2,209 Hits in 7.4 sec

Neural representations of compositional structures: representing and manipulating vector spaces with spiking neurons

Terrence C. Stewart, Trevor Bekolay, Chris Eliasmith
2011 Connection science  
In particular, we describe methods for constructing spiking networks that can represent and manipulate structured, symbol-like representations.  ...  This paper re-examines the question of localist vs. distributed neural representations using a biologically realistic framework based on the central notion of neurons having a preferred direction vector  ...  Figure 1 is dependent on the neural model used to convert current into spikes. For this paper, we use the standard Leaky Integrate-and-Fire neuron. 2.  ... 
doi:10.1080/09540091.2011.571761 fatcat:bigxgbwpmfdkla27bb4sui3tam

Perceptual Spaces: Mathematical Structures to Neural Mechanisms

Q. Zaidi, J. Victor, J. McDermott, M. Geffen, S. Bensmaia, T. A. Cleland
2013 Journal of Neuroscience  
A central goal of neuroscience is to understand how populations of neurons build and manipulate representations of percepts that provide useful information about the environment.  ...  This symposium explores the fundamental properties of these representations and the perceptual spaces in which they are organized.  ...  At the top of the hierarchy is familiar Euclidean geometry and its non-Euclidean relatives, which allow representing stimuli as vectors, with well-defined sizes and angles.  ... 
doi:10.1523/jneurosci.3343-13.2013 pmid:24198350 pmcid:PMC3818541 fatcat:g7o4kys47nfgxmmy6a5hsmyxdq

Structured sequence processing and combinatorial binding: neurobiologically and computationally informed hypotheses

Ryan Calmus, Benjamin Wilson, Yukiko Kikuchi, Christopher I. Petkov
2019 Philosophical Transactions of the Royal Society of London. Biological Sciences  
The model integrates established vector symbolic additive and conjunctive binding operators with neurobiologically plausible oscillatory dynamics, and is compatible with modern spiking neural network simulation  ...  The neural mechanisms for segmenting continuous streams of sensory input and establishing representations of dependencies remain largely unknown, as do the transformations and computations occurring between  ...  We thank Elizabeth Buffalo, Pascal Fries, Karl Friston, Tim Griffiths, David Poeppel, Mark Stokes and Chris Summerfield for inspiring discussions on complex combinatorial binding.  ... 
doi:10.1098/rstb.2019.0304 pmid:31840585 pmcid:PMC6939361 fatcat:3arw5gklwvc5jaxrxfhilya6le

Analogical mapping and inference with binary spatter codes and sparse distributed memory

Blerim Emruli, Ross W. Gayler, Fredrik Sandin
2013 The 2013 International Joint Conference on Neural Networks (IJCNN)  
Vector symbolic architectures (VSAs) are a class of connectionist models for the representation and manipulation of compositional structures, which can be used to model analogy.  ...  Analogies require complex, relational representations of learned structures, which is challenging for both symbolic and neurally inspired models.  ...  Given that the representation and manipulation of structure is central to analogy and that VSAs can easily represent and manipulate structure there have been multiple applications of VSAs to different  ... 
doi:10.1109/ijcnn.2013.6706829 dblp:conf/ijcnn/EmruliGS13 fatcat:bm34lczbhbaebobtu2vlz5bxyq

Behavioral decomposition reveals rich encoding structure employed across neocortex [article]

Bartul Mimica, Tuce Tombaz, Claudia Battistin, Jingyi Guo Fuglstad, Benjamin Adric Dunn, Jonathan Robert Whitlock
2022 bioRxiv   pre-print
Momentary actions, such as rearing or turning, were represented ubiquitously and could be decoded from all sampled structures.  ...  However, more elementary and continuous features, such as pose and movement, followed region-specific organization, with neurons in visual and auditory cortices preferentially encoding mutually distinct  ...  Widerøe of the Norwegian University of Science and Technology (NTNU) MRI Core Facility and S. Eggen for veterinary oversight.  ... 
doi:10.1101/2022.02.08.479515 fatcat:rrb22y32yrgghmwbooe34ohnae

Simulating and Predicting Dynamical Systems With Spatial Semantic Pointers

Aaron R. Voelker, Peter Blouw, Xuan Choo, Nicole Sandra-Yaffa Dumont, Terrence C. Stewart, Chris Eliasmith
2021 Neural Computation  
While neural networks are highly effective at learning task-relevant representations from data, they typically do not learn representations with the kind of symbolic structure that is hypothesized to support  ...  and (2) be integrated with deep neural networks to predict the future of physical trajectories.  ...  The SPA provides methods for realizing and manipulating these representations in spiking (and nonspiking) neural networks.  ... 
doi:10.1162/neco_a_01410 pmid:34310679 fatcat:ebiydkydyvh2hbiboeiyxgev6m

Neural blackboard architectures of combinatorial structures in cognition

Frank van der Velde, Marc de Kamps
2006 Behavioral and Brain Sciences  
These architectures are briefly discussed, together with an example of a combinatorial structure in which the blackboard architectures for language and vision are combined.  ...  Human cognition is unique in the way in which it relies on combinatorial (or compositional) structures.  ...  Reduced vector coding Gayler and Sommer & Kanerva argue for an alternative solution of combinatorial structures in terms of reduced representations in high-dimensional vector spaces.  ... 
doi:10.1017/s0140525x06009022 pmid:16542539 fatcat:yjti7u4gxnaojk6m6gspdasmka

