Russell Greiner

Possible Summer Jobs (2016) for Undergraduate Students

If you are interested in any of the positions mentioned below, please send me ...

a cover letter, specifying which positions most interest you and indicating why you feel you qualify;
your academic transcript;
your CV, including a description of any previous research or industrial jobs you have held; and
a list of your references, preferably including their email addresses.

Note that we are only looking for students from the University of Alberta, or other local universities.

Russell Greiner
Email: rgreiner@ualberta.ca
359 Athabasca
Phone: (780) 492-4828 (admin)

See also slides (Nov 2015) for description of these, and for other ideas.

Patient Specific Survival Prediction
Intelligent Diabetes Management
Patient-specific Cancer Treatment
Human Metabolome Project
Budgeted Learning

Patient Specific Survival Prediction

We consider the challenge of predicting the survival time for individual patients -- or actually, the distribtion over survival time (like a Kaplan-Meier Curve, but for an individual, not a class of patients.) This led to our PSSP tool; check out this website; and description.

We are looking for a student to help us extend this:
(1) to deal with multiple observation times, for the same patient;
(2) to better explain the individual decisions: which features determine whether someone will die at 3 months, versus dying at 5 years, etc.

Project Webpage

Intelligent Diabetes Management

Patients with TypeI diabetes must regulate their own insulin, by administering insulin injections, several times a day. The amount is based on a parameterized formula, that involves their current blood glucose level and anticipated carbohydrate consumption, etc. This project seeks ways to improve that formula, for each individual patient, based on his/her specific logs. See project webpage.

Patient-specific Cancer Treatment

Learn which treatment should be most effective for each specific (cancer) patient, based on

Genetic data -- SNP (single nucleotide polyomorphism) profiles
Proteomic data -- microarrays (showing gene expression levels)
Metabolomic data -- eg, NMR urinalysis

In collaboration with Medical Researchers at the Cross Cancer Institute.
(For more information, see details wrt a PostDoc ad)

Human Metabolome Project

Predict chemical endpoints of metabolic processes (eg, after drinking coffee, what small molecules will be in your blood, or urine)
Given an NMR (resp, Mass Spec) spectrum, predict what compound(s) were present; see Bayesil (NMR) or CFM-ID (MS/MS).
Predict properties of novel metabolites, based on information about existing metabolites.
Find properties of novel metabolites, based on text in articles, books, website

Project Webpage

Budgeted Learning

Learning tasks typically begin with a data sample --- eg, symptoms and test results for a set of patients, together with their clinical outcomes. By contrast, many real-world studies begin with no actual data, but instead with a budget --- funds that can be used to collect the relevant information. For example, one study has allocated $30 thousand to develop a system to diagnose cancer, based on a battery of patient tests, each with its own (known) costs and (unknown) discriminative powers. Given our goal of identifying the most accurate classifier, what is the best way to spend the $30 thousand? Should we indiscriminately run every test on every patient, until exhausting the budget? Or, should we selectively, and dynamically, determine which tests to run on which patients? We call this task budgeted learning.

There are many open questions, both theoretic and empirical.

Formal analysis: is the task NP-hard given certain standard assumptions?
Are there any algorithms that are PTAS (approximation algorithms)?
Better heuristic algorithms and other efficiency tricks
Further empirical studies, over other datasets

Project webpage