A plausible neural circuit for decision making and its formation based on reinforcement learning

Wei, Hui; Dai, Dawei

doi:10.1007/s11571-017-9426-4

Cited by 10 publications

(4 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, "why are some classes easily predicted, while others are not?" (3) We proposed a method to construct a partially understandable neural model by controlling the formation of the neural-path.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Understanding a Deep Neural Network Based on Neural-Path Coding

et al. 2020

Self Cite

View full text Add to dashboard Cite

Recently, the deep neural network has achieved considerable success in the field of machine learning. However, for most tasks, the current neural network models are still considered as the "black-box" structure [1][2], essentially because the knowledge a neural network learns from data is often unpredictable and unexplainable. Similar to neural coding in the brain, one neuron may be simultaneously involved in encoding one or even several tasks. Information may be transmitted between neurons along a particular neural-path (neural circuit) to end neurons that encode a specific decision [3]. Inspired by this, we aim to explore the reason for the performance of a neural model based on a neural-path. First, for a trained neural model, we quantify the neural-path in which the role of a neuron was assumed to control the amount of information that can be passed through; Second, we define a Euclidean distance (ED) for every two neuralpaths, and by analyzing the EDs between two classes of a neural classifier, we explain the ease of prediction for some classes, whereas others are not. We performed extensive experiments for architectures of ResNet and DenseNet models on several benchmark datasets, and determined that the shorter the distance between the neural-paths of two classes, the easier it is to make mistakes. Finally, we proposed a method for controlling the formation of a "neural-path" to build a partially understandable neural model. For the new ResNet model, each feature map in redirect layers was assigned to participate in encoding only one class. The feasibility of the method is also verified through experiments. INDEX TERMS Interpretable neural network, Neural-path, ResNet, DenseNet.

show abstract

Section: Discussionmentioning

confidence: 99%

“…Hide layers (3) Term Vlk[i] denotes the average node value of the i th node in the k th layer in which vlk denotes the node value (activation value) for the sample Xj. Term n denotes the number of samples of class [1].…”

Section: B Define Neural-pathmentioning

confidence: 99%

Understanding a Deep Neural Network Based on Neural-Path Coding

et al. 2020

Self Cite

View full text Add to dashboard Cite

show abstract

“…Inspired from the physiological characteristics of biological neurons, method (Wei et al 2017), proposed the neural circuit based on some control rules, for the development of better decision making.…”

Section: Cognitive Aspects Based Saliency Models For General Objectsmentioning

confidence: 99%

Guiding attention of faces through graph based visual saliency (GBVS)

et al. 2019

View full text Add to dashboard Cite

In a general scenario, while attending a scene containing multiple faces or looking towards a group photograph, our attention does not go equal towards all the faces. It means, we are naturally biased towards some faces. This biasness happens due to availability of dominant perceptual features in those faces. In visual saliency terminology it can be called as 'salient face'. Human's focus their gaze towards a face which carries the 'dominating look' in the crowd. This happens due to comparative saliency of the faces. Saliency of a face is determined by its feature dissimilarity with the surrounding faces. In this context there is a big role of human psychology and its cognitive science too. Therefore, enormous researches have been carried out towards modeling the computer vision system like human's vision. This paper proposed a graphical based bottom up approach to point up the salient face in the crowd or in an image having multiple faces. In this novel method, visual saliencies of faces have been calculated based on the intensity values, facial areas and their relative spatial distances. Experiment has been conducted on gray scale images. In order to verify this experiment, three level of validation has been done. In the first level, our results have been verified with the prepared ground truth. In the second level, intensity scores of proposed saliency maps have been cross verified with the saliency score. In the third level, saliency map is validated with some standard parameters. The results are found to be interesting and in some aspects saliency predictions are like human vision system. The evaluation made with the proposed approach shows moderately boost up results and hence, this idea can be useful in the future modeling of intelligent vision (robot vision) system.

show abstract

“…Most of these studies focus either on non‐associative learning, that is, habituation (Hasani et al, 2017) or on improved chemotaxis (Demin & Vityaev, 2014), often without taking into account the biological structure of the governing neuronal circuits. However, mathematical models of neuronal circuits have been recruited to explore learning in other systems (Garst‐Orozco et al, 2014; Wei et al, 2017) or as stand‐alone computational work (Maass et al, 2007).…”

Section: Introductionmentioning

confidence: 99%

Modelling learning in Caenorhabditis elegans chemosensory and locomotive circuitry for T‐maze navigation

Sakelaris

Sun

et al. 2022

Eur J of Neuroscience

View full text Add to dashboard Cite

Recently, a new type of Caenorhabditis elegans associative learning was reported, where nematodes learn to reach a target arm in an empty T-maze, after they have successfully located reward (food) in the same side arm of a similar, baited, training maze. Here, we present a simplified mathematical model of C. elegans chemosensory and locomotive circuitry that replicates C. elegans navigation in a T-maze and predicts the underlying mechanisms generating maze learning. Based on known neural circuitry, the model circuit responds to food-released chemical cues by modulating motor neuron activity that drives simulated locomotion. We show that, through modulation of interneuron activity, such a circuit can mediate maze learning by acquiring a turning bias, even after a single training session. Simulated nematode maze navigation during training conditions in food-baited mazes and during testing conditions in empty mazes is validated by comparing simulated behaviour with new experimental video data, extracted through the implementation of a custom-made maze tracking algorithm. Our work provides a mathematical framework for investigating the neural mechanisms underlying this novel learning behaviour in C. elegans. Model results predict neuronal components involved in maze and spatial learning and identify target neurons and potential neural mechanisms for future experimental investigations into this learning behaviour.

show abstract

A plausible neural circuit for decision making and its formation based on reinforcement learning

Cited by 10 publications

References 40 publications

Understanding a Deep Neural Network Based on Neural-Path Coding

Understanding a Deep Neural Network Based on Neural-Path Coding

Guiding attention of faces through graph based visual saliency (GBVS)

Modelling learning in Caenorhabditis elegans chemosensory and locomotive circuitry for T‐maze navigation

Contact Info

Product

Resources

About