{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,24]],"date-time":"2026-07-24T14:55:19Z","timestamp":1784904919347,"version":"3.55.0"},"reference-count":112,"publisher":"The Royal Society","issue":"2315","license":[{"start":{"date-parts":[[2026,2,26]],"date-time":"2026-02-26T00:00:00Z","timestamp":1772064000000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"},{"start":{"date-parts":[[2026,2,26]],"date-time":"2026-02-26T00:00:00Z","timestamp":1772064000000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/100018703","name":"HORIZON EUROPE European Innovation Council","doi-asserted-by":"publisher","id":[{"id":"10.13039\/100018703","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100000266","name":"Engineering and Physical Sciences Research Council","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100000266","id-type":"DOI","asserted-by":"publisher"}]},{"name":"Department of Science Innovation and Technology"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,2,26]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>Artificial intelligence (AI) research today is largely driven by ever-larger neural network models trained on graphics processing units (GPUs). This paradigm has yielded remarkable progress, but it also risks entrenching a hardware lottery in which algorithmic choices succeed primarily because they align with current hardware, rather than because they are inherently superior. In particular, the dominance of transformer architectures running on GPU clusters has led to an arms race of scaling up models, resulting in exorbitant computational costs and energy usage. At the same time, today\u2019s AI models often remain unreliable in the sense that they cannot properly quantify uncertainty in their decisions\u2014for example, large language models (LLMs) tend to produce incorrect outputs with high confidence. This paper argues that achieving more efficient and reliable AI will require embracing a set of principles that are well aligned with the goals of neuromorphic engineering, which are in turn inspired by how the brain processes information. Specifically, we outline six key neuromorphic principles, spanning algorithms, architectures and hardware, that can inform the design of future AI systems: (i) the use of stateful, recurrent models; (ii) extreme dynamic sparsity, possibly down to spike-based processing; (iii) backpropagation-free on-device learning and fine-tuning; (iv) probabilistic decision-making; (v) in-memory computing (IMC); and (vi) hardware-software co-design via stochastic computing. We discuss each of these principles in turn, surveying relevant prior work and pointing to directions for research.<\/jats:p>\n                  <jats:p>This article is part of the discussion meeting issue \u2018Bits, neurons and qubits for sustainable AI\u2019.<\/jats:p>","DOI":"10.1098\/rsta.2024.0522","type":"journal-article","created":{"date-parts":[[2026,2,28]],"date-time":"2026-02-28T20:48:32Z","timestamp":1772311712000},"source":"Crossref","is-referenced-by-count":2,"title":["Towards efficient and reliable artificial intelligence through neuromorphic principles"],"prefix":"10.1098","volume":"384","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-2960-6909","authenticated-orcid":false,"given":"Bipin","family":"Rajendran","sequence":"first","affiliation":[{"name":"Northeastern University London 1 , ,","place":["London, UK"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Osvaldo","family":"Simeone","sequence":"additional","affiliation":[{"name":"Northeastern University London 1 , ,","place":["London, UK"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Bashir","family":"Al-Hashimi","sequence":"additional","affiliation":[{"name":"King\u2019s College London 2 , , ,","place":["London, England, UK"],"department":["Department of Engineering"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"175","published-online":{"date-parts":[[2026,2,26]]},"reference":[{"key":"2026030110264341900_R1","article-title":"On the opportunities and risks of foundation models","author":"Bommasani","year":"2021","journal-title":"arXiv Preprint arXiv:2108.07258."},{"key":"2026030110264341900_R2","article-title":"LLaMA: open and efficient foundation language models","author":"Touvron","year":"2023","journal-title":"arXiv Preprint arXiv:2302.13971."