Directorate: School of Engineering & Physical Sciences
Grade: Grade 7 (£36,924-£46,485)
Contract Type: Full Time (1FTE), Fixed Term (18 Months)
Rewards and Benefits: 33 days annual leave, plus 9 buildings closed days (and Christmas Eve when it falls on a weekday) for all full time staff.
Detailed description
The PDRA will work as part of the Prosperity Partnership between Heriot-Watt University, the University of Edinburgh and Leonardo UK. This project aims to address the current limitations of traditional frame-based sensors and associated processing pipelines with a new family of algorithmic architectures that mimic more closely the behaviours of biological brains.
Spiking neural networks (SNNs) can offer increased processing speed and reduced power consumption, especially when implemented on dedicated hardware (neuromorphic chips or FPGAs). Standard SNNs are typically fed with spiking data (e.g., streams of binary values), and the output of each layer remains spiking. They are particularly well suited for several new sensors, such as event-cameras, or single-photon detectors (used for single-photon Lidar), that natively produce event-like data that is compatible with spiking networks.
SNNs appear promising at the conceptual level but they currently suffer from limitations preventing the broad deployment. First, efficient and scalable training procedures are still needed. Second, deployment of SNNs on actual neuromorphic hardware can come with additional structural constraints on the SNNs.
In this project, we will investigate Bayesian methods to train deterministic SNNs or probabilistic SNNs. Bayesian deep learning methods have been shown to accelerate and improve the training procedure of SNNs.
Research Environment
The project is in collaboration with two partners: (i) IDCOM at the University of Edinburgh, which develops theory, algorithms and hardware for the next generation of signal processing, imaging and communication systems and (ii) Leonardo UK, providing application expertise and end-user feedback. The applicant will be expected to work directly with Dr. Yoann Altmann, Prof. Steve McLaughlin, and Prof. Mike Davies, as well as other researchers.
The PDRA will report on project progress and outcomes to the Prosperity Partnership Management Group.
Key Duties and Responsibilities
* Conduct scientific research; analyze and interpret data; oversee activities of project students; communicate with other investigators; prepare scientific papers; present research at conferences; supervise junior group members.
* Contribute to experimental design and procedure; implement experiments, theoretical models, and data analysis.
* Assist in the maintenance of experimental facilities, liaise with external collaborators, develop student research skills, and contribute to teaching.
* Participate in outreach activities, which may include interviews and public talks.
Education, Qualifications and Experience
Essential Criteria
* PhD in a relevant area of Engineering, Mathematics, Physics or related subject.
* Ability to articulate research work in written reports and oral presentations.
* Proven academic ability and technical competence in computational data science and analysis/modeling of results.
* Theoretical or experimental experience relevant to the project.
* Ability to work independently and as part of a team.
Desirable Criteria
* Experience in writing scientific papers.
* Understanding of statistical machine learning methods relevant to the project.
* Experience in programming (e.g., Python) and data analysis.
* Ability to guide other researchers.
* Willingness to learn new digital skills.
Travel
Monthly travel within Edinburgh (UoE, Leonardo) is expected, along with attendance at national and international conferences.
How to Apply
Please submit a cover letter describing your interest and suitability for the post, along with your full CV, via the Heriot-Watt online recruitment system. Applications can be submitted until midnight on Thursday 12th December 2024.
For further information and an informal discussion, please contact Dr. Yoann Altmann via email y.altmann@hw.ac.uk.
#J-18808-Ljbffr