More real than reality: using deep learning and generative models to resolve how people make sense of other people’s behaviour

Studentship opportunity

This studentship is funded by the ESRC through the Scottish Graduate School of Social Science (SGSSS)

Institution

University of Aberdeen

Pathway

Psychology

Studentship

Steers

Mode of study

Full-time / Part-time

Application deadline

4pm, 14 April 2022

4pm, 13 April 2023

Project details

This project sits at the intersection of AI/deep learning and psychological research into people’s ability for making sense of the behaviour of other people. When people observe the actions of others, they perceive them not only in terms of visually apparent body movements, but in terms of the meaning behind these movements: a friend’s excitement when opening a present, their disgust when brushing away a spider, or simply that they “want” the thing they’re currently reaching for. Yet, while such attributions of meaning to behaviour underpin all social interactions, the underlying mechanisms are surprisingly unclear.

This project builds upon recent advances in AI/deep learning to test novel theoretical frameworks that may explain how these attributions occur and implement them in AI/deep learning networks. These frameworks propose that people’s understanding of others’ actions emerges not from a simple reading of their behaviour, but from active attempts to project meaning onto it, to test whether what we think about others reflects what they actually do. The project harnesses the fact that strikingly similar projection mechanisms underpin recent advances in deep learning /generative models for artificial 2D/3D image generation (e.g., diffusion models) that achieve performance good-enough to fool human observers in mistaking artificial for real images. This alignment of mechanisms makes it possible for the first time to implement the proposed mechanisms in deep-learning architectures, and to test: (1) whether these mechanisms capture human-like performance and biases, (2) whether such models can, like human observers, combine information about the seen behaviour with outside contextual information, and, ultimately, (3) whether their output is good enough to fool human observers into mistaking artificially generated action videos with real ones.

The student will be part of an interdisciplinary team, with Prof Patric Bach from the School of Psychology as primary supervisor and Dr Georgios Leontidis from the School of Natural and Computing Sciences as secondary supervisor. They will be part of Bach’s Action Prediction Lab and be able to draw on its long-term expertise in human social interaction and perception, as well as of the University’s Interdisciplinary Centre for Data & Artificial Intelligence, headed by Georgios Leontidis.

About the University

Aberdeen is Scotland’s third oldest University and the fifth oldest in the UK. It was founded in 1495, is now ranked as one of the UK’s top 20 Universities (Guardian University Guide 2022; Times and Sunday Times Good University Guide) and one of the world’s Top 200 (Times World University Rankings 2023). It was named Scottish University of the Year in the Times and Sunday Times Good University Guide 2019.

The student will become part of a multidisciplinary team of academics, researchers and other PhD students bridging the School of Psychology and the School of Natural and Computing Sciences. The School of Psychology has an excellent record of high impact research across the themes of Cognition, Perception, and Social Cognition. It provides a vibrant research environment with weekly research seminars, specialist research groups meetings, and state-of-the-art research infrastructure (e.g., psychophysics, motion capture, eye tracking, EEG, TMS). The School of Natural and Computing Sciences is internationally renowned for research and teaching excellence. Computing Science’s strategic focus lies across the themes of Artificial Intelligence (AI) and Trust, Identity, Privacy and Security (TIPS), with nearly £4M of recent funding from research councils, charities, the EU, and industry. The candidate will have access to its state-of-the-art facilities, High Performance Computing resources, GPU equipment and funding for other project costs.

In 2020, the University launched its Aberdeen 2040 strategy, which outlines its interdisciplinary research focus on five challenge areas, of which Data and AI is a core component. Part of this initiative is the Interdisciplinary Centre for Data and AI, which is led by Dr Georgios Leontidis.

Eligibility

Applicants must meet the following eligibility criteria:

Applicants will have a first degree (undergraduate) at 2:1 or above and have a demonstrable interest in the topic area under investigation.
Applicants can have a Masters degree, however this is not a requirement.
Applicants can study part-time or full-time.

This PhD studentship is offered to UK or international students who are available to start in late September/early October 2023. It is fully funded for either 3 years, or for 4 years (1+3), where the PhD is combined with an initial, additional year of training on the MRes in Psychology, alongside project-related work. The 1+3 route provides substantial training in experiment design and statistical analysis within human experimental psychology, which is essential for the project and therefore required for candidates that do not already have prior research training. Students should be passionate about developing deep learning/generative models and algorithms that capture human social perception, and should have qualifications in Computer Science, Engineering, Mathematics or related disciplines equal to an honour’s degree, at first or upper second class level from a UK institution or equivalent from a non-UK institution. We also consider students who have an honours degree (1^st or upper 2^nd level from a UK institution or equivalent from non-UK institution) in Psychology, as long as they meet the other skills criteria. We expect candidates to have strong skills in linear algebra, mathematics, programming (preferably, python), machine/deep learning principles, problem solving, and computer vision for generating content (2D and 3D models). These skills should be coupled with demonstrable interest in understanding how AI can capture human-like processing.