},{"key":"2026030110264341900_R3","unstructured":"Crawford\n              K\n            \n            et al\n          . 2025This town\u2019s minerals make A.I. possible. Then came hurricane helene. The New York Times (Opinion). Seehttps:\/\/www.nytimes.com\/2025\/09\/26\/opinion\/ai-quartz-mining-hurricane-helene.html (accessed23-October-2025)."},{"key":"2026030110264341900_R4","unstructured":"Guo\n              C\n            , PleissG, SunY, WeinbergerKQ. 2017On calibration of modern neural networks. InProc. of the 34th Int. Conf. on Machine Learning (ICML), pp. 1321\u20131330. https:\/\/dblp.org\/rec\/conf\/icml\/2017.html."},{"key":"2026030110264341900_R5","doi-asserted-by":"crossref","first-page":"897","DOI":"10.1109\/TCCN.2023.3261300","article-title":"Robust bayesian learning for reliable wireless AI: framework and applications","volume":"9","author":"Zecchin","year":"2023","journal-title":"IEEE Trans. Cogn. Commun. Netw."},{"key":"2026030110264341900_R6","doi-asserted-by":"crossref","first-page":"58","DOI":"10.1145\/3467017","article-title":"The hardware lottery","volume":"64","author":"Hooker","year":"2021","journal-title":"Commun. ACM"},{"key":"2026030110264341900_R7","volume-title":"Seven and a half lessons about the brain","author":"Barrett","year":"2020"},{"key":"2026030110264341900_R8","article-title":"DeepSeek-V2: a strong, economical, and efficient mixture-of-experts language model","author":"DeepSeek","year":"2024","journal-title":"arXiv Preprint arXiv:2405.04434."},{"key":"2026030110264341900_R9","article-title":"FP8 formats for deep learning","author":"Micikevicius","year":"2022","journal-title":"arXiv Preprint arXiv:2209.05433."},{"key":"2026030110264341900_R10","article-title":"LLM.int8(): 8-bit matrix multiplication for transformers at scale","author":"Dettmers","year":"2022","journal-title":"arXiv Preprint arXiv:2208.07339."},{"key":"2026030110264341900_R11","article-title":"SmoothQuant: accurate and efficient post-training quantization for large language models","author":"Xiao","year":"2024","journal-title":"arXiv Preprint arXiv: 2211.10438."},{"key":"2026030110264341900_R12","unstructured":"Esser\n              SK\n            , McKinstryJL, BablaniD, AppuswamyR, ModhaDS. 2019Learned step size quantization. InProc. of Int. Conf. on Learning Representations (ICLR)."},{"key":"2026030110264341900_R13","doi-asserted-by":"crossref","unstructured":"Rastegari\n              M\n            , OrdonezV, RedmonJ, FarhadiA. 2016XNOR-Net: ImageNet classification using binary convolutional neural networks. InProc. of European Conf. on Computer Vision (ECCV), pp. 525\u2013542.","DOI":"10.1007\/978-3-319-46493-0_32"},{"key":"2026030110264341900_R14","article-title":"DoReFa-Net: training low bitwidth convolutional neural networks with low bitwidth gradients","author":"Zhou","year":"2016","journal-title":"arXiv Preprint arXiv:1606.06160."},{"key":"2026030110264341900_R15","article-title":"BitNet: scaling 1-bit transformers for large language models","author":"Wang","year":"2023","journal-title":"arXiv Preprint arXiv:2310.11453."},{"key":"2026030110264341900_R16","article-title":"Mamba: linear-time sequence modeling with selective state spaces","author":"Gu","year":"2023","journal-title":"arXiv Preprint arXiv:2312.00752."},{"key":"2026030110264341900_R17","article-title":"Retentive network: a successor to transformer for large language models","author":"Sun","year":"2023","journal-title":"arXiv Preprint arXiv:2307.08621."},{"key":"2026030110264341900_R18","unstructured":"Peng\n              B\n            \n            et al\n          . 2023RWKV: reinventing RNNs for the transformer era (eds BouamorH, PinoJ, BaliK). InFindings of the Association for Computational Linguistics: EMNLP 2023, pp. 14048\u201314077. Singapore: Association for Computational Linguistics. (doi:10.18653\/v1\/2023.findings-emnlp.936)"},{"key":"2026030110264341900_R19","unstructured":"Gu\n              A\n            , GoelK, R\u00e9C. 2022Efficiently modeling long sequences with structured state spaces. InProc. of int. conf. on learning representations."},{"key":"2026030110264341900_R20","doi-asserted-by":"crossref","unstructured":"Katti\n              P\n            , SkatchkovskyN, SimeoneO, RajendranB, Al-HashimiBM. 