{"main_file_link":[{"open_access":"1","url":"https://proceedings.neurips.cc/paper/2021/file/88e1ce84f9feef5a08d0df0334c53468-Paper.pdf"}],"project":[{"name":"What’s in a memory? Spatiotemporal dynamics in strongly coupled recurrent neuronal networks.","_id":"c084a126-5a5b-11eb-8a69-d75314a70a87","grant_number":"214316/Z/18/Z"}],"title":"Online learning of neural computations from sparse temporal feedback","quality_controlled":"1","date_published":"2021-12-01T00:00:00Z","citation":{"ieee":"L. Braun and T. P. Vogels, “Online learning of neural computations from sparse temporal feedback,” in <i>Advances in Neural Information Processing Systems - 35th Conference on Neural Information Processing Systems</i>, Virtual, Online, 2021, vol. 20, pp. 16437–16450.","ama":"Braun L, Vogels TP. Online learning of neural computations from sparse temporal feedback. In: <i>Advances in Neural Information Processing Systems - 35th Conference on Neural Information Processing Systems</i>. Vol 20. Neural Information Processing Systems Foundation; 2021:16437-16450.","mla":"Braun, Lukas, and Tim P. Vogels. “Online Learning of Neural Computations from Sparse Temporal Feedback.” <i>Advances in Neural Information Processing Systems - 35th Conference on Neural Information Processing Systems</i>, vol. 20, Neural Information Processing Systems Foundation, 2021, pp. 16437–50.","short":"L. Braun, T.P. Vogels, in:, Advances in Neural Information Processing Systems - 35th Conference on Neural Information Processing Systems, Neural Information Processing Systems Foundation, 2021, pp. 16437–16450.","apa":"Braun, L., &#38; Vogels, T. P. (2021). Online learning of neural computations from sparse temporal feedback. In <i>Advances in Neural Information Processing Systems - 35th Conference on Neural Information Processing Systems</i> (Vol. 20, pp. 16437–16450). Virtual, Online: Neural Information Processing Systems Foundation.","ista":"Braun L, Vogels TP. 2021. Online learning of neural computations from sparse temporal feedback. Advances in Neural Information Processing Systems - 35th Conference on Neural Information Processing Systems. NeurIPS: Neural Information Processing Systems vol. 20, 16437–16450.","chicago":"Braun, Lukas, and Tim P Vogels. “Online Learning of Neural Computations from Sparse Temporal Feedback.” In <i>Advances in Neural Information Processing Systems - 35th Conference on Neural Information Processing Systems</i>, 20:16437–50. Neural Information Processing Systems Foundation, 2021."},"intvolume":"        20","year":"2021","month":"12","publication_identifier":{"isbn":["9781713845393"],"issn":["1049-5258"]},"conference":{"location":"Virtual, Online","start_date":"2021-12-06","end_date":"2021-12-14","name":"NeurIPS: Neural Information Processing Systems"},"author":[{"first_name":"Lukas","last_name":"Braun","full_name":"Braun, Lukas"},{"full_name":"Vogels, Tim P","id":"CB6FF8D2-008F-11EA-8E08-2637E6697425","first_name":"Tim P","orcid":"0000-0003-3295-6181","last_name":"Vogels"}],"status":"public","date_updated":"2022-06-20T07:12:58Z","oa_version":"Published Version","page":"16437-16450","type":"conference","acknowledgement":"We would like to thank Professor Dr. Henning Sprekeler for his valuable suggestions and Dr. Andrew Saxe, Milan Klöwer and Anna Wallis for their constructive feedback on the manuscript. Lukas Braun was supported by the Network of European Neuroscience Schools through their NENS Exchange Grant program, by the European Union through their European Community Action Scheme for the Mobility of University Students, the Woodward Scholarship awarded by Wadham College, Oxford and the Medical Research Council [MR/N013468/1]. Tim P. Vogels was supported by a Wellcome Trust Senior Research Fellowship [214316/Z/18/Z].","article_processing_charge":"No","user_id":"2DF688A6-F248-11E8-B48F-1D18A9856A87","department":[{"_id":"TiVo"}],"publisher":"Neural Information Processing Systems Foundation","day":"01","date_created":"2022-06-19T22:01:59Z","volume":20,"abstract":[{"text":"Neuronal computations depend on synaptic connectivity and intrinsic electrophysiological properties. Synaptic connectivity determines which inputs from presynaptic neurons are integrated, while cellular properties determine how inputs are filtered over time. Unlike their biological counterparts, most computational approaches to learning in simulated neural networks are limited to changes in synaptic connectivity. However, if intrinsic parameters change, neural computations are altered drastically. Here, we include the parameters that determine the intrinsic properties,\r\ne.g., time constants and reset potential, into the learning paradigm. Using sparse feedback signals that indicate target spike times, and gradient-based parameter updates, we show that the intrinsic parameters can be learned along with the synaptic weights to produce specific input-output functions. Specifically, we use a teacher-student paradigm in which a randomly initialised leaky integrate-and-fire or resonate-and-fire neuron must recover the parameters of a teacher neuron. We show that complex temporal functions can be learned online and without backpropagation through time, relying on event-based updates only. Our results are a step towards online learning of neural computations from ungraded and unsigned sparse feedback signals with a biologically inspired learning mechanism.","lang":"eng"}],"scopus_import":"1","_id":"11453","oa":1,"language":[{"iso":"eng"}],"publication":"Advances in Neural Information Processing Systems - 35th Conference on Neural Information Processing Systems","publication_status":"published"}