scholarly journals How and Why the Cerebellum Recodes Input Signals: An Alternative to Machine Learning

2021 ◽  
pp. 107385842098679
Author(s):  
Mike Gilbert ◽  
R. Chris Miall
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Purkinje Cells ◽  
Granule Cells ◽  
Learning Algorithm ◽  
Mossy Fiber ◽  
High Dimensional ◽  
Input Signals ◽  
Compute Data ◽  
The Mean

Mossy fiber input to the cerebellum is received by granule cells where it is thought to be recoded into internal signals received by Purkinje cells, which alone carry the output of the cerebellar cortex. In any neural network, variables are contained in groups of signals as well as signals themselves—which cells are active and how many, for example, and statistical variables coded in rates, such as the mean and range, and which rates are strongly represented, in a defined population. We argue that the primary function of recoding is to confine translation to an effect of some variables and not others—both where input is recoded into internal signals and the translation downstream of internal signals into an effect on Purkinje cells. The cull of variables is harsh. Internal signaling is group coded. This allows coding to exploit statistics for a reliable and precise effect despite needing to work with high-dimensional input which is a highly unpredictably variable. An important effect is to normalize eclectic input signals, so that the basic, repeating cerebellar circuit, preserved across taxa, does not need to specialize (within regional variations). With this model, there is no need to slavishly conserve or compute data coded in single signals. If we are correct, a learning algorithm—for years, a mainstay of cerebellar modeling—would be redundant.

scholarly journals Unraveling the deep learning gearbox in optical coherence tomography image segmentation towards explainable artificial intelligence

2021 ◽  
Vol 4 (1) ◽  
Author(s):  
Peter M. Maloca ◽  
Philipp L. Müller ◽  
Aaron Y. Lee ◽  
Adnan Tufail ◽  
Konstantinos Balaskas ◽  
...  
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Optical Coherence Tomography ◽  
Image Segmentation ◽  
Convolutional Neural Network ◽  
Learning Algorithm ◽  
Ground Truth ◽  
Optical Coherence Tomography Image ◽  
Optical Coherence ◽  
Tomography Image

AbstractMachine learning has greatly facilitated the analysis of medical data, while the internal operations usually remain intransparent. To better comprehend these opaque procedures, a convolutional neural network for optical coherence tomography image segmentation was enhanced with a Traceable Relevance Explainability (T-REX) technique. The proposed application was based on three components: ground truth generation by multiple graders, calculation of Hamming distances among graders and the machine learning algorithm, as well as a smart data visualization (‘neural recording’). An overall average variability of 1.75% between the human graders and the algorithm was found, slightly minor to 2.02% among human graders. The ambiguity in ground truth had noteworthy impact on machine learning results, which could be visualized. The convolutional neural network balanced between graders and allowed for modifiable predictions dependent on the compartment. Using the proposed T-REX setup, machine learning processes could be rendered more transparent and understandable, possibly leading to optimized applications.

Automating Well Log Correlation Workflow Using Soft Attention Convolutional Neural Networks

2021 ◽  
Author(s):  
Aria Abubakar ◽  
Mandar Kulkarni ◽  
Anisha Kaul
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Convolutional Neural Network ◽  
Gamma Ray ◽  
Learning Algorithm ◽  
Machine Learning Algorithm ◽  
Well Log ◽  
Machine Learning Model ◽  
Well Correlation ◽  
Geological Formations

Abstract In the process of deriving the reservoir petrophysical properties of a basin, identifying the pay capability of wells by interpreting various geological formations is key. Currently, this process is facilitated and preceded by well log correlation, which involves petrophysicists and geologists examining multiple raw log measurements for the well in question, indicating geological markers of formation changes and correlating them with those of neighboring wells. As it may seem, this activity of picking markers of a well is performed manually and the process of ‘examining’ may be highly subjective, thus, prone to inconsistencies. In our work, we propose to automate the well correlation workflow by using a Soft- Attention Convolutional Neural Network to predict well markers. The machine learning algorithm is supervised by examples of manual marker picks and their corresponding occurrence in logs such as gamma-ray, resistivity and density. Our experiments have shown that, specifically, the attention mechanism allows the Convolutional Neural Network to look at relevant features or patterns in the log measurements that suggest a change in formation, making the machine learning model highly precise.