2023Bayesian inference on binary spiking networks leveraging nanoscale device stochasticity. In2023 IEEE Int. Symposium on Circuits and Systems (ISCAS). (doi:10.1109\/ISCAS46773.2023.10181438)","DOI":"10.1109\/ISCAS46773.2023.10181438"},{"key":"2026030110264341900_R21","doi-asserted-by":"crossref","first-page":"115","DOI":"10.1007\/BF02478259","article-title":"A logical calculus of the ideas immanent in nervous activity","volume":"5","author":"McCulloch","year":"1943","journal-title":"Bull. Math. Biophys."},{"key":"2026030110264341900_R22","volume-title":"The organization of behavior","author":"Hebb","year":"1949"},{"key":"2026030110264341900_R23","unstructured":"Vaswani\n              A\n            , ShazeerN, ParmarN, UszkoreitJ, JonesL, GomezAN, Kaiser\u0141, PolosukhinI. 2017Attention is all you need. InAdvances in neural information processing systems. vol. 30. Red Hook, NY: Curran Associates, Inc."},{"key":"2026030110264341900_R24","doi-asserted-by":"crossref","DOI":"10.1017\/CBO9780511815706","volume-title":"Spiking neuron models: single neurons, populations, plasticity","author":"Gerstner","year":"2002"},{"key":"2026030110264341900_R25","doi-asserted-by":"crossref","first-page":"179","DOI":"10.1207\/s15516709cog1402_1","article-title":"Finding structure in time","volume":"14","author":"Elman","year":"1990","journal-title":"Cogn. Sci."},{"key":"2026030110264341900_R26","doi-asserted-by":"crossref","first-page":"1735","DOI":"10.1162\/neco.1997.9.8.1735","article-title":"Long short-term memory","volume":"9","author":"Hochreiter","year":"1997","journal-title":"Neural Comput."},{"key":"2026030110264341900_R27","doi-asserted-by":"crossref","unstructured":"Cho\n              K\n            , van MerrienboerB, GulcehreC, BahdanauD, BougaresF, SchwenkH, BengioY. 2014Learning phrase representations using RNN encoder\u2013decoder for statistical machine translation. InProc. of the 2014 Conf. on Empirical Methods in Natural Language Processing (EMNLP).","DOI":"10.3115\/v1\/D14-1179"},{"key":"2026030110264341900_R28","doi-asserted-by":"crossref","unstructured":"Graves\n              A\n            , MohamedAR, HintonG. 2013Speech recognition with deep recurrent neural networks. In2013 IEEE Int. Conf. on Acoustics, Speech and Signal Processing, pp. 6645\u20136649. (doi:10.1109\/ICASSP.2013.6638947)","DOI":"10.1109\/ICASSP.2013.6638947"},{"key":"2026030110264341900_R29","doi-asserted-by":"crossref","unstructured":"Mikolov\n              T\n            , Karafi\u00e1tM, BurgetL, \u010cernock\u00fdJ, KhudanpurS. 2010Recurrent neural network based language model. InProc. of interspeech 2010, pp. 1045\u20131048.","DOI":"10.21437\/Interspeech.2010-343"},{"key":"2026030110264341900_R30","doi-asserted-by":"crossref","first-page":"157","DOI":"10.1109\/72.279181","article-title":"Learning long-term dependencies with gradient descent is difficult","volume":"5","author":"Bengio","year":"1994","journal-title":"IEEE Trans. Neural Netw."},{"key":"2026030110264341900_R31","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1145\/3530811","article-title":"Efficient transformers: a survey","volume":"55","author":"Tay","year":"2020","journal-title":"ACM Comput. Surv"},{"key":"2026030110264341900_R32","unstructured":"Gupta\n              A\n            , GuA, BerantJ. 2022Diagonal state spaces are as effective as structured state spaces. InProc. of the 36th Int. Conf. on Neural Information Processing Systemsser, ser. NIPS \u201922."},{"key":"2026030110264341900_R33","unstructured":"Beck\n              M\n            , P\u00f6ppelK, SpanringM, AuerA, PrudnikovaO, KoppM, KlambauerG, BrandstetterJ, HochreiterS. 2024xLSTM: extended long short-term memory. InProc. of the 38th Int. Conf. on Neural Information Processing Systems, ser. NIPS \u201924."},{"key":"2026030110264341900_R34","article-title":"ParaRNN: unlocking parallel training of nonlinear rnns for large language models","author":"Danieli","year":"2025","journal-title":"arXiv Preprint arXiv:2510.21450."},{"key":"2026030110264341900_R35","unstructured":"Katharopoulos\n              A\n            , VyasA, PappasN, FleuretF. 2020Transformers are RNNs: fast autoregressive transformers with linear attention. InProc. of Int. Conf. on Machine Learning."},{"key":"2026030110264341900_R36","unstructured":"Choromanski\n              K\n            \n            et al\n          . 2021Rethinking attention with performers. InInt. Conf. on Learning Representations."