The Cheriton School of Computer Science offers enrichment opportunities to exceptional students through the Undergraduate Research Assistantships (URA) program to earn $600 a term for part-time research with a faculty member. Starting from F18, the School will top up $400 for students who meet the URA eligibility requirements and are accepted by the CS URA program. The money is paid once near the end of the term.
Application deadlines per term
- Winter applications — Last day of the month in January
- Spring applications — Last day of the month in May
- Fall applications — Last day of the month in September
We will not accept any applications once deadline has passed
Eligibility
Any student who has completed their second year in the Faculty of Mathematics with a cumulative average (CAV) of at least 80% is eligible. Preference is given to students enrolled in Computer Science major plans. A student can only do one URA per term. A student cannot be on a co-op term while doing a URA.
Note
Please refer to the Student Success Office for this information:
https://uwaterloo.ca/international-students/social-insurance-number
For an Engineering student to do a URA with a CS professor, please follow the application procedures explained on the Engineering URA web page: https://uwaterloo.ca/engineering/ura
- International Students need an SIN # to work in Canada
- Engineering Students need to follow Engineering URA procedures
How do I apply?
- Check eligibility stated above. Make sure you are eligible before taking the following steps.
- Find a project, send a resume and recent grade report to the professor, request an appointment, and reach a verbal agreement on the details of the project.
- Complete the Registration form - Once you complete the Registration form you will be directed to a webpage with the URA Application Form
- Complete the URA Application Form (PDF)
- Submit the form- Deadline one month after term starts
The URA Application Form must be fully completed and submitted to receive your URA payment.
A faculty member must sign and fill out the account number on the URA Application Form.
URA supervisors
Faculty members post openings each term so interested and eligible students can apply. Students can also contact faculty members to see if they are interested in supervising a URA.
The following faculty members have confirmed that they are willing to be contacted by potential URAs.
URA research projects
Title: HoloClean — A Machine Learning Framework for Data Quality and Enrichment
HoloClean is a statistical inference engine to impute, clean, and enrich data. As a weakly supervised machine learning system, HoloClean leverages available quality rules, value correlations, reference data, and multiple other signals to build a probabilistic model that accurately captures the data generation process, and uses the model in a variety of data curation tasks. HoloClean allows data practitioners and scientists to save the enormous time they spend in building piecemeal cleaning solutions, and instead effectively communicate their domain knowledge in a declarative way to enable accurate analytics, predictions, and insights from noisy, incomplete, and erroneous data.
Check out the details and the open source project at www.holoclean.io.
Please contact Ihab Ilyas if you are interested in this URA opportunity.
Please visit Ihab Ilyas's website.
Title: C∀ (Cforall) Programming Language and Runtime System
The C∀ project is an open-source project extending ISO C with modern safety and productivity features, while still ensuring backwards compatibility with C and its programmers. C∀ is designed to have an orthogonal feature-set based closely on the C programming paradigm (non-object-oriented) and these features can be added incrementally to an existing C code-base allowing programmers to learn C∀ on an as-needed basis. In many ways, C∀ is to C as Scala is to Java, providing a research vehicle for new typing and control-flow capabilities on top of a highly popular programming language allowing immediate dissemination. There are many small development and evaluation activities within the C∀ project suitable for URAs. A URA candidate should be interested in programming languages and associated runtime internals with experience in C/C++ programming.
If you are interested, please email Peter Buhr at: pabuhr@uwaterloo.ca
Title: Digital Watermarking of Machine Learning Models
Digital watermarks are secret messages embedded into cover data, such that only a secret key holder can retrieve the message. Digital watermarks can be used to prove ownership of the cover data, e.g., by embedding a message "This model belongs to the University of Waterloo". Digital watermarks are supposed to be imperceptible, i.e. an adversary with access to the modified cover data should not be able to tell whether a message was embedded, and robust, i.e., an adversary should not be able to remove the watermark from the modified cover data without destroying the utility of the cover data. In this work we aim to embed messages in machine learning models, such as neural networks. We investigate a) new attacks to detect or remove watermarks from existing algorithms, b) new watermarking algorithms that can withstand those attacks.
