About me

I am pursuing my Masters degree in Computer Science at the University of British Columbia. Previously, I completed my undergraduate education in Computer Science and Engineering from the Indian Institute of Technology Guwahati in 2016.

I am a part of the Computer Vision Lab at UBC, working under the guidance of Prof. Leonid Sigal. Previously, I worked as a Research Assistant under the guidance of Prof. Partha Talukdar at the Machine and Language Learning (MaLL) Lab, which is located in the Indian Institute of Science.

I am interested in Machine Learning, especially its application in Computer Vision and Natural Language Processing.

I am 24 years old. I was born in the city of Aurangabad in India. I lived in Guwahati for four years while I was persuing my undergraduate education. Currently I stay in the beautiful city of Vancouver.

I enjoy playing basketball, eating, watching movies, programming and playing videogames. I recently have developed a keen interest in Fantasy Sports.


  • Research
  • Jan 18 - Apr 18
Enhanced Visual Dialog
We looked at the task of cooperative visual dialog, wherein two agents have a natural language conversation about a particular image. Contrary to previous work, we proposed an end-to-end differentiable deep neural framework with multi-modal attention for this task.
  • Research
  • Jan 18 - Apr 18
Toxic Comment Classification
We analyzed the performance of word & sentence attention mechanisms on the task of toxic comment classification. Through extensive experimentation, we demonstrated the effectiveness of deep hierarchical attention models in identifying abusive text.
  • Research
  • Sep 17 - Dec 17
Multi-Armed Bandits for Influence Maximization
We proposed a bandits framework to learn probabilities for influence maximization in Independent Cascade and Topic-Aware Independent Cascade models. We used a variant of Online Expectation Maximization algorithm to scale the framework to larger networks.
  • Research
  • Jun 16 - Jun 17
Improving Distantly Supervised Relation Extraction using Word and Entity Based Attention
We proposed two novel word attention models for distantly supervised relation extraction: (1) a Bi-directional Gated Recurrent Unit (Bi-GRU) based word attention model (BGWA), (2) an entity-centric attention model (EA). We also introduced GDS, a new distant supervision dataset for relation extraction. GDS removes test data noise present in all previous distant supervision benchmark datasets, making credible automatic evaluation possible. Through extensive experiments on multiple real-world datasets, we demonstrated the effectiveness of the proposed methods.
  • Improving Distantly Supervised Relation Extraction using Word and Entity Based Attention
This work was presented at the AKBC workshop at NIPS 2017.
  • Research
  • Aug 15 - Apr 16
Faster Cluster Estimation for K-Means Algorithm
My Bachelors' Thesis Project at IIT Guwahati was persued under the guidance of Prof. Amit Awekar. This work aims at reasonably estimating the final cluster centroids for K-means algorithm in lesser amount of time. We propose a heuristic that attempts to reduce the number of point to centroid distance and bound computations.
  • Faster Cluster Estimation for K-Means Algorithm
This work was published at the European Conference on Information Retrieval 2017, held at Aberdeen, Scotland UK.
  • Mar 16 - Apr 16
Classification and Analysis of Medical Data using Machine Learning
An extensive study on two medical data sets, namely Diabetes and Breast Cancer, to identify the features that play an important role in identifying the class of patients that need medical attention. This work focuses on developing high recall classifiers to predict patient re-admission using their medical history. We also used Apriori algorithm to identify the association rules in the data sets.
  • Mar 16 - Apr 16
Question Answering System for Text Document
Developed an answering system for specific questions, which also includes questions that require a certain level of semantic deduction, asked by the user. The input for the system consists of a pure text document and the knowledge domain is limited to this input. The system leverages the keywords in the input text and the question, and a graph created using the dependency relations in the input sentences and the question, to find the answers.
  • Research
  • May 15 - Jul 15
Measuring Data Sources based on Sufficiency and Relevancy
This work was done at Adobe Research under the guidance of Dr. Niyati Chhaya. We came up with a novel ranking method for data tables based on their relevance and a new concept of sufficiency for a given query. Our goal was to help journalists in the field of data driven journalism by providing them with complete and precise information, which would allow them to analyse causal factors and effects of various events efficiently. We came up with the correlated topic space for the query using query expansion on a learned weighted domain graph. Our contribution was coming up with a new metric which, for each data source, measured its coverage of the query topic space. This metric, called sufficiency, was used in addition to query-document relevance for ranking.
  • Research
  • May 15 - Jul 15
Measuring Source Quality for Data Sources
This work was done at Adobe Research under the guidance of Dr. Niyati Chhaya. The project tried to define parameters for ranking data sources on the basis on their quality. We defined new parameters for measuring the quality of a particular source and formulated an empirical definition of the source quality.
  • Oct 15 - Nov 15
Serial, Parallel and Distributed Implementation of Hierarchical Clustering
I implemented the serial, shared memory parallel and distributed versions of single link hierarchical clustering in Python.
Read More
  • Jan 15 - Mar 15
C++ To MIPS Compiler
I developed a compiler for a language similar to C++ that could parse basic codes containing recursion, loops and conditional statements. It converted the C++-similar code to MIPS code, which could then be run on a MIPS simulator
Read More
  • Research
  • May 14 - Present
Classifying Twitter Idioms and Predicting their Popularity
This work was done at CNeRG, IIT Kharagpur under the guidance of Prof. Animesh Mukherjee. The project was aimed at understanding the reason for the popularity of a particular type of hashtags called Idioms, classifying them from other hashtags and predicting their popularity after a few months from their first appearance. The motivation behind studying Idioms, which are a concatenation of atleast two words that have some underlying theme associated with them (like #IWishICould, #10ThingsAboutMe), was to understand the opinions and sentiments of people.
  • Oct 14 - Nov 14
Smart Rack
The objective was to convert a simple rack into a smart rack that had features like Counting the number of objects kept inside the rack, issuing a warning when there is a shortage of objects or when an object has expired or as a reminder to the user to take out a particular object, ensuring that the rack cannot be accessed by everyone, etc.
Read More
  • Aug 14
IITG MUN Website
Designed and developed the official website for the second edition of the IIT Guwahati Modal United Nations Conference to be held at Techniche, the annual Techno-management festival of IIT Guwahati. The website was made using HTML, CSS and simple JavaScript.
Read More
  • Feb 14 - Mar 14
Assembler, Linker, Loader for 8085 Microprocessor
This project aims at converting simple assembly language code defined on an instruction set into 8085 assembly code, linking different files and their variables and loading it in an appropriate location in the memory. A user friendly graphic user interface facilitates analyzing each step to understand the changes that are made to the input code. The interface has an inbuilt text editor that allows editing the code before assembling.
Read More


