{"title":"白日梦 Hopfield 网络及其对相关数据的惊人功效","authors":"Ludovica Serricchio, Dario Bocchi, Claudio Chilin, Raffaele Marino, Matteo Negri, Chiara Cammarota, Federico Ricci-Tersenghi","doi":"arxiv-2405.08777","DOIUrl":null,"url":null,"abstract":"To improve the storage capacity of the Hopfield model, we develop a version\nof the dreaming algorithm that perpetually reinforces the patterns to be stored\n(as in the Hebb rule), and erases the spurious memories (as in dreaming\nalgorithms). For this reason, we called it Daydreaming. Daydreaming is not\ndestructive and it converges asymptotically to stationary retrieval maps. When\ntrained on random uncorrelated examples, the model shows optimal performance in\nterms of the size of the basins of attraction of stored examples and the\nquality of reconstruction. We also train the Daydreaming algorithm on\ncorrelated data obtained via the random-features model and argue that it\nspontaneously exploits the correlations thus increasing even further the\nstorage capacity and the size of the basins of attraction. Moreover, the\nDaydreaming algorithm is also able to stabilize the features hidden in the\ndata. Finally, we test Daydreaming on the MNIST dataset and show that it still\nworks surprisingly well, producing attractors that are close to unseen examples\nand class prototypes.","PeriodicalId":501066,"journal":{"name":"arXiv - PHYS - Disordered Systems and Neural Networks","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2024-05-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Daydreaming Hopfield Networks and their surprising effectiveness on correlated data\",\"authors\":\"Ludovica Serricchio, Dario Bocchi, Claudio Chilin, Raffaele Marino, Matteo Negri, Chiara Cammarota, Federico Ricci-Tersenghi\",\"doi\":\"arxiv-2405.08777\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"To improve the storage capacity of the Hopfield model, we develop a version\\nof the dreaming algorithm that perpetually reinforces the patterns to be stored\\n(as in the Hebb rule), and erases the spurious memories (as in dreaming\\nalgorithms). For this reason, we called it Daydreaming. Daydreaming is not\\ndestructive and it converges asymptotically to stationary retrieval maps. When\\ntrained on random uncorrelated examples, the model shows optimal performance in\\nterms of the size of the basins of attraction of stored examples and the\\nquality of reconstruction. We also train the Daydreaming algorithm on\\ncorrelated data obtained via the random-features model and argue that it\\nspontaneously exploits the correlations thus increasing even further the\\nstorage capacity and the size of the basins of attraction. Moreover, the\\nDaydreaming algorithm is also able to stabilize the features hidden in the\\ndata. Finally, we test Daydreaming on the MNIST dataset and show that it still\\nworks surprisingly well, producing attractors that are close to unseen examples\\nand class prototypes.\",\"PeriodicalId\":501066,\"journal\":{\"name\":\"arXiv - PHYS - Disordered Systems and Neural Networks\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-05-14\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - PHYS - Disordered Systems and Neural Networks\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2405.08777\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv - PHYS - Disordered Systems and Neural Networks","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2405.08777","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Daydreaming Hopfield Networks and their surprising effectiveness on correlated data
To improve the storage capacity of the Hopfield model, we develop a version
of the dreaming algorithm that perpetually reinforces the patterns to be stored
(as in the Hebb rule), and erases the spurious memories (as in dreaming
algorithms). For this reason, we called it Daydreaming. Daydreaming is not
destructive and it converges asymptotically to stationary retrieval maps. When
trained on random uncorrelated examples, the model shows optimal performance in
terms of the size of the basins of attraction of stored examples and the
quality of reconstruction. We also train the Daydreaming algorithm on
correlated data obtained via the random-features model and argue that it
spontaneously exploits the correlations thus increasing even further the
storage capacity and the size of the basins of attraction. Moreover, the
Daydreaming algorithm is also able to stabilize the features hidden in the
data. Finally, we test Daydreaming on the MNIST dataset and show that it still
works surprisingly well, producing attractors that are close to unseen examples
and class prototypes.