An ideal candidate has some experience with programming machine learning models. Good knowledge of statistics is helpful. You will learn how to run systematic experiments to evaluate security attacks.
If you are interested, please email Florian Kerschbaum at: florian.kerschbaum@uwaterloo.ca with a resume and an unofficial transcript.
Title: Expressive Human-Computer Interaction
The Exii research group led by Prof. Daniel Vogel does all kinds of exciting Human-Computer Interaction research on different input and interaction technologies. Everything from pen based computing, touchscreens, and mobile devices, to large displays, virtual reality, augmented reality, and spatial augmented reality. We typically build our prototypes using tools like 3D game engines (mostly Unity), web development, or the Processing language. Our projects often leverage a range of techniques and technologies like computer graphics (hit testing algorithms, transforms, rendering), applied computer vision (mostly OpenCV), applied statistics (usually in Jupyter Notebook with a mix of Python, Pandas, and R), 3D fabrication (3D printing, physical mockups), electronics prototyping (usually Arduino based), applied machine learning (e.g. deep learning, classifiers like SVM), and motion tracking (Vicon or AR markers). A single project typically uses only a subset of these, and we don’t expect URAs to necessarily have these skills already. This is a great opportunity to learn new techniques and technologies while you work on cutting edge Human-Computer Interaction research.
Interested students should view a list of current projects and submit an application:
https://goo.gl/forms/bjMkN0Rryjupry2V2
Title: Differentially Private Statistical Estimation
Differential privacy is the current gold standard for data privacy. Recent developments have resulted in theoretically principled algorithms for private statistical estimation in high-dimensional settings. The goal of this project is to develop these theoretical methods into realizable and practical algorithms, which are able to privately estimate parameters of high-dimensional distributions in both synthetic and real data settings.
The ideal URA will be very comfortable with probability and linear algebra, in addition to being confident in Python.
Interested students should email Gautaum Kamath: gckamath@uwaterloo.ca.
Title: Intelligent Connectivity for the Internet-of-Things
Although interest in connected devices has surged in recent years, barriers still remain in realizing the dream of the Internet of Things (IoT). The main challenge in delivering the IoT systems stems from a huge diversity in their energy, scalability, and data-rate constraints. The ICON lab at the Cheriton School of Computer Science is conducting research to address these challenges. The student will do research on designing new connectivity and sensing systems for virtual reality and smart home applications.
We are looking for students with one or multiple of the following skills:
- Experience with C/C++ programming, and interests in designing systems
- Familiarity with Micro-Controller, Raspberry Pi or FPGA programming
- Hardware design (familiarity with PCB design would be a plus)
- Research experience in wireless networks and systems
Interested students should email Omid Abari: omid.abari@uwaterloo.ca.
Title: Solar-powered off-grid embedded computing
We have developed a system that uses a solar panel and a battery to allow off-grid computing (details can be found here: http://blizzard.cs.uwaterloo.ca/iss4e/projectsenergy-access/ ). I am looking for a student who can (a) write scripts to diagnose faults in the system, such as shade falling on panels or battery issues and (b) come up with innovative uses for the system, such as a stand-alone Wikipedia server. Familiarity with Raspberry Pi a plus, but optional. Please send your CV and transcript to Prof. S. Keshav at keshav@uwaterloo.ca.
Title: Wrist-worn hand and finger gesture recognition study
The HCI lab at the Cheriton School of Computer Science is looking for a student interested in user study in hand and finger gesture interaction using wrist-worn sensors. The project requires excellent coding skills and some experience with hardware prototyping Additional experience with Arduino / Processing / User study is appreciated, but not necessary.
If you are interested, please email Keiko Katsuragawa at kkatsura@uwaterloo.ca.
