Una-May O’Reilly holds a BSc from the University of Calgary, and a MCS and PhD (1995) from Carleton University (Ottawa, Canada). O’Reilly joined MIT Computer Science Artificial Intelligence Laboratory (CSAIL) as a Post-Doctoral Associate in 1996. Now, she is the leader of the AnyScale Learning For All (ALFA) group, editor for Data Analytics and Knowledge Discovery for Genetic Programming and Evolvable Machines, editor for Evolutionary Computation, and action editor for the Journal of Machine Learning Research at MIT. Some of her achievements include: the EvoStar Award for Outstanding Contribution of Evolutionary Computation (2013), Fellow of the International Society of Genetic and Evolutionary Computation, and ACM sig-EVO.
Industry Impact
O’Reilly focuses her research on scalable machine learning, evolutionary algorithms, and frameworks for large scale knowledge mining, prediction and analytics. The vision is to develop data-driven machine learning systems that advance the quality of healthcare, the understanding of cyber arm races, and the delivery of online education.
As the healthcare industry is transforming by the ability to record massive amounts of information about individual patients, the enormous volume of data being collected is almost impossible to analyze. Machine learning provides ways to instantly find patterns and reason about data, which enables healthcare professionals to move to personalized care. For example, healthcare providers can take advantage in being able to foresee hospital re-admission for chronically ill patients. Being able to recognize those patients that have the probability of being re-admitted can have superior support after being discharged.
Gigabeats: Data science for medical sensor data
This project investigates machine learning to interpret and exploit repositories holding waveform data e.g. arterial blood pressure, ECG and EEG. This includes “BeatDB” where the development of a fast and scalable framework for compiling machine learning data sets from waveform repositories and “Trajectories Like Mine” where the development of a sub-linear time method called Locality Sensitive Hashing to find the nearest neighbors in waveform space. The goal is to help medical researchers and clinicians understand the growing repositories of waveform and signal data collected from critically ill patients.
Adversarial Cyber Security
Cyberspace has become a competitive platform occupied by intelligent, adaptive adversaries. Defenders and attackers engage in arm races as both sides take turns crafting new responses to each other’s actions. The goal is to understand the nature of cyber security arms between malicious and bonafide parties. With this new technical approach, the aim is to frame a robust optimization problem where the objectives of the two sides conflict and the positive gains of one side imply negative outcomes for the other. The co-optimization problem is solved with co-evolutionary algorithms.
MOOC Learner Project: Data science for E-Learning
The MOOC Learner Project addresses data science and machine learning challenges arising from investigating and MOOC learner clickstream, video watching, and forum data. A part of this research investigates software engineering practices that support extending data schemas. The aim is to provide learning scientists, instructional designers and online education specialists with open source software that enables them to efficiently extract teaching and learning insights from the data gathered when students learn on the edX or open edX platform.