The Role of Temporal Structure in Human Vision

Randolph Blake, Sang-Hun Lee
2005 Behavioral & Cognitive Neuroscience Reviews  
They start with an overview of evidence bearing on temporal acuity of human vision, covering studies dealing with temporal integration and temporal differentiation.  ...  The authors conclude with a brief discussion of neurophysiological implications of these results.  ...  recreate a usable representation of the composite.  ... 
doi:10.1177/1534582305276839 pmid:15886401 fatcat:yvuatkje4ngwfi4dvi2kix3evy

Spike sequences and their consequences

Zoltán Nádasdy
2000 Journal of Physiology - Paris  
Spatio-temporal patterns of spikes have an advantage of representing information by their spike composition similar to words of languages.  ...  Their hidden dependency structure can be revealed by spike 'sequences', defined as a set of neurons which fire in a specific temporal order with certain delay between successive spikes.  ...  The author was supported by The Lady Davis Postdoctoral Fellowship and the Kranzberg Fund at the Hebrew University of Jerusalem.  ... 
doi:10.1016/s0928-4257(00)01103-7 pmid:11165916 fatcat:3qltafkoxjfydpjoezacxrctvq

The AHA! Experience: Creativity Through Emergent Binding in Neural Networks

Paul Thagard, Terrence C. Stewart
2010 Cognitive Science  
Many kinds of creativity result from combination of mental representations.  ...  How do people's brains come up with new ideas, theories, technologies, organizations, and aesthetic accomplishments? What neural processes underlie the wonderful AHA!  ...  Acknowledgements: Our research has been supported by the Natural Sciences and Engineering Research Council of Canada and SHARC/NET.  ... 
doi:10.1111/j.1551-6709.2010.01142.x pmid:21428991 fatcat:rsezlspwqjatpd2tn3bkh2zssq

A geometric framework to predict structure from function in neural networks [article]

Tirthabir Biswas, James E. Fitzgerald
2021 arXiv   pre-print
The structural connectivity matrix of synaptic weights between neurons is a critical determinant of overall network function, but quantitative links between neural network structure and function are complex  ...  Neural computation in biological and artificial networks relies on the nonlinear summation of many inputs.  ...  In neural networks, the structure of synaptic connectivity critically shapes the functional responses of neurons [4, 5] , and large-scale techniques for measuring neural network structure and function  ... 
arXiv:2010.09660v3 fatcat:tajxv5cfhncotgwe2tnkxcinuq

Hybrid preference machines based on inspiration from neuroscience

Stefan Wermter, Christo Panchev
2002 Cognitive Systems Research  
It is a first hybrid framework which allows a link between spiking neural networks, connectionist preference machines and symbolic finite state machines.  ...  In the past, a variety of computational problems have been tackled with different connectionist network approaches.  ...  an n-dimensional space. vertically and horizontally, and exploits recurrent That way, a symbolic machine represented a higher, and pulsed neural networks for more neuron-like more abstract representation  ... 
doi:10.1016/s1389-0417(01)00061-4 fatcat:u3xrqmrdirbztoyvkxi2k3iheu

Spike-based symbolic computations on bit strings and numbers [article]

Ceca Kraisnikovic, Wolfgang Maass, Robert Legenstein
2021 bioRxiv   pre-print
In addition, we present new results which show that surprisingly small spiking neural networks can perform symbolic computations on bit sequences and numbers and even learn such computations using a biologically  ...  We review the current state of research on spike-based symbolic computations of this type.  ...  Funding: This research was partially supported by the Human Brain Project (Grant Agreement number 785907 and 945539), the SYNCH project (Grant Agreement number 824162) of the European Union, and under  ... 
doi:10.1101/2021.07.14.452347 fatcat:wy4aynhxrnavnc67higeo3dgqu

A mechanism for the cortical computation of hierarchical linguistic structure

Andrea E. Martin, Leonidas A. A. Doumas, David Poeppel
2017 PLoS Biology  
Such redundancy in the cortical and machine signals is indicative of formal and mechanistic alignment between representational structure building and "cortical" oscillations.  ...  A single mechanism-using time to encode information across a layered network-generates the kind of (de)compositional representational hierarchy that is crucial for human language and offers a mechanistic  ...  The ability to learn structured representations contrasts with current Bayesian models, which assume structured representations a priori or fit them from a specified space of possible representations predefined  ... 
doi:10.1371/journal.pbio.2000663 pmid:28253256 pmcid:PMC5333798 fatcat:654avixjpnfplpxy3p5auz5ln4

Learning structured representations from experience [chapter]

Leonidas A.A. Doumas, Andrea E. Martin
2018 The psychology of learning and motivation  
DORA begins with representations of objects as flat feature vectors.  ...  In particular, because the model represents relations as weights in feature space and not in terms of any specific invariant properties, it has difficulty with full relational reasoning.  ... 
doi:10.1016/bs.plm.2018.10.002 fatcat:s25msu3l7ra7dhhsbdg55rtsni
« Previous Showing results 1 — 15 out of 2,209 results