scholarly journals Efficient generation of reciprocal signals by inhibition

2012 ◽  
Vol 107 (9) ◽  
pp. 2453-2462 ◽  
Author(s):  
Sung-min Park ◽  
Esra Tara ◽  
Kamran Khodakhah
Keyword(s):  
Purkinje Cells ◽  
Granule Cell ◽  
Granule Cells ◽  
Mossy Fiber ◽  
Cell Activity ◽  
Common Mechanism ◽  
Input Output ◽  
Firing Rates ◽  
Neuronal Computation ◽  
The Brain

Reciprocal activity between populations of neurons has been widely observed in the brain and is essential for neuronal computation. The different mechanisms by which reciprocal neuronal activity is generated remain to be established. A common motif in neuronal circuits is the presence of afferents that provide excitation to one set of principal neurons and, via interneurons, inhibition to a second set of principal neurons. This circuitry can be the substrate for generation of reciprocal signals. Here we demonstrate that this equivalent circuit in the cerebellar cortex enables the reciprocal firing rates of Purkinje cells to be efficiently generated from a common set of mossy fiber inputs. The activity of a mossy fiber is relayed to Purkinje cells positioned immediately above it by excitatory granule cells. The firing rates of these Purkinje cells increase as a linear function of mossy fiber, and thus granule cell, activity. In addition to exciting Purkinje cells positioned immediately above it, the activity of a mossy fiber is relayed to laterally positioned Purkinje cells by a disynaptic granule cell → molecular layer interneuron pathway. Here we show in acutely prepared cerebellar slices that the input-output relationship of these laterally positioned Purkinje cells is linear and reciprocal to the first set. A similar linear input-output relationship between decreases in Purkinje cell firing and strength of stimulation of laterally positioned granule cells was also observed in vivo. Use of interneurons to generate reciprocal firing rates may be a common mechanism by which the brain generates reciprocal signals.

scholarly journals CAN A MACHINE LEARNING ALGORITHM IDENTIFY SARS-COV-2 VARIANTS BASED ON CONVENTIONAL rRT-PCR? PROOF OF CONCEPT

2021 ◽  
Author(s):  
jorge cabrera Alvargonzalez ◽  
Ana Larranaga Janeiro ◽  
Sonia Perez ◽  
Javier Martinez Torres ◽  
Lucia martinez lamas ◽  
...  
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Supervised Classification ◽  
Learning Algorithm ◽  
Support Vector ◽  
Classification Algorithms ◽  
Machine Learning Algorithm ◽  
Proof Of Concept ◽  
The Past ◽  
Number Of Cycles

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has been and remains one of the major challenges humanity has faced thus far. Over the past few months, large amounts of information have been collected that are only now beginning to be assimilated. In the present work, the existence of residual information in the massive numbers of rRT-PCRs that tested positive out of the almost half a million tests that were performed during the pandemic is investigated. This residual information is believed to be highly related to a pattern in the number of cycles that are necessary to detect positive samples as such. Thus, a database of more than 20,000 positive samples was collected, and two supervised classification algorithms (a support vector machine and a neural network) were trained to temporally locate each sample based solely and exclusively on the number of cycles determined in the rRT-PCR of each individual. Finally, the results obtained from the classification show how the appearance of each wave is coincident with the surge of each of the variants present in the region of Galicia (Spain) during the development of the SARS-CoV-2 pandemic and clearly identified with the classification algorithm.

scholarly journals A Comparison of Machine Learning Methods in a High-Dimensional Classification Problem

2014 ◽  
Vol 5 (3) ◽  
pp. 82-96 ◽  
Author(s):  
Marijana Zekić-Sušac ◽  
Sanja Pfeifer ◽  
Nataša Šarlija
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Classification Accuracy ◽  
Classification Problem ◽  
High Dimensional ◽  
Nearest Neighbour ◽  
Learning Methods ◽  
Machine Learning Methods ◽  
Dimensional Classification ◽  
Artificial Neural

Abstract Background: Large-dimensional data modelling often relies on variable reduction methods in the pre-processing and in the post-processing stage. However, such a reduction usually provides less information and yields a lower accuracy of the model. Objectives: The aim of this paper is to assess the high-dimensional classification problem of recognizing entrepreneurial intentions of students by machine learning methods. Methods/Approach: Four methods were tested: artificial neural networks, CART classification trees, support vector machines, and k-nearest neighbour on the same dataset in order to compare their efficiency in the sense of classification accuracy. The performance of each method was compared on ten subsamples in a 10-fold cross-validation procedure in order to assess computing sensitivity and specificity of each model. Results: The artificial neural network model based on multilayer perceptron yielded a higher classification rate than the models produced by other methods. The pairwise t-test showed a statistical significance between the artificial neural network and the k-nearest neighbour model, while the difference among other methods was not statistically significant. Conclusions: Tested machine learning methods are able to learn fast and achieve high classification accuracy. However, further advancement can be assured by testing a few additional methodological refinements in machine learning methods.

