Compared various deep learning software packages like Torch, Theano, Caffe on different metrics (accuracy, timing, memory usage, etc.) for different deep algorithms.

• Extending the work of Ghosh et al to global and multi-climate model data.
• Developing a parallel C (MPI) code that emulates the original MATLAB script.
• Next goal will be downloading the massive climate precipitation NetCDF data from the servers.

• Used deepnet package to learn a 3-layer Boltzmann machine for pretraining.
• Dataset to be used: pain, schizophrenia, cocaine addiction.

This research internship was fully supported by MITACS Globalink Fellowship Grant.

• Developing a cross platform app Edmonton Automated Sugar Intelligence (EASI) which recommends insulin dose based on the patient's input and settings set up by his/her diabetologist.

• Developed a tool which automatically adjusts the amount of insulin to be injected.

• Modified T1DMS Simulator to use it with 20 in-silico patients.

• Demonstrated that a supervised learning approach is better than certain reinforcement

  learning techniques like SARSA, Actor-Critic Model.

Sana is a standard-focused open-source end-to-end telemedicine system that facilitates the capture of medical data and physiological signals through a fully programmable work ow interface.

• Built a database that contains rapid search and a training set for (future) automated A.I. classification of images, audio and video.
• Constructed a web based protocol for Clinicians.
• Worked on integrating SANA with OpenMRS (a Medical Record System) for portability.

• Developing an intelligent mobile application that collects patient input through tests (Gait,Handwriting, Speech, Tremors) and provides interactive feedback to patients and monitoring summaries for physicians.

• Discovery of hidden relationships through unsupervised learning (clustering), leading to the identification of dierent Parkinson's disease variations.

• Automated ranking of different inputs (attributes) by measuring their information content.

• Used University of California, Irvine’s PD Telemonitoring dataset.

• Proposed a ’collaborative’ deep learning model which tightly couples convolutional neural networks and probabilistic matrix factorization.
• Better performance than Collaborative Filtering was shown using MovieLens dataset.

• Scrutinized linguistic content from Facebook, extracting keywords to analyze personality.
• Used a linear SVM for classifying sex and ridge regression for predicting age.

Designed a hybrid algorithm using Fuzzy Logic allowing user to incorporate information such as traffic levels, weather conditions and construction zones to find an optimal shortest distance between two points with the help of Djikstras Algorithm.