Probabilistic Inference in Discrete Spaces Can Be Implemented into Networks of LIF Neurons
The means by which cortical neural networks are able to efficiently solve inference problems remains an open question in computational neuroscience. Recently, abstract models of Bayesian computation in neural circuits have been proposed, but they lack a mechanistic interpretation at the single-cell...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2015-02-01
|
| Series: | Frontiers in Computational Neuroscience |
| Subjects: | |
| Online Access: | http://journal.frontiersin.org/Journal/10.3389/fncom.2015.00013/full |
| _version_ | 1818130560667615232 |
|---|---|
| author | Dimitri eProbst Mihai Alexandru Petrovici Ilja eBytschok Johannes eBill Dejan ePecevski Johannes eSchemmel Karlheinz eMeier |
| author_facet | Dimitri eProbst Mihai Alexandru Petrovici Ilja eBytschok Johannes eBill Dejan ePecevski Johannes eSchemmel Karlheinz eMeier |
| author_sort | Dimitri eProbst |
| collection | DOAJ |
| description | The means by which cortical neural networks are able to efficiently solve inference problems remains an open question in computational neuroscience. Recently, abstract models of Bayesian computation in neural circuits have been proposed, but they lack a mechanistic interpretation at the single-cell level. In this article, we describe a complete theoretical framework for building networks of leaky integrate-and-fire neurons that can sample from arbitrary probability distributions over binary random variables. We test our framework for a model inference task based on a psychophysical phenomenon (the Knill-Kersten optical illusion) and further assess its performance when applied to randomly generated distributions. As the local computations performed by the network strongly depend on the interaction between neurons, we compare several types of couplings mediated by either single synapses or interneuron chains. Due to its robustness to substrate imperfections such as parameter noise and background noise correlations, our model is particularly interesting for implementation on novel, neuro-inspired computing architectures, which can thereby serve as a fast, low-power substrate for solving real-world inference problems. |
| first_indexed | 2024-12-11T08:06:59Z |
| format | Article |
| id | doaj.art-79003be675a640d2ab3a40fd6cd6b27c |
| institution | Directory Open Access Journal |
| issn | 1662-5188 |
| language | English |
| last_indexed | 2024-12-11T08:06:59Z |
| publishDate | 2015-02-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Computational Neuroscience |
| spelling | doaj.art-79003be675a640d2ab3a40fd6cd6b27c2022-12-22T01:14:58ZengFrontiers Media S.A.Frontiers in Computational Neuroscience1662-51882015-02-01910.3389/fncom.2015.00013119535Probabilistic Inference in Discrete Spaces Can Be Implemented into Networks of LIF NeuronsDimitri eProbst0Mihai Alexandru Petrovici1Ilja eBytschok2Johannes eBill3Dejan ePecevski4Johannes eSchemmel5Karlheinz eMeier6University of HeidelbergUniversity of HeidelbergUniversity of HeidelbergGraz University of TechnologyGraz University of TechnologyUniversity of HeidelbergUniversity of HeidelbergThe means by which cortical neural networks are able to efficiently solve inference problems remains an open question in computational neuroscience. Recently, abstract models of Bayesian computation in neural circuits have been proposed, but they lack a mechanistic interpretation at the single-cell level. In this article, we describe a complete theoretical framework for building networks of leaky integrate-and-fire neurons that can sample from arbitrary probability distributions over binary random variables. We test our framework for a model inference task based on a psychophysical phenomenon (the Knill-Kersten optical illusion) and further assess its performance when applied to randomly generated distributions. As the local computations performed by the network strongly depend on the interaction between neurons, we compare several types of couplings mediated by either single synapses or interneuron chains. Due to its robustness to substrate imperfections such as parameter noise and background noise correlations, our model is particularly interesting for implementation on novel, neuro-inspired computing architectures, which can thereby serve as a fast, low-power substrate for solving real-world inference problems.http://journal.frontiersin.org/Journal/10.3389/fncom.2015.00013/fulltheoretical neuroscienceNeural codingGraphical ModelsMCMCneuromorphic hardwareBayesian theory |
| spellingShingle | Dimitri eProbst Mihai Alexandru Petrovici Ilja eBytschok Johannes eBill Dejan ePecevski Johannes eSchemmel Karlheinz eMeier Probabilistic Inference in Discrete Spaces Can Be Implemented into Networks of LIF Neurons Frontiers in Computational Neuroscience theoretical neuroscience Neural coding Graphical Models MCMC neuromorphic hardware Bayesian theory |
| title | Probabilistic Inference in Discrete Spaces Can Be Implemented into Networks of LIF Neurons |
| title_full | Probabilistic Inference in Discrete Spaces Can Be Implemented into Networks of LIF Neurons |
| title_fullStr | Probabilistic Inference in Discrete Spaces Can Be Implemented into Networks of LIF Neurons |
| title_full_unstemmed | Probabilistic Inference in Discrete Spaces Can Be Implemented into Networks of LIF Neurons |
| title_short | Probabilistic Inference in Discrete Spaces Can Be Implemented into Networks of LIF Neurons |
| title_sort | probabilistic inference in discrete spaces can be implemented into networks of lif neurons |
| topic | theoretical neuroscience Neural coding Graphical Models MCMC neuromorphic hardware Bayesian theory |
| url | http://journal.frontiersin.org/Journal/10.3389/fncom.2015.00013/full |
| work_keys_str_mv | AT dimitrieprobst probabilisticinferenceindiscretespacescanbeimplementedintonetworksoflifneurons AT mihaialexandrupetrovici probabilisticinferenceindiscretespacescanbeimplementedintonetworksoflifneurons AT iljaebytschok probabilisticinferenceindiscretespacescanbeimplementedintonetworksoflifneurons AT johannesebill probabilisticinferenceindiscretespacescanbeimplementedintonetworksoflifneurons AT dejanepecevski probabilisticinferenceindiscretespacescanbeimplementedintonetworksoflifneurons AT johanneseschemmel probabilisticinferenceindiscretespacescanbeimplementedintonetworksoflifneurons AT karlheinzemeier probabilisticinferenceindiscretespacescanbeimplementedintonetworksoflifneurons |