Title: High-Performance Multi-Core Runtime Systems
The KOS project builds an experimental operating system kernel to investigate the intersection of design simplicity and performance in runtime systems. An important part of this study is carried out by building a nimble user-level threading runtime system for existing OS platforms, such as Linux and FreeBSD. Several development and evaluation activities are suitable for URAs in the context of this project, such as support for dynamic stack sizing, or fine-tuning the performance of synchronization operations. A URA candidate should be interested in system-level software and runtime internals, have experience with C/C++ programming, and be familiar with the POSIX system call interface.
If you are interested, please email Martin Karsten at: mkarsten@uwaterloo.ca.
Title: Cloud and Distributed Big Data Systems
If you are a senior undergraduate student interested in working with large scale distributed systems that involve managing big data and storage in the cloud to achieve scalability and deliver good performance, please e-mail Khuzaima Daudjee at: kdaudjee AT uwaterloo.ca
Title: Android Traffic Interception with PrivacyGuard
PrivacyGuard is an Android app that intercepts the traffic of all apps on a smartphone without requiring root access. By intercepting traffic, PrivacyGuard can provide useful features, such as notifying the user when an app leaks sensitive information (https://github.com/cryspuwaterloo/privacyguard). The goal of this URA is to add new features to PrivacyGuard.
Typically URAs involving PrivacyGuard consist of two parts: First, the student addresses some limitations of PrivacyGuard to become familiar with the code base. Second, the student, in collaboration with the supervisor, designs and develops a new feature for PrivacyGuard.
Android development experience is required. CS 350 or CS 456 would be useful.
Interested students should contact Urs Hengartner: urs.hengartner@uwaterloo.ca
Title: Enhancing a Main-Memory DBMS
We are looking for up to three students to fill URA positions to work on a main-memory database management system. The system is mainly coded in common lisp, but also has components written in C, Bison, Flex and csh. It is a "compiling" DBMS that has a simple extensible runtime and operates by translating SQL-like data manipulation requests directly to C code that interfaces with this runtime. The underlying data model is a generalization of the relational data model in which, among other things, arbitrary inheritance hierarchies may be declared.
During the course of the project, students will increase their familiarity with a range of topics, among which: semantic data modelling, query optimization, A* search, transaction compilation, legacy data integration, rule-based translation and source code synthesis.
Ideal applicants will have a solid background in algorithms and data structures, and a particular interest in program translation. Experience with programming in a functional style would also be an asset.
Interested students should contact Grant Weddell: gweddell@cs.uwaterloo.ca
Title: Techniques for Visual Data Storyelling
This project aims to explore novel interactive visualizations to support storytelling and communication with data insights. The URA will design and develop tools that integrate machine learning and data mining techniques with advanced visual interfaces. An idea candidate should be comfortable in web app development (e.g., JavaScript, NodeJS, D3) and data analytics (e.g., Python, Jupyter Notebook).
If you are interested, please email Jian Zhao at jianzhao@uwaterloo.ca with a resume and an unofficial transcript.
Title: Semantic Mining of Mathematical Text
Extracting semantic information from natural language text is recognized as an important problem in artificial intelligence, and there are many developed techniques. The present project is to extract semantic information from mathematical text. It is the long-term goal of the “International Mathematical Knowledge Trust” to capture the mathematical knowledge represented in the complete research literature of mathematics. There are several levels of semantic information, from general keyword spotting all the way to formalizing the mathematical statements in a language such as CoQ or Lean. The present project will perform n-gram analysis on the abstracts of mathematical article reviews provided by ZB-Math.
Interested students should contact Stephen Watt: smwatt@uwaterloo.ca
Title: Deep Learning for Handwriting Feature Recognition
One successful method for recognizing handwritten characters is based on representing them as parametric curves (x(s), y(s)) where x(s) and y(s) give the x and y positions of the pen as a function of arclength. Expressing x(s) and y(s) as truncated Legendre series or Legendre-Sobolev series has proven quite successful for recognition. These series use orthogonal polynomials as the basis, rather than monomials s^k. Recognition methods using proximity of coefficient vectors has proven quite successful. The present project is to explore how accuracy can be improved by using coefficient vectors to train neural networks for recognition.
