Applications are invited for a PhD fellowship/scholarship at Graduate School of Natural Sciences, Aarhus University, Denmark, within the Computer Science programme. The position is available from August 2024 or later.
Title:
Theoretical Understanding of Classic Learning Algorithms
Research area and project description:
Over the past decades, analysing data has become ubiquitous in all areas of science, industry, and society in general. In particular machine learning techniques for recognising patterns and making predictions from data, have become absolutely central.
Classic learning techniques, such as Boosting, Bagging, Random Forest and Support Vector Machines excel at learning tasks where data is scarce, expensive to come by, or tabular. Since many smaller businesses and research groups cannot collect huge data sets, nor spend the computing resources needed to train large deep learning models, classic learning algorithms are essential, and making the most out of limited training data is key to training accurate models. Classic tree-building learning algorithms, such as Random Forest and Gradient Boosting, have the added benefit of requiring less data cleaning and providing better out-of-the-box performance than deep learning approaches.
Research challenges and objectives
Despite the overwhelming practical successes of machine learning, many fundamental theoretical questions remain unanswered and better classic learning algorithms, requiring less training data to make accurate predictions, may still be developed. In light of this, the first research challenge of the project is:
How can we improve classic learning algorithms to make better predictions from less data?
We address this challenge from three different directions:
Moreover, what makes this project unique is that we are not only interested in developing better algorithms, but also in understanding the fundamental limitations of learning. We therefore also address the following challenge:
Can we prove, via lower bounds, that the algorithms we develop are optimal in terms of accuracy and running time?
To address the above challenges, this project will develop classic learning algorithms that are provably more efficient in terms of training time, samples needed, accuracy and parallelization. We will complement these algorithms with lower bounds establishing their optimality and shedding light on the fundamental barriers in learning.
For technical reasons, you must upload a project description. When - as here - you apply for a specific project, please simply copy the project description above, and upload it as a PDF in the application. If you wish to, you can indicate an URL where further information can be found.
Qualifications and specific competences:
The candidate should have a Computer Science or Mathematics background, with a focus on theoretical/math courses when it comes to electives. Having passed a Machine Learning course is a requirement. Relevant algorithms courses are also strongly appreciated.
Place of employment and place of work:
The place of employment is Aarhus University, and the place of work is Department of Computer Science, Åbogade 34, 8200 Aarhus N, Denmark.
Contacts:
Applicants seeking further information for this project are invited to contact: Professor Kasper Green Larsen, larsen@cs.au.dk
How to apply:
For information about application requirements and mandatory attachments, please see the Application guide. Please read the Application guide thoroughly before applying.
When ready to apply, go to https://phd.nat.au.dk/for-applicants/apply-here/ (Note, the online application system opens 1 March 2024)
Please note:
At the Faculty of Natural Science at Aarhus University, we strive to support our scientific staff in their career development. We focus on competency development and career clarification and want to make your opportunities transparent. On our website, you can find information on all types of scientific positions, as well as the entry criteria we use when assessing candidates. You can also read more about how we can assist you in your career planning and development.
Aarhus University’s ambition is to be an attractive and inspiring workplace for all and to foster a culture in which each individual has opportunities to thrive, achieve and develop. We view equality and diversity as assets, and we welcome all applicants. All interested candidates are encouraged to apply, regardless of their personal background.
