{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,20]],"date-time":"2026-05-20T16:27:49Z","timestamp":1779294469619,"version":"3.51.4"},"reference-count":20,"publisher":"MDPI AG","issue":"7","license":[{"start":{"date-parts":[[2025,7,4]],"date-time":"2025-07-04T00:00:00Z","timestamp":1751587200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"name":"Ministry of Science, Technological Development and Innovation of the Republic of Serbia","award":["451-03-136\/2025-03\/200102"],"award-info":[{"award-number":["451-03-136\/2025-03\/200102"]}]},{"name":"Ministry of Science, Technological Development and Innovation of the Republic of Serbia","award":["101160293"],"award-info":[{"award-number":["101160293"]}]},{"name":"European Union\u2019s Horizon 2023 research and innovation program","award":["451-03-136\/2025-03\/200102"],"award-info":[{"award-number":["451-03-136\/2025-03\/200102"]}]},{"name":"European Union\u2019s Horizon 2023 research and innovation program","award":["101160293"],"award-info":[{"award-number":["101160293"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Information"],"abstract":"<jats:p>The 32-bit floating-point (FP32) format has many useful applications, particularly in computing and neural network systems. The classic 32-bit fixed-point (FXP32) format often introduces lower quality of representation (i.e., precision), making it unsuitable for real deployment, despite offering faster computations and reduced computational cost, which positively impacts energy efficiency. In this paper, we propose a switched FXP32 format able to compete with or surpass the widely used FP32 format across a wide variance range. It actually proposes switching between the possible values of key parameters according to the variance level of the data modeled with the Laplacian distribution. Precision analysis is achieved using the signal-to-quantization noise ratio (SQNR) as a performance metric, introduced based on the analogy between digital formats and quantization. Theoretical SQNR results provided in a wide range of variance confirm the design objectives. Experimental and simulation results obtained using neural network weights further support the approach. The strong agreement between the experiment, simulation, and theory indicates the efficiency of this proposal in encoding Laplacian data, as well as its potential applicability in neural networks.<\/jats:p>","DOI":"10.3390\/info16070574","type":"journal-article","created":{"date-parts":[[2025,7,4]],"date-time":"2025-07-04T03:47:26Z","timestamp":1751600846000},"page":"574","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":2,"title":["Switched 32-Bit Fixed-Point Format for Laplacian-Distributed Data"],"prefix":"10.3390","volume":"16","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-9003-2545","authenticated-orcid":false,"given":"Bojan","family":"Deni\u0107","sequence":"first","affiliation":[{"name":"Faculty of Electronic Engineering, University of Ni\u0161, Aleksandra Medvedeva 4, 18000 Ni\u0161, Serbia"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8267-9541","authenticated-orcid":false,"given":"Zoran","family":"Peri\u0107","sequence":"additional","affiliation":[{"name":"Faculty of Electronic Engineering, University of Ni\u0161, Aleksandra Medvedeva 4, 18000 Ni\u0161, Serbia"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7508-0277","authenticated-orcid":false,"given":"Milan","family":"Din\u010di\u0107","sequence":"additional","affiliation":[{"name":"Faculty of Electronic Engineering, University of Ni\u0161, Aleksandra Medvedeva 4, 18000 Ni\u0161, Serbia"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0009-0002-6523-4740","authenticated-orcid":false,"given":"Sofija","family":"Peri\u0107","sequence":"additional","affiliation":[{"name":"Faculty of Electronic Engineering, University of Ni\u0161, Aleksandra Medvedeva 4, 18000 Ni\u0161, Serbia"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-0748-4672","authenticated-orcid":false,"given":"Nikola","family":"Simi\u0107","sequence":"additional","affiliation":[{"name":"Faculty of Technical Sciences, University of Novi Sad, Trg Dositeja Obradovi\u0107a 6, 21102 Novi Sad, Serbia"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6419-2062","authenticated-orcid":false,"given":"Marko","family":"An\u0111elkovi\u0107","sequence":"additional","affiliation":[{"name":"IHP\u2014Leibniz-Institut f\u00fcr Innovative Mikroelektronik, Im Technologiepark 25, 15236 Frankfurt (Oder), Germany"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"1968","published-online":{"date-parts":[[2025,7,4]]},"reference":[{"key":"ref_1","unstructured":"(2019). Standard for Floating-Point Arithmetic (Standard No. IEEE 754-2019)."},{"key":"ref_2","doi-asserted-by":"crossref","unstructured":"Mienye, I.D., Swart, T.G., and Obaido, G. (2024). Recurrent neural networks: A comprehensive review of architectures, variants, and applications. Information, 15.","DOI":"10.20944\/preprints202408.0748.v1"},{"key":"ref_3","doi-asserted-by":"crossref","unstructured":"Aymone, F.M., and Pau, D.P. (2024). Benchmarking in-sensor machine learning computing: An extension to the MLCommons-Tiny suite. Information, 15.","DOI":"10.3390\/info15110674"},{"key":"ref_4","first-page":"3021","article-title":"A novel quantization and model compression approach for hardware accelerators in edge computing","volume":"80","author":"He","year":"2024","journal-title":"Comput. Mater. Contin."