Interested students should contact Stephen Watt: smwatt@uwaterloo.ca
Title: Machine Learning of Handwriting using Chebyshev-Sobolev Approximation
One successful method for recognizing handwritten characters is based on representing them as parametric curves (x(s), y(s)) where x(s) and y(s) give the x and y positions of the pen as a function of arclength. So far, expressing x(s) and y(s) as truncated Legendre series or Legendre-Sobolev series has proven quite successful. These series use orthogonal polynomials as the basis, rather than monomials s^k. This project will be to explore whether recognition can be improved using the Chebyshev and Chebyshev-Sobolev polynomials as basis functions.
Interested students should contact Stephen Watt: smwatt@uwaterloo.ca
Title: Algorithms for Symbolic Polynomials
Computer algebra systems, such as Maple, are very good at performing algebraic operation on polynomials with terms of fixed degree, but are much weaker when exponents involve variables. A new class of algorithms work with so-called “symbolic polynomials”, which are polynomial-like objects where the exponents are themselves integer valued polynomials, e.g. x^(n^2+3n) – 5y^(2n-1). The present project will develop or enhance an existing Maple package for algorithms on symbolic polynomials, including polynomial greatest common divisors and factorizations. Experience with Maple is required.
Interested students should contact Stephen Watt: smwatt@uwaterloo.ca
Title: Handwriting Neatening
Just as face averaging can be used to produce faces that are perceived as attractive, character averaging can be used to produce attractive characters written in a user’s own distinctive style. Representing handwritten characters as a particular type of feature vector allows this averaging to be well-defined and efficient. This can allow handwriting samples to be adjusted for uniformity of size, slant, pressure, etc. The present project will involve developing a tool to mark up character samples. Experience with Typescript and Node.js will be an asset to the URA candidate, and experience with calligraphy will be a plus.
Interested students should contact Stephen Watt: smwatt@uwaterloo.ca
Title: Suggestion Systems for Handwriting Recognition
On-line handwriting recognition with tablet or phone has the advantage that it is not necessary for the system to present a unique recognition result. It is possible to present a set of best guesses and have the user accept the top result or make a correction. This means that a user’s acceptance and correction history can be used to continuously train and individualize the recognition. This project is divided in two parts: the first is to explore methods to evolve the support set of symbol samples as character suggestions are accepted and rejected. The second is to understand to what degree handwriting “styles” can be used to predict preferred forms of symbols, i.e if a writer draws one character in a certain way, can this be used to predict how they will write other characters. Experience with Typescript and Node.js will be an asset to the URA candidate.
Interested students should contact Stephen Watt: smwatt@uwaterloo.ca
Title: NAN-Boxing for Read-Free Javascript Code
Programming language interpreters, such as those for Javascript, typically allocate data objects in memory augmented with header words that describe the object type and size. This means that type testing, array bounds checks and other operations require following a pointer and reading memory. This is much slower than register-to-register operations. One technique that is used to improve performance is so-called “NAN-boxing” which encodes pointers in the “not-a-number” space of floating point numbers. This project explores the performance gains that can be made by encoding additional information with the pointer, including certain values and size information. A URA candidate should be familiar with at least one systems level programming language, such as C/C++, and working with program-managed memory.
Interested students should contact Stephen Watt: smwatt@uwaterloo.ca
Title: Learner Modelling for On-Line STEM Education
On-line education is undergoing a transformation in which the traditional “repeat until you get the quiz right” model is starting to be replaced with multiple pre-programmed paths with detours. This project will explore using machine learning techniques on quiz data to identify classes of learners with correlated learning needs. This can allow individualized coaching by the on-line instruction system, and steer students toward explanations given in terms they have shown they understand. A URA candidate should have experience with, or be willing to learn about, R or python and SVMs.
Interested students should contact Stephen Watt: smwatt@uwaterloo.ca
For more information, please contact:
Shoshannah Holdom
Office: MC 4014
Phone: (519) 888-4567 ext. 36826
Email: sholdom@uwaterloo.ca
