Black-box and surrogate optimization for tuning spiking neural models of striatum plasticity
The basal ganglia (BG) is a brain structure that has long been proposed to play an essential role in action selection, and theoretical models of spiking neurons have tried to explain how the BG solves this problem. A recently proposed functional and biologically inspired network model of the striatu...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2022-10-01
|
| Series: | Frontiers in Neuroinformatics |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fninf.2022.1017222/full |
| _version_ | 1811340203292360704 |
|---|---|
| author | Nicolás C. Cruz Álvaro González-Redondo Juana L. Redondo Jesús A. Garrido Eva M. Ortigosa Pilar M. Ortigosa |
| author_facet | Nicolás C. Cruz Álvaro González-Redondo Juana L. Redondo Jesús A. Garrido Eva M. Ortigosa Pilar M. Ortigosa |
| author_sort | Nicolás C. Cruz |
| collection | DOAJ |
| description | The basal ganglia (BG) is a brain structure that has long been proposed to play an essential role in action selection, and theoretical models of spiking neurons have tried to explain how the BG solves this problem. A recently proposed functional and biologically inspired network model of the striatum (an important nucleus of the BG) is based on spike-timing-dependent eligibility (STDE) and captured important experimental features of this nucleus. The model can recognize complex input patterns and consistently choose rewarded actions to respond to such sensory inputs. However, model tuning is challenging due to two main reasons. The first is the expert knowledge required, resulting in tedious and potentially biased trial-and-error procedures. The second is the computational cost of assessing model configurations (approximately 1.78 h per evaluation). This study addresses the model tuning problem through numerical optimization. Considering the cost of assessing solutions, the selected methods stand out due to their low requirements for solution evaluations and compatibility with high-performance computing. They are the SurrogateOpt solver of Matlab and the RBFOpt library, both based on radial basis function approximations, and DIRECT-GL, an enhanced version of the widespread black-box optimizer DIRECT. Besides, a parallel random search serves as a baseline reference of the outcome of opting for sophisticated methods. SurrogateOpt turns out to be the best option for tuning this kind of model. It outperforms, on average, the quality of the configuration found by an expert and works significantly faster and autonomously. RBFOpt and the random search share the second position, but their average results are below the option found by hand. Finally, DIRECT-GL follows this line becoming the worst-performing method. |
| first_indexed | 2024-04-13T18:38:28Z |
| format | Article |
| id | doaj.art-ef7a5b26480d42f68104f4302ec5f511 |
| institution | Directory Open Access Journal |
| issn | 1662-5196 |
| language | English |
| last_indexed | 2024-04-13T18:38:28Z |
| publishDate | 2022-10-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Neuroinformatics |
| spelling | doaj.art-ef7a5b26480d42f68104f4302ec5f5112022-12-22T02:34:49ZengFrontiers Media S.A.Frontiers in Neuroinformatics1662-51962022-10-011610.3389/fninf.2022.10172221017222Black-box and surrogate optimization for tuning spiking neural models of striatum plasticityNicolás C. Cruz0Álvaro González-Redondo1Juana L. Redondo2Jesús A. Garrido3Eva M. Ortigosa4Pilar M. Ortigosa5Department of Computer Engineering, Automation and Robotics, University of Granada, Granada, SpainDepartment of Computer Engineering, Automation and Robotics, University of Granada, Granada, SpainDepartment of Informatics, University of Almería, ceiA3 Excellence Agri-food Campus, Almeria, SpainDepartment of Computer Engineering, Automation and Robotics, University of Granada, Granada, SpainDepartment of Computer Engineering, Automation and Robotics, University of Granada, Granada, SpainDepartment of Informatics, University of Almería, ceiA3 Excellence Agri-food Campus, Almeria, SpainThe basal ganglia (BG) is a brain structure that has long been proposed to play an essential role in action selection, and theoretical models of spiking neurons have tried to explain how the BG solves this problem. A recently proposed functional and biologically inspired network model of the striatum (an important nucleus of the BG) is based on spike-timing-dependent eligibility (STDE) and captured important experimental features of this nucleus. The model can recognize complex input patterns and consistently choose rewarded actions to respond to such sensory inputs. However, model tuning is challenging due to two main reasons. The first is the expert knowledge required, resulting in tedious and potentially biased trial-and-error procedures. The second is the computational cost of assessing model configurations (approximately 1.78 h per evaluation). This study addresses the model tuning problem through numerical optimization. Considering the cost of assessing solutions, the selected methods stand out due to their low requirements for solution evaluations and compatibility with high-performance computing. They are the SurrogateOpt solver of Matlab and the RBFOpt library, both based on radial basis function approximations, and DIRECT-GL, an enhanced version of the widespread black-box optimizer DIRECT. Besides, a parallel random search serves as a baseline reference of the outcome of opting for sophisticated methods. SurrogateOpt turns out to be the best option for tuning this kind of model. It outperforms, on average, the quality of the configuration found by an expert and works significantly faster and autonomously. RBFOpt and the random search share the second position, but their average results are below the option found by hand. Finally, DIRECT-GL follows this line becoming the worst-performing method.https://www.frontiersin.org/articles/10.3389/fninf.2022.1017222/fullmodel tuningsurrogate optimizationblack-box optimizationstriatumreinforcement learningspiking neural networks |
| spellingShingle | Nicolás C. Cruz Álvaro González-Redondo Juana L. Redondo Jesús A. Garrido Eva M. Ortigosa Pilar M. Ortigosa Black-box and surrogate optimization for tuning spiking neural models of striatum plasticity Frontiers in Neuroinformatics model tuning surrogate optimization black-box optimization striatum reinforcement learning spiking neural networks |
| title | Black-box and surrogate optimization for tuning spiking neural models of striatum plasticity |
| title_full | Black-box and surrogate optimization for tuning spiking neural models of striatum plasticity |
| title_fullStr | Black-box and surrogate optimization for tuning spiking neural models of striatum plasticity |
| title_full_unstemmed | Black-box and surrogate optimization for tuning spiking neural models of striatum plasticity |
| title_short | Black-box and surrogate optimization for tuning spiking neural models of striatum plasticity |
| title_sort | black box and surrogate optimization for tuning spiking neural models of striatum plasticity |
| topic | model tuning surrogate optimization black-box optimization striatum reinforcement learning spiking neural networks |
| url | https://www.frontiersin.org/articles/10.3389/fninf.2022.1017222/full |
| work_keys_str_mv | AT nicolasccruz blackboxandsurrogateoptimizationfortuningspikingneuralmodelsofstriatumplasticity AT alvarogonzalezredondo blackboxandsurrogateoptimizationfortuningspikingneuralmodelsofstriatumplasticity AT juanalredondo blackboxandsurrogateoptimizationfortuningspikingneuralmodelsofstriatumplasticity AT jesusagarrido blackboxandsurrogateoptimizationfortuningspikingneuralmodelsofstriatumplasticity AT evamortigosa blackboxandsurrogateoptimizationfortuningspikingneuralmodelsofstriatumplasticity AT pilarmortigosa blackboxandsurrogateoptimizationfortuningspikingneuralmodelsofstriatumplasticity |