The work package I am in-charge of is primarily involved in the design of better algorithms for spiking neural networks. To this end, successful candidates will conduct research in one or more of the following areas:
- Neuronal encoding: how to better encode external stimuli into spike based representations to facilitate decoding with high fidelity and also better learning performance (in terms of accuracy and power efficiency).
- Supervised learning: given that spiking neural networks are asynchronous and sparse in their activities, the design of supervised learning algorithms that can fully capitalize on these properties becomes critical.
- Mapping of state-of-art deep learning networks to spiking networks. Neuromorphic learning algorithms are still solving fairly simple problems compared to deep learning. For this, we would like to systematically borrow from the deep learning community networks and learning algorithms that can quickly boost the capabilities of spiking neural networks.
- Unsupervised learning. STDP is well suited for unsupervised learning in spiking neural networks, and we would like to further advance STDP learning in spiking neural networks (both theory and applications).
Preference will be given to candidates who can document knowledge in deep learning, spiking neural networks or signal processing (with interest in spiking neural networks).
Candidates must have a PhD (for scientists) or MS/BS (for engineers) in computer science, computational neuroscience or related fields. Strong programming and quantitative skills are highly desired. Candidates should be proficient in spoken and written English.
The appointment will be for 3 years, and extended for another 1 year, after review.The start date is flexible and applications will be considered on a rolling basis until the positions are filled.
