Facebook engineering manager and professor of Software Engineering at the University College London (UCL), Mark Harman was recently awarded a European Research Council (ERC) Advanced Grant, to apply to research focused on evolving program improvement collaborators (EPIC) in automated tools that offer specifically-evolved, explained and experimentally-justified advice on software improvements that optimize operational performance, while maintaining and/or extending functionality.
Funding frontier research and ideas, ERC grants are considered to be the most prestigious in the EUs Framework of programs for Research and Innovation. They are granted to excellent researchers in Europe working on pioneering and ambitious projects. The grant will be distributed through his work at UCL. Kicking off the research focus for the grant, Mark is announcing three fully-funded PhD Studentships. Students will be based at the UCL CREST centre and with research focused on the general area of Search Based Software Engineering (SBSE). Students will be supervised by Harman, together with a supervisory team, including Dr. Federica Sarro (http://www0.cs.ucl.ac.uk/staff/F.Sarro/) and Dr. Earl Barr (http://earlbarr.com), both also eminent software engineering researchers at UCL.
At Facebook, Mark manages the team working on the application of SBSE to automated software test design. His joint appointments foster the collaboration between academic research and industry application, where students may see the impact of their research at Facebook scale.
The main idea is that evolutionary computation can evolve software improvement collaborators; automated tools that offer specifically-evolved, explained and experimentally-justified advice on software improvements that optimise operational performance, while maintaining and/or extending functionality. This “Epi-Collaborator” will make suggestions, including transplantation of code from a donor system to a host, grafting of entirely new features grown (evolved) by the Epi-Collaborator, and identification and optimisation of tuneable deep parameters. One feature (and an important scientific and technical challenge for the project) is that these suggestions need to be backed by automatically-constructed quantitative evidence that justifies, explains and documents improvements.
The EPIC project thereby aims to introduce a new way of developing software, as a collaboration between human and machine, integrated into typically continuous integration code review repo frameworks. Rather than seeking to replace human intelligence with artificial intelligence, EPIC thus seeks to understand and exploit the complementary strengths of each: humans’ domain and contextual insights and machines’ ability to intelligently search large search spaces.
PhD students wishing to apply for one of the Studentships can apply on the UCL site here, by 28th April 2018.