{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/the-heidelberg-spiking-data-sets-for-the","title":"The Heidelberg Spiking Data Sets for the Systematic Evaluation of Spiking Neural Networks","arxiv_id":null,"date":"2022-07-07","proceeding":"Transactions on Neural Networks and Learning Systems 2022 7","authors":["Benjamin Cramer","Yannik Stradmann","Johannes Schemmel","and Friedemann Zenke"],"abstract":"Spiking neural networks are the basis of versatile\r\nand power-efficient information processing in the brain. Although\r\nwe currently lack a detailed understanding of how these networks\r\ncompute, recently developed optimization techniques allow us\r\nto instantiate increasingly complex functional spiking neural\r\nnetworks in-silico. These methods hold the promise to build more\r\nefficient non-von-Neumann computing hardware and will offer\r\nnew vistas in the quest of unraveling brain circuit function. To\r\naccelerate the development of such methods, objective ways to\r\ncompare their performance are indispensable. Presently, however,\r\nthere are no widely accepted means for comparing the com-\r\nputational performance of spiking neural networks. To address\r\nthis issue, we introduce two spike-based classification data sets,\r\nbroadly applicable to benchmark both software and neuro-\r\nmorphic hardware implementations of spiking neural networks.\r\nTo accomplish this, we developed a general audio-to-spiking\r\nconversion procedure inspired by neurophysiology. Furthermore,\r\nwe applied this conversion to an existing and a novel speech data\r\nset. The latter is the free, high-fidelity, and word-level aligned\r\nHeidelberg digit data set that we created specifically for this\r\nstudy. By training a range of conventional and spiking classifiers,\r\nwe show that leveraging spike timing information within these\r\ndata sets is essential for good classification accuracy. These results\r\nserve as the first reference for future performance comparisons\r\nof spiking neural networks.","url_abs":"https://ieeexplore.ieee.org/document/9311226","url_pdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9311226","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"the-heidelberg-spiking-data-sets-for-the","repo_url":"https://github.com/electronicvisions/lauscher","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}