scholarly journals The human as delta-rule learner.

Decision ◽  
2020 ◽  
Vol 7 (1) ◽  
pp. 55-66
Author(s):  
Sangil Lee ◽  
Joshua I. Gold ◽  
Joseph W. Kable
Keyword(s):  
Delta Rule

scholarly journals 'Tg-.DELTA. Rule' Applied to Semiconducting Vanadate Glasses Containing Different Amounts of Fe2O3.

1999 ◽  
Vol 107 (1245) ◽  
pp. 408-412 ◽  
Author(s):  
Tetsuaki NISHIDA ◽  
Jun IWASHITA ◽  
Shiro KUBUKI
Keyword(s):  
Delta Rule ◽  
Vanadate Glasses

Artificial Neural Networks Used As a Rule Selector for Fuzzy Logic Controllers

1993 ◽  
Author(s):  
Shou-Heng Huang ◽  
Ron M. Nelson
Keyword(s):  
Fuzzy Logic ◽  
Fuzzy Logic Controller ◽  
Absolute Error ◽  
Operating Conditions ◽  
Delta Rule ◽  
Rule Sets ◽  
Artificial Neural ◽  
Artificial Neural Network Ann ◽  
Set Up ◽  
And Control

Abstract A feedforward, three-layer, partially-connected artificial neural network (ANN) is proposed to be used as a rule selector for a rule-based fuzzy logic controller. This will allow the controller to adapt to various control modes and operating conditions for different plants. A principal advantage of an ANN over a look up table is that the ANN can make good estimates to fill in for missing data. The control modes, operating conditions, and control rule sets are encoded into binary numbers as the inputs and outputs for the ANN. The General Delta Rule is used in the backpropagation learning process to update the ANN weights. The proposed ANN has a simple topological structure and results in a simple analysis and relatively easy implementation. The average square error and the maximal absolute error are used to judge if the correct connections between neurons are set up. Computer simulations are used to demonstrate the effectiveness of this ANN as a rule selector.

scholarly journals Learning reward frequency over reward probability: A tale of two learning rules

2018 ◽  
Author(s):  
Hilary Don ◽  
A Ross Otto ◽  
Astin Cornwall ◽  
Tyler Davis ◽  
Darrell A. Worthy
Keyword(s):  
Average Reward ◽  
Binary Outcome ◽  
Delta Rule ◽  
Human Choice ◽  
Presentation Frequency ◽  
Learning Rules ◽  
Expected Values ◽  
Value Learning ◽  
Average Rewards ◽  
Delta Model

Learning about reward and expected values of choice alternatives is critical for adaptive behavior. Although human choice is affected by the presentation frequency of reward-related alternatives, this is overlooked by some dominant models of value learning. For instance, the delta rule learns average rewards, whereas the decay rule learns cumulative rewards for each option. In a binary-outcome choice task, participants selected between pairs of options that had reward probabilities of .65 (A) versus .35 (B) or .75 (C) versus .25 (D). Crucially, during training there were twice as many AB trials as CD trials, therefore option A was associated with higher cumulative reward, while option C gave higher average reward. Participants then decided between novel combinations of options (e.g., AC). Participants preferred option A, a result predicted by the Decay model, but not the Delta model. This suggests that expected values are based more on total reward than average reward.

The Stochastic Delta Rule: Faster and More Accurate Deep Learning Through Adaptive Weight Noise

2020 ◽  
Vol 32 (5) ◽  
pp. 1018-1032 ◽  
Author(s):  
Noah Frazier-Logue ◽  
Stephen José Hanson
Keyword(s):  
Prediction Error ◽  
Saddle Points ◽  
Model Averaging ◽  
Random Variable ◽  
Mean Values ◽  
Delta Rule ◽  
Binomial Random Variable ◽  
Optimal Network ◽  
Training Error ◽  
Feature Discovery

Multilayer neural networks have led to remarkable performance on many kinds of benchmark tasks in text, speech, and image processing. Nonlinear parameter estimation in hierarchical models is known to be subject to overfitting and misspecification. One approach to these estimation and related problems (e.g., saddle points, colinearity, feature discovery) is called Dropout. The Dropout algorithm removes hidden units according to a binomial random variable with probability [Formula: see text] prior to each update, creating random “shocks” to the network that are averaged over updates (thus creating weight sharing). In this letter, we reestablish an older parameter search method and show that Dropout is a special case of this more general model, stochastic delta rule (SDR), published originally in 1990. Unlike Dropout, SDR redefines each weight in the network as a random variable with mean [Formula: see text] and standard deviation [Formula: see text]. Each weight random variable is sampled on each forward activation, consequently creating an exponential number of potential networks with shared weights (accumulated in the mean values). Both parameters are updated according to prediction error, thus resulting in weight noise injections that reflect a local history of prediction error and local model averaging. SDR therefore implements a more sensitive local gradient-dependent simulated annealing per weight converging in the limit to a Bayes optimal network. We run tests on standard benchmarks (CIFAR and ImageNet) using a modified version of DenseNet and show that SDR outperforms standard Dropout in top-5 validation error by approximately 13% with DenseNet-BC 121 on ImageNet and find various validation error improvements in smaller networks. We also show that SDR reaches the same accuracy that Dropout attains in 100 epochs in as few as 40 epochs, as well as improvements in training error by as much as 80%.

Modelling and Optimization of CO2 Absorption in Pneumatic Contactors Using Artificial Neural Networks Developed with Clonal Selection-Based Algorithm

2015 ◽  
Vol 16 (2) ◽  
pp. 97-110 ◽  
Author(s):  
Petronela Cozma ◽  
Elena Niculina Drăgoi ◽  
Ioan Mămăligă ◽  
Silvia Curteanu ◽  
Walter Wukovits ◽  
...  
Keyword(s):  
Neural Network ◽  
Immune System ◽  
Bubble Column ◽  
Clonal Selection ◽  
Back Propagation ◽  
Maximum Efficiency ◽  
Co2 Absorption ◽  
Delta Rule ◽  
Local Search Procedure ◽  
Artificial Neural

AbstractOur research focuses on the application of airlift contactors (ALRs) for the decontamination of CO2-containing gas streams, such as biogas. To assess the performance of ALRs during CO2 absorption, a complex experimental programme was applied in a laboratory-scale rectangular pneumatic contactor, able to operate either as a bubble column or as an airlift reactor. Using the experimental data, a model based on artificial neural network (ANN) was developed. The algorithm for determining the optimal neural network model and for reactor optimization is clonal selection (CS), belonging to artificial immune system class, which is a new computational intelligence paradigm based on the principles of the vertebrate immune system. To improve its capabilities and the probability for highly suitable models and input combinations, addressing maximum efficiency, a Back-Propagation (BK) algorithm – a supervised learning method based on the delta rule – is used as a local search procedure. It is applied in a greedy manner for the best antibody found in each generation. Since the highest affinity antibodies are cloned in the next generation, the effect of BK on the suitability of the individuals propagates into a large proportion of the population. In parallel with the BK hybridization of the basic CS–ANN combination, a series of normalization procedures are included for improving the overall results provided by the new algorithm called nCS-MBK (normalized Clonal Selection-Multilayer Perceptron Neural Network and Back-Propagation algorithm). The optimization allowed for achieving the optimal reactor configuration, which leads to a maximum amount of CO2 dissolved in water.

Delta Rule and Backpropagation

2019 ◽  
pp. 621-633 ◽  
Author(s):  
Leonardo Vanneschi ◽  
Mauro Castelli
Keyword(s):  
Delta Rule
Sign in / Sign up

Export Citation Format

Share Document