Intelligent Systems Lab
ISL research is focused on learning high quality models from data. Unlabeled data is modeled with clustering algorithms. Mixtures of labeled an unlabeled data are addressed with semi-supervised learning approaches. Of particular interest is big data for which Deep Neural Networks are often useful. Ensembles of different (and the same) types of models are considered, imbalanced data is continually addressed. Imprecision in intelligent systems is also a research topic. Some recent work focuses on learning prognostic models from medical images and clinical data, learning models of activity in very large information networks and clustering data in a network environment. An overall goal is to be able to group large sets of unlabeled data in useful ways, uncovering small, but important groups where they exist. Another goal is to be able to make accurate predictions from potentially large (at least partially) labeled data sets.
Some recent paper titles (if you do not find these or others of interest through Google, please ask).
- Rahul Paul, Lawrence Hall, Dmitry Goldgof , Matthew Schabath, and Robert Gillies, Predicting Nodule Malignancy using a CNN Ensemble Approach, IJCNN 2018.
M. Zhou, J. Scott, B. Chaudhury, L.O. Hall, D. Goldgof, K.W. Yeom, M. Iv, Y. Ou, J.
Kalpathy-Cramer, S. Napel, R. Gillies, O. Gevaert, R. Gatenby, Radiomics in Brain Tumor: Image
Assessment, Quantitative Feature Descriptors and Machine-learning Approaches, American Journal of Neuroradiology, American Journal of Neuroradiology Oct 2017,
Renhao Liu, Lawrence O. Hall, Kevin W. Bowyer, Dmitry B. Goldgof, Robert Gatenby, and Kaoutar Ben Ahmed, Synthetic Minority Image Over-sampling Technique: How to Improve AUC for Glioblastoma Patient Survival Prediction, IEEE International Conference on SMC, Oct. 2017.
Lab Director: Lawrence O. Hall