China
Africa
Explore chapters and articles related to this topic
Neurons
Published in Nassir H. Sabah, Neuromuscular Fundamentals, 2020
It has been estimated that approximately 20% of the synaptic contacts in the brain are inhibitory. Note that a single-line diagram of neuronal connections as in Figure 7.3 should be interpreted in terms of neuron types rather than literally. Thus, in Figure 7.3a, the incoming excitation may excite a given excitatory neuron, but the inhibitory neuron excited by this incoming excitation may inhibit other excitatory neurons of the same type as the given excitatory neuron. Similarly, in Figure 7.3b, the output of the excitatory neuron excites an inhibitory interneuron, which may then inhibit other excitatory neurons of the same type. In center-surround-inhibition, an excited neuron is surrounded by inhibited neurons of the same type. In lateral inhibition, the inhibited neurons are located mainly laterally with respect to the excited neuron of the same type. Examples of these types of inhibition are given in connection with Figure 12.14.
L
Published in Philip A. Laplante, Comprehensive Dictionary of Electrical Engineering, 2018
latch a small temporary holding cell for a value, the value on the input wires is buffered upon occurrence of some event, such as a clock pulse or rising edge of a separate latch signal. latency (1) total time taken for a bit to pass through the network from origin to destination. (2) the time between positioning a read/write head over a track of data and when the beginning of the track of data passes under the head. lateral (1) a lateral on a primary distribution line is a short tap from the main distribution line which serves a local set of loads. Single phase laterals are common in residential districts. (2) a three-phase or single-phase power line which supplies the distribution transformers along a street. See feeder. lateral inhibition in the human visual system, the inhibitory effect between nearby cells which acts to enhance changes (temporal or spatial) in the stimulus. lateral superlattice refers to a lithographically defined structure in which a periodic (superlattice) potential is induced onto the surface of a normal semiconductor (or metallic) system. Since the periodic potential is induced in the lateral variations of the surface, it is called a lateral superlattice. lateral wave wave generated by a beam of bounded extent incident at an angle close to the critical angle. It manifests by producing a lateral shift of the bounded reflected wave. lattice constant the length of the sides of the three dimensional unit cell in a crystal.
Precise and efficient pose estimation of stacked objects for mobile manipulation in industrial robotics challenges
Published in Advanced Robotics, 2019
Gi Hyun Lim, Nuno Lau, Eurico Pedrosa, Filipe Amaral, Artur Pereira, José Luís Azevedo, Bernardo Cunha
To support single arm manipulation of blank piles, precise perception, especially pose estimation of blank piles, is an integral part of the autonomous packaging system. Based on spatial knowledge such as location of two workbenches in map, the system is required to approach and detect two pallets on a workbench and roughly estimate the pose of blank piles, then to precisely estimate the pose of blank pile for picking up, and to detect the blank magazine in the other workbench for feeding the blank pile smoothly. For fine pose estimation of a blank pile which is tightly aligned with surrounding piles, drawing pin filtering and pinhole filtering processes are performed. The shape of the drawing pin filter and thresholding can be regarded as lateral inhibition of amacrine cells in the retina. They manage the number of active pixels neighboring in the lateral direction.
Computer-aided automatic approach for denoising of magnetic resonance images
Published in Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization, 2021
The network was trained using the feature normalised using two different normalisation. The main path consists of local response normalisation while the parallel path which is weighted residual path consists of group normalisation. The main path creates competition for large activities between neurons response using different kernels which are inspired by real neurons. The LRN uses lateral inhibition, i.e. the capacity of a neuron to reduce the activity of its neighbours. The intention of the lateral inhibition is to carry out local contrast enhancement so that locally maximum pixel values are used as excitation for the next layers (Aqeel Anwar) (Vijendra Singh). The pixel values are square normalised in feature maps within the local neighbourhood. The high-frequency features are detected efficiently by the LRN. It diminishes the response which is uniformly large in the local neighbourhood and performs normalisation in multiple directions. Thus LRN reduces the effect of uniformly distributed Rician noise in the MR images as it has the capability of dampening the common uniform high-frequency effects. The common and uniform high-frequency effects in the neighbourhood (noise) once taken into account are dampened with the use of LRN as noise is uniformly distributed over the range of pixels. The hyperparameters of local response normalisation were fixed as: k = 2, α = 0.1, β = 0.75. The constants k, α and β are hyper-parameters whose values are determined using a validation set (Krizhevsky et al. 2017). K is used to avoid any singularities (division by zero), α is used as a normalisation constant, while β is a contrast constant. In local response normalisation, denotes the activity of a neuron computed by applying kernel i at position (x, y) and then applying the RELU. The response-normalised activity is given by equation 7:
Logics for unconventional computing
Published in International Journal of Parallel, Emergent and Distributed Systems, 2018
Traditionally, symbolic logic was regarded as foundations of mathematics. In the two-volume work Grundlagen der Mathematik (Foundations of Mathematics) written by David Hilbert and Paul Bernays and originally published in 1934 and 1939, there was proposed an axiomatic system as a basic theory for mathematics. In this system, a mathematical proof was considered a discrete process that can be automatized. Since that work by Hilbert and Bernays, all symbolic-logical systems were based on this discrete treatment of mathematical proofs. Nevertheless, in neuroscience each cognition is treated not only as discrete (concentrated on details), but also as analogue (concentrated on the whole picture). The first mechanism to perceive information is called a lateral inhibition and the second a lateral activation. Andrew Schumann and Alexander V. Kuznetsov show in their paper Foundations of Mathematics under Neuroscience Conditions of Lateral Inhibition and Lateral Activation that mathematicians can deal not only with a logical way of automatic proving from some axioms (the lateral inhibition in mathematics), but also with combining proof trees on tree forests by using some analogies as inference metarules (the lateral activation in mathematics). This means that the conventional foundations of mathematics developed by Hilbert and Bernays are focused just on lateral inhibition effects in proof cognitions, but the new foundations are possible, too, which consider proofs as analogue emergent processes (the lateral activation effects in proof cognitions). Hence, symbolic logics in unconventional computing mean that we can change logics themselves to make them more applicable to simulating natural processes. For example, cognitive processes are not only discrete (i.e. they proceed not only in accordance with a lateral inhibition), but also analogue (i.e. they proceed also in accordance with a lateral activation). As a consequence, an appropriate symbolic logic can be either discrete or analogue, as well.