University of British Columbia
Master of Science (M.Sc)
Computer Science
Sept 17 - Present
Indian Institute of Technology Guwahati
Bachelor of Technology (B.Tech)
Computer Science and Engineering
CGPA: 8.65

Jul 12 - Jun 16
Saint Johns Senior Secondary School
Senior Secondary
Percentage (Class XII Examination): 94.2%

Jun 10 - Apr 12
Nath Valley School
Secondary and Primary
Percentage (Class X Examination): 95.6%

Jun 00 - Apr 10
Work Experience
MaLL Lab, Indian Institute of Science
Research Assistant

I worked as a Research Assistant under the guidance of Prof. Partha Talukdar in the area of Temporal Information Processing, Relation Extraction and Dialogue Generation.

Jun 16 - Jun 17
Adobe Research, India
Research Intern

I interned at Adobe Research India at Bengaluru under the guidance of Dr. Niyati Chhaya. I worked on novel projects in the field of Data Journalism for a period of three months.

May 15 - Jul 15
CNeRG, IIT Kharagpur
Research Intern

I interned at the CNeRG lab at IIT Kharagpur. I worked on a project that aimed at understanding Information Propagation And Opinion Formation among people through Twitter Idioms.

May 14 - Jul 14
Techniche, IIT Guwahati
Head, Industrial Conclave and Literary Module

I was the head of the Industrial Conclave and Literary Module and a core team memeber of Techniche-2014, the annual Techno – management festival of IIT Guwahati. We invited eminent personalities from the industrial world and organized different literary events like Spelling Bee, Parliamentary Debate and Model United Nations (MUN) conference. I was also incharge of the infrastructural requirements during the festival that witnessed a footfall of around 30,000.

Sep 13 - Sep 14
Qualified for the ACM International Collegiate Programming Contest (ICPC) Amritapuri Onsite Regionals.
Qualified IIT-JEE among the top 0.24% out of 5 lakh candidates.
Cleared ISAT (Entrance exam for Indian Institute of Space Science and Technology) with an All India Rank of 425.
Qualified for the award of Scholarship of Higher Education (SHE) by virtue of performance within top 1% of the Class XII CBSE Examination under the INSPIRE Scheme by the Central Board of Secondary Education.
Recommended for the KVPY (Kishore Vaigyanik Protsahan Yojana) Fellowship.
Qualified for the Indian National Mathematics Olympiad (INMO).


If you like my work or you have a project or an idea you’d like to discuss, please feel free to contact me.

You can reach me at : siddhesh166@gmail.com or skhandel@cs.ubc.ca
This website is designed and developed by Siddhesh Khandelwal