},{"key":"2026030110264341900_R37","article-title":"It\u2019s all connected: a journey through test-time memorization, attentional bias, retention, and online optimization","author":"Behrouz","year":"2025","journal-title":"arXiv Preprint arXiv:2504.13173."},{"key":"2026030110264341900_R38","doi-asserted-by":"crossref","first-page":"31788","DOI":"10.52202\/079017-0999","article-title":"Transformers to SSMs: distilling quadratic knowledge to subquadratic models","volume":"37","author":"Bick","year":"2024","journal-title":"Adv. Neural Inf. Process. Syst"},{"key":"2026030110264341900_R39","unstructured":"Dehghani\n              M\n            , GouwsS, VinyalsO, UszkoreitJ, Kaiser\u0141. 2019Universal transformers. InProc. of Int. Conf. on Learning Representations."},{"key":"2026030110264341900_R40","article-title":"Scaling latent reasoning via looped language models","author":"Zhu","year":"2025","journal-title":"arXiv Preprint arXiv:2510.25741."},{"key":"2026030110264341900_R41","doi-asserted-by":"crossref","first-page":"82","DOI":"10.1109\/MM.2018.112130359","article-title":"Loihi: a neuromorphic manycore processor with on-chip learning","volume":"38","author":"Davies","year":"2018","journal-title":"IEEE Micro."},{"key":"2026030110264341900_R42","unstructured":"Bal\n              M\n            , SenguptaA. 2025P-SpikeSSM: harnessing probabilistic spiking state space models for long-range dependency tasks. InProc. of Int. Conf. on Learning Representations."},{"key":"2026030110264341900_R43","doi-asserted-by":"crossref","DOI":"10.1093\/oso\/9780190905385.001.0001","volume-title":"The brain from inside out","author":"Buzs\u00e1ki","year":"2019"},{"key":"2026030110264341900_R44","doi-asserted-by":"crossref","first-page":"481","DOI":"10.1016\/j.conb.2004.07.007","article-title":"Sparse coding of sensory inputs","volume":"14","author":"Olshausen","year":"2004","journal-title":"Curr. Opin. Neurobiol."},{"key":"2026030110264341900_R45","first-page":"217","article-title":"Possible principles underlying the transformations of sensory messages","volume":"1","author":"Barlow","year":"1961","journal-title":"Sensory Communication"},{"key":"2026030110264341900_R46","doi-asserted-by":"crossref","first-page":"1133","DOI":"10.1097\/00004647-200110000-00001","article-title":"An energy budget for signaling in the grey matter of the brain","volume":"21","author":"Attwell","year":"2001","journal-title":"J. Cereb. Blood Flow Metab."},{"key":"2026030110264341900_R47","unstructured":"Evci\n              U\n            , GaleT, MenickJ, CastroPS, ElsenE. 2020Rigging the lottery: making all tickets winners. InProc. of the 37th Int. Conf. on Machine Learning. JMLR.org."},{"key":"2026030110264341900_R48","doi-asserted-by":"crossref","first-page":"1659","DOI":"10.1016\/S0893-6080(97)00011-7","article-title":"Networks of spiking neurons: the third generation of neural network models","volume":"10","author":"Maass","year":"1997","journal-title":"Neural Netw."},{"key":"2026030110264341900_R49","doi-asserted-by":"crossref","first-page":"607","DOI":"10.1038\/s41586-019-1677-2","article-title":"Towards spike-based machine intelligence with neuromorphic computing","volume":"575","author":"Roy","year":"2019","journal-title":"Nature"},{"key":"2026030110264341900_R50","doi-asserted-by":"crossref","first-page":"652","DOI":"10.1109\/JPROC.2014.2304638","article-title":"The SpiNNaker project","volume":"102","author":"Furber","year":"2014","journal-title":"Proc. IEEE"},{"key":"2026030110264341900_R51","doi-asserted-by":"crossref","first-page":"502","DOI":"10.1109\/TMLCN.2025.3556634","article-title":"Neuromorphic wireless split computing with multi-level spikes","volume":"3","author":"Wu","year":"2025","journal-title":"Trans. Mach. Learn. Comm. Netw."},{"key":"2026030110264341900_R52","article-title":"The era of 1-bit LLMs: all large language models are in 1.58 bits","author":"Ma","year":"2024","journal-title":"arXiv Preprint arXiv:2402.17764."},{"key":"2026030110264341900_R53","doi-asserted-by":"crossref","first-page":"243","DOI":"10.1145\/3007787.3001163","article-title":"EIE: efficient inference engine on compressed deep neural network","volume":"44","author":"Han","year":"2016","journal-title":"ACM SIGARCH Comput. Archit. News"},{"key":"2026030110264341900_R54","unstructured":"Shazeer\n              N\n            , MirhoseiniA, MaziarzK, DavisA, LeQV, HintonG, DeanJ. 