[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-127361-en":3,"doc-seo-127361-105":30,"detail-sidebar-cat-0-en-105":92},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},127361,962085564549,"Genevieve","https://ap-avatar.wpscdn.com/davatar_085a072bc5b1113ac321206ff7593b45",8,"Research & Report","Implementing Engrams from a Machine Learning Perspective - The Relevance of a Latent Space - Short Comment","The note links biological “engram” implementation to autoencoders over recurrent neural networks, focusing on the role of the latent space. It analyzes how latent-space dimensionality relates to the complexity of encoded information and proposes a way biological systems could index latent representations to assemble parts of engrams into episodic memory. It also connects species-level connectome differences to cognitive capacities, concluding that human cognition may be structurally limited unlike machine learning systems.","arXiv :2407 . 166 16v 1 [ cs .NE] 23 Jul 2024  \nIMPLEMENTING ENGRAMS FROM A MACHINE LEARNING PERSPECTIVE: THE RELEVANCE OF A LATENT SPACE  \nSHORT COMMENT  \n Jesus Marco de Lucas  \njesus.marco@csic .es  \nAdvanced Computing and e-Science Group  \nInstituto de Física de Cantabria (IFCA) CSIC-Universidad de Cantabria, Santander, ES 39005, SPAIN  \nJuly 24, 2024  \nABSTRACT  \nIn our previous work, we proposed that engrams in the brain could be biologically implemented as autoencoders over recurrent neural networks. These autoencoders would comprise basic excitatory/inhibitory motifs, with credit assignment deriving from a simple homeostatic criterion. This brief note examines the relevance of the latent space in these autoencoders. We consider the relationship between the dimensionality of these autoencoders and the complexity of the information being encoded. We discuss how observed differences between species in their connectome could be linked to their cognitive capacities. Finally, we link this analysis with a basic but often overlooked fact:  \nhuman cognition is likely limited by our own brain structure. However, this limitation does not apply to machine learning systems, and we should be aware of the need to learn how to exploit this augmented vision of the nature.  \nKeywords engrams · RNN · autoencoders · latent space · connectome · concept neurons · cognition  \n1 Introduction  \nIn a previous work (Marco de Lucas, 2023) we have proposed that our brain processes the information through different neural networks, extracting the relevant information and storing it in a compressed form, following an architecture similar to an autoencoder, a key structure used in machine learning, and supporting the possible existence of \"concept cells\" (Quiroga, 2012) . We have also discussed (Peña et al., 2024) how a simple XOR neuronal motif, built withexcitatory and inhibitory neurons, could implement the analogy of a loss function in computational neural networks, solving the credit assignment problem following only a basic homeostatic criterion. Although this previous work was applied to basic engrams, following a model based in the simple C. Elegans neurons, it can be extended to models with more complex architectures, including those with spiking neurons, as it has been done in its computational counterpart (Bidollahkhani et al., 2023) for liquid time constant (LTC) neural networks (Hasani et al., 2021) .  \nIn this new work we will analyse how the structure of the latent space of these neural networks may have an impact on the cognitive potential. First, we will show how the complexity of the data being processed must be related to the structure of the neural network that processes it, and in a first approximation to the dimension of its latent space. We will then propose how a biological system may build an index on top of this latent space, and how different concepts could be linked as parts of an engram to compose episodic memory. We will discuss the implications on the cognitive potential of different species, comparing some basic ideas about their connectomes. Finally, we will reflect on the implications of a structural limit in the capacity of the human brain, which does not apply to machine learning systems.  \n2 Structure of an autoencoder and intrinsic dimension of data  \nThe processing of information in a neural network following the architecture of an autoencoder, and trained in unsupervised mode, is organized in different sections (see figure 1):  \n-The encoder section, that handles the incoming data and processes it through a series of “hidden layers”, extracting the most relevant features of each data to be able to identify it correctly when a new instance is presented.  \n-The latent space, a layer where the output of the encoder section is connected, with a reduced dimension.  \n-The decoder section, that enables recovering an “image” of the data, from a given point in the latent space. The decoder section may be completely symm","cbCairwF8tTJ0Vmc","https://ap.wps.com/l/cbCairwF8tTJ0Vmc","pdf",479659,1,6,"English","en",105,"# Introduction\n# Structure of an autoencoder and intrinsic dimension of data\n## Encoder, latent space, and decoder\n## Intrinsic dimension of data","[{\"question\":\"How does the work connect engrams to machine learning autoencoders?\",\"answer\":\"It proposes that engrams could be biologically implemented as autoencoders built on recurrent neural networks, where a latent space stores compressed representations used for later reconstruction and recognition.\"},{\"question\":\"What is the paper’s main focus regarding the latent space?\",\"answer\":\"It examines how the latent space structure and its dimensionality must relate to the complexity of the information being encoded, and how this may impact cognitive potential.\"},{\"question\":\"How does the document relate species differences to cognition?\",\"answer\":\"It discusses how observed differences between species in their connectomes could be linked to differences in cognitive capacities, tied to the computational properties of the proposed autoencoder-like organization.\"}]","Implementing Engrams from a Machine Learning Perspective - The Relevance of a Latent Space - Short Comment | PDF",1785938511,15,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":87,"head_meta":89,"extra_data":91,"updated_unix":28},"implementing-engrams-from-a-machine-learning-perspective-the-relevance-of-a-latent-space-short-comment","",{"@graph":36,"@context":86},[37,54,69],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/implementing-engrams-from-a-machine-learning-perspective-the-relevance-of-a-latent-space-short-comment/127361/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":63,"encodingFormat":61,"isAccessibleForFree":64,"interactionStatistic":65},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-22","2026-08-05",true,{"@type":66,"interactionType":67,"userInteractionCount":20},"InteractionCounter",{"@type":68},"ViewAction",{"@type":70,"mainEntity":71},"FAQPage",[72,78,82],{"name":73,"@type":74,"acceptedAnswer":75},"How does the work connect engrams to machine learning autoencoders?","Question",{"text":76,"@type":77},"It proposes that engrams could be biologically implemented as autoencoders built on recurrent neural networks, where a latent space stores compressed representations used for later reconstruction and recognition.","Answer",{"name":79,"@type":74,"acceptedAnswer":80},"What is the paper’s main focus regarding the latent space?",{"text":81,"@type":77},"It examines how the latent space structure and its dimensionality must relate to the complexity of the information being encoded, and how this may impact cognitive potential.",{"name":83,"@type":74,"acceptedAnswer":84},"How does the document relate species differences to cognition?",{"text":85,"@type":77},"It discusses how observed differences between species in their connectomes could be linked to differences in cognitive capacities, tied to the computational properties of the proposed autoencoder-like organization.","https://schema.org",{"og:url":52,"og:type":88,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":90,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":93},[94,98,102,106,111,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":95,"show_sort_weight":96,"slug":97},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":99,"show_sort_weight":100,"slug":101},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":103,"show_sort_weight":104,"slug":105},"Exam",70,"exam",{"id":107,"doc_module":4,"doc_module_name":46,"category_name":108,"show_sort_weight":109,"slug":110},5,"Comic",60,"comic",{"id":21,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":107,"slug":138},19,"General","general"]