Shikhar Sharma

About Me

I am a grad student at the University of Toronto where I am pursuing a Master's degree in Computer Science. I am working with Prof. Ruslan Salakhutdinov and the Machine Learning Group and my research interests are mainly Deep Learning, Machine Learning and Computer Vision.

I obtained my B.Tech. in Computer Science & Engineering from IIT Kanpur in May 2014. To know more about me you can visit my personal website.

Contact Details

Shikhar Sharma
Room 3302
10 King's College Road
Toronto, ON M5S 3G4

shikhar [at] cs [dot] toronto [dot] edu


University of Toronto

M.Sc. in Computer Science January 2016 Canada

IIT Kanpur

B.Tech. in Computer Science & Engineering May 2014 India

Select Projects

Speech Recognition

B.Tech. Project with Prof. Harish Karnick January 2014 - April 2014

Used Deep Belief Networks and Hidden Markov Models with MFCC and its first and second order derivatives as features to recognize alphabets from ISOLET dataset and words from an IIT Kanpur dataset of roughly 120,000 Hindi word recordings.

Movie Recommendation System

Course Project August 2013 - November 2013 Code and Report

Applied collaborative filtering techniques to generate recommendations and predict ratings on MovieLens dataset having 100K ratings from 943 users on 1682 movies.

Learning Context-Driven Trajectory Preferences via Sub-optimal Feedback

Cornell University

Research Intern May 2013 - July 2013 Project Page

Worked with Ashesh Jain and Prof. Ashutosh Saxena to make Baxter Research Robot learn trajectory preferences for grocery store checkout using incremental improvements suggested by the user. The project involved feature engineering, using algorithms like preference perceptrons and structural support vector machine.

Twitter Sentiment Analyzer

Yahoo! HackU 24-hr hackathon August 2012 Code and Documentation

Plotted twitter trends on Google Maps and did country-wise Sentiment Analysis on tweets. Plotted whether a product was trending positively/negatively based on the buzz generated by it on Twitter.

Logic Studio

IIT Kanpur in collaboration with Microsoft Research

Research Project April 2012 - July 2012

Worked with Prof. Amey Karkare, Prof. Subhajit Roy and Dr. Sumit Gulwani on creating an Intelligent Tutoring System for teaching First Order Logic. We utilized the Stanford CoreNLP Parser to parse and retrieve meta-data and generated first order logic formulae. We also explored generating new questions from our corpus.


Action Recognition and Video Description using Visual Attention

Shikhar Sharma Masters Thesis, University of Toronto 2016 [ PDF || BibTeX ]

Action Recognition using Visual Attention

Shikhar Sharma, Ryan Kiros and Ruslan Salakhutdinov International Conference on Learning Representations (ICLR) Workshop 2016 [ PDF || BibTeX || Project page ]

Action Recognition using Visual Attention

Shikhar Sharma, Ryan Kiros and Ruslan Salakhutdinov NIPS Workshop on Time Series 2015 [ PDF || BibTeX || Project page ]

Learning Preferences for Manipulation Tasks from Online Coactive Feedback

Ashesh Jain, Shikhar Sharma, Thorsten Joachims and Ashutosh Saxena To appear in International Journal of Robotics Research (IJRR) 2015 [ PDF || BibTeX || Project page || Video ]

Beyond Geometric Path Planning: Learning Context-Driven Trajectory Preferences via Sub-optimal Feedback

Ashesh Jain, Shikhar Sharma and Ashutosh Saxena International Symposium of Robotics Research (ISRR) 2013 [ PDF || BibTeX || Project page || Video ]

Get In Touch.

The postal address given here is the mailing address for all grad students at the University of Toronto. Therefore, do not forget to write my name too.