{"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/pre-defined-sparse-neural-networks-with","title":"Pre-Defined Sparse Neural Networks with Hardware Acceleration","arxiv_id":"1812.01164","date":"2018-12-04","proceeding":null,"authors":["Sourya Dey","Kuan-Wen Huang","Peter A. Beerel","Keith M. Chugg"],"abstract":"Neural networks have proven to be extremely powerful tools for modern\nartificial intelligence applications, but computational and storage complexity\nremain limiting factors. This paper presents two compatible contributions\ntowards reducing the time, energy, computational, and storage complexities\nassociated with multilayer perceptrons. Pre-defined sparsity is proposed to\nreduce the complexity during both training and inference, regardless of the\nimplementation platform. Our results show that storage and computational\ncomplexity can be reduced by factors greater than 5X without significant\nperformance loss. The second contribution is an architecture for hardware\nacceleration that is compatible with pre-defined sparsity. This architecture\nsupports both training and inference modes and is flexible in the sense that it\nis not tied to a specific number of neurons. For example, this flexibility\nimplies that various sized neural networks can be supported on various sized\nField Programmable Gate Array (FPGA)s.","url_abs":"http://arxiv.org/abs/1812.01164v1","url_pdf":"http://arxiv.org/pdf/1812.01164v1.pdf","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":"pre-defined-sparse-neural-networks-with","repo_url":"https://github.com/souryadey/predefinedsparse-nnets","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"pre-defined-sparse-neural-networks-with","repo_url":"https://github.com/usc-hal/predefinedsparse-nnets","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","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}