{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,9,18]],"date-time":"2025-09-18T10:42:54Z","timestamp":1758192174468,"version":"3.44.0"},"reference-count":22,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2025,9,17]],"date-time":"2025-09-17T00:00:00Z","timestamp":1758067200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Artif. Intell."],"abstract":"<jats:p>With the adoption of machine learning models in various practical domains, there is a growing need for evaluating and increasing model robustness. Hyperdimensional computing (HDC) is a neurosymbolic computational paradigm that represents symbols as high dimensional vectors and symbolic operations as vector operations, seamlessly interfacing between neuro- and symbolic components of a model. However, there is a notable gap in HDC research regarding the robustness of HDC models to input perturbations. This study presents a novel theoretical framework tailored to evaluate the robustness of hyperdimensional classifiers against perturbations in the input space. In particular, our proposed measure of robustness gives a theoretical upper bound for the magnitude of noise a model can tolerate without changing its prediction for any given data point. We also propose a method to enhance the robustness of the model based on our proposed measure of robustness. Our approach introduces several methods to calculate model robustness as a function of the specific dataset and type of hyperdimensional encoding used. The results show that the average robustness of HDC models increases under the proposed optimization scheme while maintaining accuracy by varying the variance of the Gaussian distribution used to encode hypervectors. The practical effectiveness of our proposed measure of robustness is also demonstrated.<\/jats:p>","DOI":"10.3389\/frai.2025.1637105","type":"journal-article","created":{"date-parts":[[2025,9,17]],"date-time":"2025-09-17T11:40:05Z","timestamp":1758109205000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Lipschitz-based robustness estimation for hyperdimensional learning"],"prefix":"10.3389","volume":"8","author":[{"given":"Calvin","family":"Yeung","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Hamza","family":"Errahmouni Barkam","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Zhuowen","family":"Zou","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Sanggeon","family":"Yun","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Nathaniel D.","family":"Bastian","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Mohsen","family":"Imani","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"1965","published-online":{"date-parts":[[2025,9,17]]},"reference":[{"key":"B1","first-page":"854","article-title":"\u201cParseval networks: Improving robustness to adversarial examples,\u201d","author":"Cisse","year":"2017","journal-title":"Proceedings of the 34th International Conference on Machine Learning"},{"key":"B2","first-page":"1310","article-title":"\u201cCertified adversarial robustness via randomized smoothing,\u201d","author":"Cohen","year":"2019","journal-title":"Proceedings of the 36th International Conference on Machine Learning"},{"key":"B3","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1501.07627","article-title":"Representing objects, relations, and sequences","author":"Gallant","year":"2015","journal-title":"arXiv"},{"key":"B4","author":"Gayler","year":"1998","journal-title":"Multiplicative Binding, Representation Operators"},{"key":"B5","first-page":"56","article-title":"\u201cOnlineHD: robust, efficient, and single-pass online learning using hyperdimensional system,\u201d","author":"Hern\u00e1ndez-Cano","year":"2021","journal-title":"2021 Design, Automation"},{"key":"B6","first-page":"358","article-title":"\u201cFully distributed representation\u201d","volume-title":"Real World Computing Symposium (RWC)","author":"Kanerva","year":"1997"},{"key":"B7","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1145\/3538531","article-title":"A survey on hyperdimensional computing aka vector symbolic architectures, part I: models and data transformations","volume":"55","author":"Kleyko","year":"2023","journal-title":"ACM Comput. Surv"},{"key":"B8","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1109\/NICE61972.2024.10549719","article-title":"\u201cCompositional factorization of visual scenes with convolutional sparse coding