2017Outrageously large neural networks: the sparsely-gated mixture-of-experts layer. InInt. Conf. on Learning Representations. Red Hook, NY: Curran Associates, Inc."},{"key":"2026030110264341900_R55","first-page":"1","article-title":"Switch transformers: scaling to trillion parameter models with simple and efficient sparsity","volume":"23","author":"Fedus","year":"2022","journal-title":"J. Mach. Learn. Res"},{"key":"2026030110264341900_R56","first-page":"341","volume-title":"Proceedings of machine learning and systems","author":"Korthikanti","year":"2023"},{"key":"2026030110264341900_R57","unstructured":"Malladi\n              S\n            , GaoT, NichaniE, DamianA, LeeJD, ChenD, AroraS. 2023Fine-tuning language models with just forward passes. InAdvances in neural information processing systems, vol. 36, pp. 53038\u201353075,"},{"key":"2026030110264341900_R58","article-title":"Memory-efficient backpropagation for fine-tuning LLMs on resource-constrained mobile devices","author":"Song","year":"2025","journal-title":"arXiv Preprint arXiv:2510.03425."},{"key":"2026030110264341900_R59","first-page":"482","article-title":"An overview of the simultaneous perturbation method for efficient optimization","volume":"19","author":"Spall","year":"1998","journal-title":"Johns Hopkins APL Tech. Dig."},{"key":"2026030110264341900_R60","doi-asserted-by":"crossref","first-page":"2788","DOI":"10.1109\/TIT.2015.2409256","article-title":"Optimal rates for zero-order convex optimization: the power of two function evaluations","volume":"61","author":"Duchi","year":"2015","journal-title":"IEEE Trans. Inf. Theory"},{"key":"2026030110264341900_R61","unstructured":"Liu\n              Y\n            , ZhuZ, GongC, ChengM, HsiehCJ, YouY. 2024Sparse MeZO: less parameters for better performance in zeroth-order LLM fine-tuning. InInt. Conf. on Machine Learning (ICML)."},{"key":"2026030110264341900_R62","doi-asserted-by":"crossref","unstructured":"Yang\n              Y\n            , ZhenK, BanijamaliE, MouchtarisA, ZhangZ. 2024Adazeta: adaptive zeroth-order tensor-train adaption for memory-efficient large language models fine-tuning. InEmpirical Methods in Natural Language Processing (EMNLP).","DOI":"10.18653\/v1\/2024.emnlp-main.56"},{"key":"2026030110264341900_R63","doi-asserted-by":"crossref","first-page":"64","DOI":"10.1109\/MSP.2019.2935234","article-title":"An introduction to probabilistic spiking neural networks: probabilistic models, learning rules, and applications","volume":"36","author":"Jang","year":"2019","journal-title":"IEEE Signal Process. Mag."},{"key":"2026030110264341900_R64","article-title":"Direct feedback alignment provides learning in deep neural networks","volume":"29","author":"N\u00f8kland","year":"2016","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"2026030110264341900_R65","article-title":"More than memory savings: zeroth-order optimization mitigates forgetting in continual learning","author":"Yu","year":"2025","journal-title":"arXiv Preprint arXiv:2510.21019."},{"key":"2026030110264341900_R66","article-title":"The forward-forward algorithm: some preliminary investigations","author":"Hinton","year":"2022","journal-title":"arXiv Preprint arXiv:2212.13345."},{"key":"2026030110264341900_R67","unstructured":"Lakshminarayanan\n              B\n            , PritzelA, BlundellC. 2017Simple and scalable predictive uncertainty estimation using deep ensembles. InAdvances in neural information processing systems (NeurIPS)."},{"key":"2026030110264341900_R68","doi-asserted-by":"crossref","DOI":"10.1017\/9781009072205","volume-title":"Machine learning for engineers","author":"Simeone","year":"2022"},{"key":"2026030110264341900_R69","unstructured":"Gal\n              Y\n            , GhahramaniZ. 2016Dropout as a bayesian approximation: representing model uncertainty in deep learning. InProc. of the 33rd Int. Conf. on Machine Learning (ICML), pp. 1050\u20131059."},{"key":"2026030110264341900_R70","unstructured":"Blundell\n              C\n            , CornebiseJ, KavukcuogluK, WierstraD. 2015Weight uncertainty in neural networks. InProc. of the 32nd Int. Conf. on Machine Learning."},{"key":"2026030110264341900_R71","unstructured":"Jiang\n              M\n            , RuanY, HuangS, LiaoS, PitisS, GrosseRB, BaJ. 2023Calibrating language models via augmented prompt ensembles. InInt. Conf. on Machine Learning."