the images those mislead the CNN models.
• Large Scale Optimization:
• Developing new metaheuristics to handle large number of interrelated variables.
• Improving existing metahuristics.
• Developing variable decomposition methods to group correlated variables.

• Constrained Optimization:
• Developing new metaheuristics to handle low volume of feasible region.
• Improving existing metahuristics.
• Developing constraint handling techniques.
• Developing repair techniques.

• Unconstrained Optimization:
• Developing new search strategies for existing algorithms, especially DE, CMAES, and GA.
• Developing local search strategies.
• Developing parameter adaptation techniques to automatically tune parameters.

• Multi- and Many-Optimization:
• Developing new search strategies for handling multiple objectives.
• Developing new selection schemes to deal with irregular Pareto fronts.
• Developing diversity enhancement schemes.