About the Project
The goal of this PhD is to provide a framework for enabling humans and machines to work together for situational understanding, by exploiting their respective strengths. Situational understanding refers to the ability to relate relevant information, identify gaps in the available information, and form logical conclusions that enable decisions and actions. This is the case of many complex, dynamic areas, such as driving vehicles (including airplanes), health care, emergency response, military command and control operations,..., where the final human decision makers needs to be supported in their understanding of the environment for taking effective decisions.
To take decisions, humans typically rely on their experience with similar situations as well as their knowledge about the domain, which allows them to operate without full information, and to explain these decisions to others. On the other hand, processing large amounts of information is often beyond their skills.
This is where machines can help, as modern data analysis and machine learning algorithms are able to efficiently handle large quantities of information to support inductive reasoning, i.e., to infer general rules and patterns from specific observations. However, many of these techniques are limited when it comes to taking into account the knowledge and experience accumulated by human experts, to identifying gaps in the available information, and to producing human- understandable explanations of their decisions, which are all crucial when humans and machines have to work together to make decisions that impact the real world. Furthermore, these techniques often require large amounts of training data, which may not be available in practice.
For this reason, we will investigate more logical-oriented approaches to machine learning, such as probabilistic logic programming and Bayesian networks.
This PhD will investigate techniques to bring together the strengths of humans and machines in a new framework that supports effective human-in-the-loop situational understanding, building upon and extending existing work in fields such as artificial intelligence, machine learning, and reasoning under uncertainty. It will be carried out within the Distributed Analytics and Information Sciences International Technology Alliance (DAIS ITA), a collaboration between Cardiff University, IBM, Airbus, BAE Systems, University College London, University of California Los Angeles, and other UK and US partners - see https://dais-ita.org/ . At Cardiff University the PhD will be supervised by members of the Crime and Security Research Institute and the School of Computer Science and Informatics.
Funding Notes
The Studentship covers tuition fees at UK/EU level (£4,284 p.a. 2018/19), and a maintenance stipend at the RCUK rate (£14,777 p.a. 2018/19).
Residency eligibility: Open to all, but an international student would be required to fund the difference between overseas tuition fee (£19,950 p.a. 2018/19) and above UK/EU tuition fee.
Academic eligibility: Applicants will have or expect to have a 1st or 2:1 Honours degree or a master's degree, in computing or a related subject. Applicants for whom English is not their first language must demonstrate an IELTS score of 6.5 minimum overall, with 6.0 minimum in each skills component.