and resonator networks,\u201d","volume-title":"2024 Neuro Inspired Computational Elements Conference","author":"Kymn","year":"2024"},{"key":"B9","doi-asserted-by":"publisher","first-page":"SC1053","DOI":"10.35848\/1347-4065\/acb1b8","article-title":"Comprehensive analysis on error-robustness of fefet computation-in-memory for hyperdimensional computing","volume":"62","author":"Matsui","year":"2023","journal-title":"Jap. J. Appl. Phys"},{"key":"B10","doi-asserted-by":"publisher","first-page":"63","DOI":"10.3389\/frobt.2020.00063","article-title":"Symbolic representation and learning with hyperdimensional computing","volume":"7","author":"Mitrokhin","year":"2020","journal-title":"Front. Robot. AI"},{"key":"B11","doi-asserted-by":"publisher","first-page":"757125","DOI":"10.3389\/fnins.2022.757125","article-title":"GrapHD: graph-based hyperdimensional memorization for brain-like cognitive learning","volume":"16","author":"Poduval","year":"2022","journal-title":"Front. Neurosci"},{"key":"B12","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1801.09344","article-title":"Certified defenses against adversarial examples","author":"Raghunathan","year":"2018","journal-title":"arXiv"},{"key":"B13","doi-asserted-by":"crossref","first-page":"64","DOI":"10.1145\/2934583.2934624","article-title":"\u201cA robust and energy-efficient classifier using brain-inspired hyperdimensional computing,\u201d","volume-title":"Proceedings of the 2016 International Symposium on Low Power Electronics and Design","author":"Rahimi","year":"2016"},{"key":"B14","first-page":"1177","article-title":"\u201cRandom features for large-scale kernel machines,\u201d","volume-title":"Proceedings of the 20th International Conference on Neural Information Processing Systems, NIPS'07","author":"Rahimi","year":"2007"},{"key":"B15","article-title":"Gluing neural networks symbolically through hyperdimensional computing","author":"Sutor","year":"2022","journal-title":"arXiv preprint"},{"key":"B16","doi-asserted-by":"publisher","first-page":"215","DOI":"10.1613\/jair.1.12664","article-title":"A theoretical perspective on hyperdimensional computing","volume":"72","author":"Thomas","year":"2022","journal-title":"J. Artif. Int. Res"},{"key":"B17","first-page":"5286","article-title":"\u201cProvable defenses against adversarial examples via the convex outer adversarial polytope,\u201d","author":"Wong","year":"","journal-title":"Proceedings of the 35th International Conference on Machine Learning"},{"key":"B18","first-page":"5286","article-title":"\u201cProvable defenses against adversarial examples via the convex outer adversarial polytope,\u201d","volume-title":"International Conference on Machine Learning","author":"Wong","year":""},{"key":"B19","doi-asserted-by":"publisher","first-page":"200551","DOI":"10.1016\/j.iswa.2025.200551","article-title":"Cognitive map formation under uncertainty via local prediction learning. Intellig","volume":"27","author":"Yeung","year":"2025","journal-title":"Syst. Appl"},{"key":"B20","article-title":"Generalized holographic reduced representations","author":"Yeung","year":"2024","journal-title":"arXiv preprint"},{"key":"B21","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1906.06316","article-title":"Towards stable and efficient training of verifiably robust neural networks","author":"Zhang","year":"2019","journal-title":"arXiv"},{"key":"B22","doi-asserted-by":"publisher","DOI":"10.1109\/ASAP52443.2021.00039","article-title":"\u201cAssessing robustness of hyperdimensional computing against errors in associative memory: (invited paper),\u201d","author":"Zhang","year":"2021","journal-title":"2021 IEEE 32nd International Conference on Application-specific Systems, Architectures and Processors (ASAP)"}],"container-title":["Frontiers in Artificial Intelligence"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2025.1637105\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,9,17]],"date-time":"2025-09-17T11:40:09Z","timestamp":1758109209000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2025.1637105\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,9,17]]},"references-count":22,"alternative-id":["10.3389\/frai.2025.1637105"],"URL":"https:\/\/doi.org\/10.3389\/frai.2025.1637105","relation":{},"ISSN":["2624-8212"],"issn-type":[{"value":"2624-8212","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,9,17]]},"article-number":"1637105"}}