Funding

As per guidance published by UKRI in October 2020, a maximum of 30% of all studentships awarded can be made to international students, with the remaining 70% going to home students.

Residential Criteria

To be classed as a home student, applicants must meet the following criteria:

Be a UK national (meeting residency requirements), or
Have settled status, or
Have pre-settled status (meeting residency requirements), or
Have indefinite leave to remain or enter.

If you do not meet the criteria above, you will be classed as an international student. To establish if you would be classed as a home student, please see pages 4 and 5 of the UKRI eligibility guidance here.

As per guidance published by UKRI in October 2020, a maximum of 30% of all studentships awarded can be made to international students, of which SGSSS has now awarded for the 2021/22 round of studentship awards. As such, this opportunity is only open to home students.

Residential Criteria

To be classed as a home student, applicants must meet the following criteria:

Be a UK national (meeting residency requirements), or
Have settled status, or
Have pre-settled status (meeting residency requirements), or
Have indefinite leave to remain or enter.

If you do not meet the criteria above, you will be classed as an international student and will not be eligible to apply. To establish if you would be classed as a home student, please see pages 4 and 5 of the UKRI eligibility guidance here.

Award details

The scholarship is available as a +3 (3 year PhD) or a 1+3 (Masters year and 3 year PhD) studentship depending on prior research training. This will be assessed as part of the recruitment process, however you can access guidance here to help you decide on which to apply for. The programme will commence in October 2023. The full ESRC studentship package includes, as advised by ESRC:

An annual maintenance grant (stipend)
Fees at the standard institutional home rate
Students can also draw on a pooled Research Training Support Grant (RTSG)

Other information

The studentship is offered either as a three-year or four-year program. The three-year program will be offered to students who already have prior research training in human experimental psychology, experimental design, and statistical analysis. The four-year program will be offered to students who do not already have this psychology experience, which will involve an initial year undertaking the Aberdeen School of Psychology MRes in Psychological Research Methods, alongside the research work.

The successful candidate will be based in the School of Psychology and will work as part of our multi-disciplinary (Psychology-Computing Science) group with a supportive team of academics, researchers, and other PhD students in the school and the Interdisciplinary Centre for Data & Artificial Intelligence. The candidate will have access to state-of-the-art facilities, High Performance Computing resources, GPU equipment and funding for other project costs, such as training and conferences.

For further information or informal discussion please contact Prof Patric Bach (patric.bach@abdn.ac.uk) or Dr Georgios Leontidis (georgios.leontidis@abdn.ac.uk).

Interested candidates can familiarize themselves with the psychological research background by reading the information page of a concurrently running Leverhulme Trust project grant lead by Prof Patric Bach, to uncover the behavioural and neuroscientific mechanisms of social perception, or read this overview article.

How to apply

Applicants must register on SGSSS Apply, completing their Equal Opportunities data.
Applicants must apply via SGSSS Apply, uploading the following documentation:

- Microsoft Word application form (converted to PDF prior to upload)
- Academic transcripts
- Academic prizes
- Referee information
- CV

Please Note:

This is not an application to the relevant University, this is an application for SGSSS (ESRC) funding.
Students do not need a Masters/PhD offer from the relevant University before they can apply for funding, i.e. this studentship.
If successful in obtaining the SGSSS (ESRC) studentship, students can only start the funded studentship once they have an unconditional Masters/PhD degree offer from the relevant University. It is your responsibility to find out the University’s application process, including when you need to secure your offer, as SGSSS plays no role in this process.

This studentship opportunity will open for applications on 9th June.

Selection process

Applications will be ranked by an internal institutional selection panel, and you will be notified if you have been shortlisted for interview on or around the 24^th of April. Interviews will take place on 4^th and 5^th of May.

This studentship award is subject to the successful candidate securing admission to a PhD programme within the University of Aberdeen. The successful candidate will be invited to apply for admission to the relevant PhD programme after they are selected for funding.

Contact details

Name

SGSSS Team

Email

team@sgsss.ac.uk

Coronavirus (COVID-19) Updates For the latest information and advice see: Coronavirus (COVID-19) frequently asked questions

Scottish Graduate School of Social Science

Scottish Graduate School of Social Science

Studentship opportunity