Facebook’s AI doctoral research programme makes inroads into UK

Facebook is expanding its artificial intelligence (AI) PhD programme to the UK through a four-year partnership with University College London (UCL).

The PhD programme – which adds to those Facebook already has underway in the US, France and Canada – will initially host four students over the coming year, with the opportunity to add more each year.

The students will be assigned mentors from the Facebook AI Research (Fair) division, and will spend time at both UCL and Facebook as they purse their AI-related projects, which Facebook has indicated will be open-sourced.

In its announcement, Facebook said the mentors would be based at its Fair lab in London, which focuses on 3D computer vision, knowledge intensive and multilingual natural language processing (NLP), and reinforcement learning (RL). It added that while focus will be on attracting students with interest in these specific areas, the programme will also welcome students conducting research in other emerging areas of AI.

Facebook’s Fair London site is already in use by several students, including third-year PhD student Patrick Lewis, who is working on teaching machines to answer natural language questions.

“Fair affiliation really helps propel your research, whether it be through access to powerful research tools, like the Fair compute cluster, or world-class researchers in your field always willing to provide feedback,” said Lewis.

“The emphasis on ‘bottom-up’ research at Fair is also a great fit for PhD students, and the management team are excellent at supporting your growth personally and facilitating the research you want to do.

“Unlike a summer internship, the Fair PhD programme provides the continuity that allows for deeper collaboration, enabling us to build out long research visions and execute on them.”

Facebook already has strong links with UCL – which is recognised by the Research Excellence Framework as a top-ranking institution in the field of computer science, and is one of four UK research insittions involved in Europe’s “AI network of excellence”, Ellis – with several of its Fair researchers already affiliated with the university.

“Given the current rate of progress, pushing the limits of AI is a team effort, where connections between groups and across industrial and academic boundaries is key to success,” said Pontus Stenetorp, leader of the computer science department’s NLP group.

“Through the arrangements of this program, our PhD students have access to the people and resources from a world-leading academic institution in AI such as UCL and from a world-leading industrial research lab such as Fair. This makes the programme something very special and should appeal to any student who seeks to kick-start a career in AI.”

Head of computer science at UCL Steve Hailes, who has been working with Facebook to launch the programme, added: “The creation of a PhD programme based partly in UCL computer science and partly in Fair, and offers students a unique opportunity to see the world from two perspectives: building on the depth of expertise in UCL’s AI Centre to undertake world-class academic research, and gaining an understanding of how to have real-world impact.

“We believe this holistic view is an extremely powerful model for PhD study and will have wide appeal to all students looking for careers in AI, whether in academic research informed by practical constraints or in innovative industrial positions.”

Academic AI research in the UK has received significant funding from the government and other private companies in recent years since the publication of the government’s Industrial Strategy whitepaper in November 2017, which placed a focus on investment in AI and innovation, as well as digital skills and infrastructure

In February 2019, for example, the UK government launched a state and IT industry-funded graduate education programme in artificial intelligence as part of its AI Sector Deal.

It included 200 AI masters places at UK universities funded by Deepmind, QuantumBlack, Cisco and BAE Systems, in collaboration with the Institute of Coding and British Computer Society; PhDs at 16 dedicated UK Research and Innovation AI Centres for Doctoral Training; and Alan Turing AI fellowships.

A further £20m funding boost for AI research projects was granted in November 2020, covering issues such as identifying cancer earlier and processing data at lightning speed.