},{"key":"2026030110264341900_R72","doi-asserted-by":"crossref","unstructured":"Nan\n              Y\n            , HeP, TandonR, XuH. 2025Can multiple responses from an LLM reveal the sources of its uncertainty?InFindings of the Association for Computational Linguistics: EMNLP 2025, pp. 15551\u201315569. Suzhou, China: Association for Computational Linguistics.","DOI":"10.18653\/v1\/2025.findings-emnlp.841"},{"key":"2026030110264341900_R73","volume-title":"Algorithmic learning in a random world","author":"Vovk","year":"2005"},{"key":"2026030110264341900_R74","doi-asserted-by":"crossref","first-page":"494","DOI":"10.1561\/2200000101","article-title":"Conformal prediction: a gentle introduction","volume":"16","author":"Angelopoulos","year":"2023","journal-title":"Found. Trends Mach. Learn."},{"key":"2026030110264341900_R75","doi-asserted-by":"crossref","first-page":"44","DOI":"10.1109\/JSAIT.2024.3368229","article-title":"Forking uncertainties: reliable prediction and model predictive control with sequence models via conformal risk control","volume":"5","author":"Zecchin","year":"2024","journal-title":"IEEE J. Sel. Areas Inf. Theory"},{"key":"2026030110264341900_R76","doi-asserted-by":"crossref","first-page":"1235","DOI":"10.1109\/TSP.2024.3358615","article-title":"Federated inference with reliable uncertainty quantification over wireless channels via conformal prediction","volume":"72","author":"Zhu","year":"2024","journal-title":"IEEE Trans. Signal Process."},{"key":"2026030110264341900_R77","article-title":"Conformal calibration: ensuring the reliability of black-box ai in wireless systems","author":"Simeone","year":"2025","journal-title":"arXiv Preprint arXiv:2504.09310."},{"key":"2026030110264341900_R78","doi-asserted-by":"crossref","unstructured":"Horowitz\n              M\n            \n          . 20141.1 computing\u2019s energy problem (and what we can do about it). InIEEE Int. Solid-State Circuits Conf. Digest of Technical Papers (ISSCC), pp. 10\u201314.","DOI":"10.1109\/ISSCC.2014.6757323"},{"key":"2026030110264341900_R79","doi-asserted-by":"crossref","first-page":"3","DOI":"10.1109\/TVLSI.2020.3020286","article-title":"Mixed-signal computing for deep neural network inference","volume":"29","author":"Murmann","year":"2021","journal-title":"IEEE Trans. VLSI Syst."},{"key":"2026030110264341900_R80","doi-asserted-by":"crossref","first-page":"333","DOI":"10.1038\/s41928-018-0092-2","article-title":"In-memory computing with resistive switching devices","volume":"1","author":"Ielmini","year":"2018","journal-title":"Nat. Electron."},{"key":"2026030110264341900_R81","doi-asserted-by":"crossref","first-page":"529","DOI":"10.1038\/s41565-020-0655-z","article-title":"Memory devices and applications for in-memory computing","volume":"15","author":"Sebastian","year":"2020","journal-title":"Nat. Nanotechnol."},{"key":"2026030110264341900_R82","doi-asserted-by":"crossref","first-page":"1027","DOI":"10.1109\/JSSC.2022.3140414","article-title":"HERMES-core\u2014a 1.59-TOPS\/mm2 PCM on 14-nm CMOS in-memory compute core using 300-ps\/LSB linearized CCO-based ADCs","volume":"57","author":"Khaddam-Aljameh","year":"2022","journal-title":"IEEE J. Solid-State Circuits"},{"key":"2026030110264341900_R83","doi-asserted-by":"crossref","first-page":"211","DOI":"10.1038\/s41586-021-04196-6","article-title":"A crossbar array of magnetoresistive memory devices for in-memory computing","volume":"601","author":"Jung","year":"2022","journal-title":"Nature"},{"key":"2026030110264341900_R84","doi-asserted-by":"crossref","first-page":"325","DOI":"10.1126\/science.adf5538","article-title":"Fusion of memristor and digital compute-in-memory processing for energy-efficient edge computing","volume":"384","author":"Wen","year":"2024","journal-title":"Science"},{"key":"2026030110264341900_R85","doi-asserted-by":"crossref","DOI":"10.1038\/s41467-024-45670-9","article-title":"Hardware implementation of memristor-based artificial neural networks","volume":"15","author":"Aguirre","year":"2024","journal-title":"Nat. Commun."},{"key":"2026030110264341900_R86","doi-asserted-by":"crossref","first-page":"1093","DOI":"10.1109\/JSSC.2020.3039206","article-title":"Z-PIM: a sparsity-aware processing-in-memory architecture with fully variable weight bit-precision for energy-efficient deep neural networks","volume":"56","author":"Kim","year":"2021","journal-title":"IEEE J. Solid-State Circuits"},{"key":"2026030110264341900_R87","doi-asserted-by":"crossref","unstructured":"Lee\n              S\n            \n            et al\n          . 