},{"key":"ref_5","unstructured":"Das, D., Mellempudi, N., Mudigere, D., Kalamkar, D., Avancha, S., Banerjee, K., Sridharan, S., Vaidyanathan, K., Kaul, B., and Georganas, E. (2018). Mixed precision training of convolutional neural networks using integer operations. arXiv."},{"key":"ref_6","doi-asserted-by":"crossref","first-page":"535","DOI":"10.1007\/s00521-021-06821-x","article-title":"S-DFP: Shifted dynamic fixed point for quantized deep neural network training","volume":"37","author":"Sakai","year":"2025","journal-title":"Neural Comput. Appl."},{"key":"ref_7","doi-asserted-by":"crossref","unstructured":"Kummer, L., Sidak, K., Reichmann, T., and Gansterer, W. (2023, January 27\u201329). Adaptive Precision Training (AdaPT): A Dynamic Fixed Point Quantized Training Approach for DNNs. Proceedings of the 2023 SIAM International Conference on Data Mining (SDM23), Minneapolis, MN, USA.","DOI":"10.1137\/1.9781611977653.ch63"},{"key":"ref_8","doi-asserted-by":"crossref","unstructured":"Alsuhli, G., Sakellariou, V., Saleh, H., Al-Qutayri, M., Mohammad, B., and Stouraitis, T. (2024). DFXP for DNN architectures. Number Systems for Deep Neural Network Architectures: Synthesis Lectures on Engineering, Science, and Technology, Springer.","DOI":"10.1007\/978-3-031-38133-1"},{"key":"ref_9","doi-asserted-by":"crossref","first-page":"20828","DOI":"10.1109\/ACCESS.2021.3054879","article-title":"Zero-centered fixed-point quantization with iterative retraining for deep convolutional neural network-based object detectors","volume":"9","author":"Sungrae","year":"2021","journal-title":"IEEE Access"},{"key":"ref_10","unstructured":"Wu, S., Li, G., Chen, F., and Shi, L. (2018). Training and inference with integers in deep neural networks. arXiv."},{"key":"ref_11","unstructured":"Banner, R., Nahshan, Y., Hoffer, E., and Soudry, D. (2018). ACIQ: Analytical clipping for integer quantization of neural networks. arXiv."},{"key":"ref_12","doi-asserted-by":"crossref","unstructured":"Peric, Z., Savic, M., Simic, N., Denic, B., and Despotovic, V. (2021). Design of a 2-bit neural network quantizer for Laplacian source. Entropy, 23.","DOI":"10.3390\/e23080933"},{"key":"ref_13","doi-asserted-by":"crossref","unstructured":"Peric, Z., Savic, M., Dincic, M., Vucic, N., Djosic, D., and Milosavljevic, S. (2021, January 25\u201327). Floating Point and Fixed Point 32-bits Quantizers for Quantization of Weights of Neural Networks. Proceedings of the 12th International Symposium on Advanced Topics in Electrical Engineering (ATEE 2021), Bucharest, Romania.","DOI":"10.1109\/ATEE52255.2021.9425265"},{"key":"ref_14","doi-asserted-by":"crossref","unstructured":"Peric, Z., and Dincic, M. (2023). Optimization of the 24-Bit fixed-point format for the Laplacian source. Mathematics, 11.","DOI":"10.3390\/math11030568"},{"key":"ref_15","doi-asserted-by":"crossref","first-page":"113037","DOI":"10.1016\/j.measurement.2023.113037","article-title":"Optimization of the fixed-point representation of measurement data for intelligent measurement systems","volume":"217","author":"Dincic","year":"2023","journal-title":"Measurement"},{"key":"ref_16","doi-asserted-by":"crossref","unstructured":"Jayant, N.C., and Noll, P. (1984). Digital Coding of Waveforms: Principles and Applications to Speech and Video, Prentice Hall.","DOI":"10.1016\/0165-1684(85)90053-2"},{"key":"ref_17","doi-asserted-by":"crossref","unstructured":"Gersho, A., and Gray, R. (1992). Vector Quantization and Signal Compression, Kluwer Academic Publishers.","DOI":"10.1007\/978-1-4615-3626-0"},{"key":"ref_18","doi-asserted-by":"crossref","unstructured":"Burgess, N., Milanovic, J., Stephens, N., Monachopoulos, K., and Mansell, D. (2019, January 10\u201312). Bfloat16 Processing for Neural Networks. Proceedings of the IEEE 26th Symposium on Computer Arithmetic (ARITH 2019), Kyoto, Japan.","DOI":"10.1109\/ARITH.2019.00022"},{"key":"ref_19","unstructured":"Lecun, Y., Cortez, C., and Burges, C. (2025, February 01). The MNIST Handwritten Digit Database. Available online: http:\/\/yann.lecun.com."},{"key":"ref_20","unstructured":"Krizhevsky, A., and Nair, V. (2025, February 01). The CIFAR-10 and CIFAR-100 Dataset. Available online: https:\/\/www.cs.toronto.edu."}],"container-title":["Information"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/2078-2489\/16\/7\/574\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,10,9]],"date-time":"2025-10-09T18:04:19Z","timestamp":1760033059000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/2078-2489\/16\/7\/574"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,7,4]]},"references-count":20,"journal-issue":{"issue":"7","published-online":{"date-parts":[[2025,7]]}},"alternative-id":["info16070574"],"URL":"https:\/\/doi.org\/10.3390\/info16070574","relation":{},"ISSN":["2078-2489"],"issn-type":[{"value":"2078-2489","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,7,4]]}}}