Android-Based Skin Cancer Recognition System Using Convolutional Neural Network

2021 ◽  
pp. 59-85
Author(s):  
Sercan Demirci ◽  
Durmuş Özkan Şahin ◽  
Ibrahim Halil Toprak
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Skin Cancer ◽  
Early Diagnosis ◽  
Learning Algorithm ◽  
Recognition System ◽  
Machine Learning Algorithm ◽  
The Sun ◽  
Human Beings ◽  
The World

Skin cancer, which is one of the most common types of cancer in the world, is a malignant growth seen on the skin due to various reasons. There was an increase in the number of the cases of skin cancer nearly 200% between 2004-2009. Since the ozone layer is depleting, harmful rays reflected from the sun cannot be filtered. In this case, the likelihood of skin cancer will increase over the years and pose more risks for human beings. Early diagnosis is very significant as in all types of cancers. In this study, a mobile application is developed in order to detect whether the skin spots photographed by using the machine learning technique for early diagnosis have a suspicion of skin cancer. Thus, an auxiliary decision support system is developed that can be used both by the clinicians and individuals. For cases that are predicted to have a risk higher than a certain rate by the machine learning algorithm, early diagnosis could be initiated for the patients by consulting a physician when the case is considered to have a higher risk by machine learning algorithm.

A Framework for an Artificial-Neural-Network-Based Electronic Nose

2022 ◽  
pp. 350-374
Author(s):  
Mudassir Ismail ◽  
Ahmed Abdul Majeed ◽  
Yousif Abdullatif Albastaki
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Artificial Neural Network ◽  
Electronic Nose ◽  
Learning Algorithm ◽  
General Purpose ◽  
Food Spoilage ◽  
Standard Data ◽  
Semiconductor Gas Sensor ◽  
Artificial Neural

Machine odor detection has developed into an important aspect of our lives with various applications of it. From detecting food spoilage to diagnosis of diseases, it has been developed and tested in various fields and industries for specific purposes. This project, artificial-neural-network-based electronic nose (ANNeNose), is a machine-learning-based e-nose system that has been developed for detection of various types of odors for a general purpose. The system can be trained on any odor using various e-nose sensor types. It uses artificial neural network as its machine learning algorithm along with an OMX-GR semiconductor gas sensor for collecting odor data. The system was trained and tested with five different types of odors collected through a standard data collection method and then purified, which in turn had a result varying from 93% to 100% accuracy.

scholarly journals Interpretation of Neural Networks Is Fragile

2019 ◽  
Vol 33 ◽  
pp. 3681-3688 ◽  
Author(s):  
Amirata Ghorbani ◽  
Abubakar Abid ◽  
James Zou
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Neural Networks ◽  
Input Data ◽  
Learning Algorithm ◽  
Hessian Matrix ◽  
Machine Learning Algorithm ◽  
Feature Importance ◽  
Adversarial Attack ◽  
Measurement Biases

In order for machine learning to be trusted in many applications, it is critical to be able to reliably explain why the machine learning algorithm makes certain predictions. For this reason, a variety of methods have been developed recently to interpret neural network predictions by providing, for example, feature importance maps. For both scientific robustness and security reasons, it is important to know to what extent can the interpretations be altered by small systematic perturbations to the input data, which might be generated by adversaries or by measurement biases. In this paper, we demonstrate how to generate adversarial perturbations that produce perceptively indistinguishable inputs that are assigned the same predicted label, yet have very different interpretations. We systematically characterize the robustness of interpretations generated by several widely-used feature importance interpretation methods (feature importance maps, integrated gradients, and DeepLIFT) on ImageNet and CIFAR-10. In all cases, our experiments show that systematic perturbations can lead to dramatically different interpretations without changing the label. We extend these results to show that interpretations based on exemplars (e.g. influence functions) are similarly susceptible to adversarial attack. Our analysis of the geometry of the Hessian matrix gives insight on why robustness is a general challenge to current interpretation approaches.

Sign in / Sign up

Export Citation Format

Share Document