2021Hardware architecture and software stack for PIM based on commercial DRAM technology: industrial product. In2021 ACM\/IEEE 48th Annual Int. Symposium on Computer Architecture (ISCA), pp. 43\u201356. (doi:10.1109\/ISCA52012.2021.00013)","DOI":"10.1109\/ISCA52012.2021.00013"},{"key":"2026030110264341900_R88","doi-asserted-by":"crossref","first-page":"338","DOI":"10.1109\/JETCAS.2022.3160455","article-title":"An overview of processing-in-memory circuits for artificial intelligence and machine learning","volume":"12","author":"Kim","year":"2022","journal-title":"IEEE J. Emerg. Sel. Topics Circuits Syst."},{"key":"2026030110264341900_R89","doi-asserted-by":"crossref","unstructured":"Kang\n              M\n            , KeelMS, ShanbhagNR, EilertS, CurewitzK. 2014An energy-efficient VLSI architecture for pattern recognition via deep embedding of computation in SRAM. In2014 IEEE Int. Conf. on Acoustics, Speech and Signal Processing (ICASSP), pp. 8326\u20138330. (doi:10.1109\/ICASSP.2014.6855225)","DOI":"10.1109\/ICASSP.2014.6855225"},{"key":"2026030110264341900_R90","first-page":"1773","article-title":"Challenges and trends of sram-based computing-in-memory for AI edge devices","volume":"68","author":"Jhang","year":"2021","journal-title":"IEEE Trans. Circuits Syst. I."},{"key":"2026030110264341900_R91","doi-asserted-by":"crossref","DOI":"10.1063\/1.5042408","article-title":"A phase-change memory model for neuromorphic computing","volume":"124","author":"Nandakumar","year":"2018","journal-title":"J. Appl. Phys."},{"key":"2026030110264341900_R92","doi-asserted-by":"crossref","unstructured":"Le Gallo\n              M\n            , TumaT, ZipoliF, SebastianA, EleftheriouE. 2016Inherent stochasticity in phase-change memory devices. In2016 46th European Solid-State Device Research Conference (ESSDERC), pp. 373\u2013376. (doi:10.1109\/ESSDERC.2016.7599664)","DOI":"10.1109\/ESSDERC.2016.7599664"},{"key":"2026030110264341900_R93","doi-asserted-by":"crossref","DOI":"10.1038\/s41467-018-04933-y","article-title":"Neuromorphic computing with multi-memristive synapses","volume":"9","author":"Boybat","year":"2018","journal-title":"Nat. Commun."},{"key":"2026030110264341900_R94","doi-asserted-by":"crossref","first-page":"041102","DOI":"10.1063\/5.0168089","article-title":"Using the IBM analog in-memory hardware acceleration kit for neural network training and inference","volume":"1","author":"Le Gallo","year":"2023","journal-title":"APL Mach. Learn."},{"key":"2026030110264341900_R95","doi-asserted-by":"crossref","unstructured":"Burr\n              G\n            \n            et al\n          . 2014Experimental demonstration and tolerancing of a large-scale neural network (165,000 synapses), using phase-change memory as the synaptic weight element. InElectron Devices Meeting (IEDM), 2014 IEEE International, p. 29. (doi:10.1109\/IEDM.2014.7047135)","DOI":"10.1109\/IEDM.2014.7047135"},{"key":"2026030110264341900_R96","doi-asserted-by":"crossref","DOI":"10.1038\/s41467-020-16108-9","article-title":"Accurate deep neural network inference using computational phase-change memory","volume":"11","author":"Joshi","year":"2020","journal-title":"Nat. Commun."},{"key":"2026030110264341900_R97","doi-asserted-by":"crossref","DOI":"10.1038\/s41467-020-18098-0","article-title":"Committee machines\u2014a universal method to deal with non-idealities in memristor-based neural networks","volume":"11","author":"Joksas","year":"2020","journal-title":"Nat. Commun"},{"key":"2026030110264341900_R98","doi-asserted-by":"crossref","DOI":"10.3389\/fnins.2020.00406","article-title":"Mixed-precision deep learning based on computational memory","volume":"14","author":"Nandakumar","year":"2020","journal-title":"Front. Neurosci."},{"key":"2026030110264341900_R99","volume-title":"Analog vlsi and neural systems","author":"Mead","year":"1989"},{"key":"2026030110264341900_R100","doi-asserted-by":"crossref","first-page":"668","DOI":"10.1126\/science.1254642","article-title":"A million spiking-neuron integrated circuit with a scalable communication network and interface","volume":"345","author":"Merolla","year":"2014","journal-title":"Science"},{"key":"2026030110264341900_R101","doi-asserted-by":"crossref","first-page":"97","DOI":"10.1109\/MSP.2019.2933719","article-title":"Low-power neuromorphic hardware for signal processing applications: a review of architectural and system-level design approaches","volume":"36","author":"Rajendran","year":"2019","journal-title":"IEEE Signal Process. Mag."},{"key":"2026030110264341900_R102","doi-asserted-by":"crossref","DOI":"10.1038\/s41467-025-61576-6","article-title":"A neuromorphic processor with on-chip learning for beyond-CMOS device integration","volume":"16","author":"Greatorex","year":"2025","journal-title":"Nat. Commun."},{"key":"2026030110264341900_R103","doi-asserted-by":"crossref","first-page":"151","DOI":"10.1038\/s41928-020-00523-3","article-title":"In situ learning using intrinsic memristor variability via markov chain monte carlo sampling","volume":"4","author":"Dalgaty","year":"2021","journal-title":"Nat. Electron."},{"key":"2026030110264341900_R104","doi-asserted-by":"crossref","DOI":"10.3390\/mi14101840","article-title":"A design methodology for fault-tolerant neuromorphic computing using bayesian neural network","volume":"14","author":"Gao","year":"2023","journal-title":"Micromachines"},{"key":"2026030110264341900_R105","doi-asserted-by":"crossref","DOI":"10.1038\/s41467-023-43317-9","article-title":"Bringing uncertainty quantification to the extreme-edge with memristor-based Bayesian neural networks","volume":"14","author":"Bonnet","year":"2023","journal-title":"Nat. Commun."},{"key":"2026030110264341900_R106","doi-asserted-by":"crossref","DOI":"10.1038\/s41467-022-30305-8","article-title":"Neural sampling machine with stochastic synapse allows brain-like learning and inference","volume":"13","author":"Dutta","year":"2022","journal-title":"Nat. Commun."},{"key":"2026030110264341900_R107","article-title":"Ferroelectric NAND for efficient hardware bayesian neural networks","volume":"16","author":"Song","year":"2025","journal-title":"Nat. Commun."},{"key":"2026030110264341900_R108","article-title":"Bayesian neural network with unified entropy source and synapse weights using 3D 16-layer Fe-diode array","volume":"16","author":"Huang","year":"2025","journal-title":"Nat. Commun."},{"key":"2026030110264341900_R109","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1145\/3609116","article-title":"Spinbayes: algorithm-hardware co-design for uncertainty estimation using Bayesian in-memory approximation on spintronic-based architectures","volume":"22","author":"Ahmed","year":"2023","journal-title":"ACM Trans. Embed. Comput. Syst."},{"key":"2026030110264341900_R110","doi-asserted-by":"crossref","unstructured":"Jang\n              H\n            , SkatchkovskyN, SimeoneO. 2021BiSNN: training spiking neural networks with binary weights via Bayesian learning. In2021 IEEE Data Science and Learning Workshop (DSLW), pp. 1\u20136. IEEE. (doi:10.1109\/DSLW51110.2021.9523415)","DOI":"10.1109\/DSLW51110.2021.9523415"},{"key":"2026030110264341900_R111","doi-asserted-by":"crossref","DOI":"10.1038\/s41467-025-59011-x","article-title":"Thermodynamic computing system for AI applications","volume":"16","author":"Melanson","year":"2025","journal-title":"Nat. Commun."},{"key":"2026030110264341900_R112","article-title":"Stochastic quantum spiking neural networks with quantum memory and local learning","author":"Chen","year":"2025","journal-title":"arXiv Preprint arXiv:2506.21324."}],"container-title":["Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/royalsocietypublishing.org\/rsta\/article-pdf\/doi\/10.1098\/rsta.2024.0522\/6117087\/rsta.2024.0522.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/royalsocietypublishing.org\/rsta\/article-pdf\/doi\/10.1098\/rsta.2024.0522\/6117087\/rsta.2024.0522.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,3,1]],"date-time":"2026-03-01T15:27:02Z","timestamp":1772378822000},"score":1,"resource":{"primary":{"URL":"https:\/\/royalsocietypublishing.org\/rsta\/article\/384\/2315\/20240522\/480517\/Towards-efficient-and-reliable-artificial"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,2,26]]},"references-count":112,"journal-issue":{"issue":"2315","published-online":{"date-parts":[[2026,2,26]]},"published-print":{"date-parts":[[2026,2,26]]}},"URL":"https:\/\/doi.org\/10.1098\/rsta.2024.0522","relation":{},"ISSN":["1364-503X","1471-2962"],"issn-type":[{"value":"1364-503X","type":"print"},{"value":"1471-2962","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,2,26]]},"article-number":"20240522"}}