Nonsynaptic Error Backpropagation in Long-Term Cognitive Networks

Nopoles, Gonzalo; Vanhoenshoven, Frank; Falcón, Rafael; Vanhoof, Koen

doi:10.1109/tnnls.2019.2910555

Cited by 18 publications

(11 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this model, each neural concept C i denotes a domain variable while weights w ji ∈ R represent the rate of change in the conditional mean of C i with respect to C j , assuming that the activation values of the remaining neurons impacting C i are fixed [4]. Hidden neurons are not allowed, as they cannot be interpreted naturally.…”

Section: Network Constructionmentioning

confidence: 99%

“…i , whereas the weights w ij and bounds L i and U i remain unaltered [4]. Overall, the learning task consists of adjusting the shape of the transfer function associated with the i-th neuron in each iteration.…”

Section: Nonsynaptic Learningmentioning

confidence: 99%

“…This principle allows experts to inject domain knowledge directly into the network in the form of weights or constraints on the neurons' activation values, yet the model is able to produce low simulation errors. Recently, the authors in [4] presented a new method termed Long-term Cognitive Networks (LTCNs) to handle longer dependencies among the problem variables. In this neural cognitive model, each neuron's transfer function parameters are learned through a Nonsynaptic Backpropagation (NSBP) algorithm.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

On the Performance of the Nonsynaptic Backpropagation for Training Long-term Cognitive Networks

et al. 2021

View full text Add to dashboard Cite

Long-term Cognitive Networks (LTCNs) are recurrent neural networks for modeling and simulation. Such networks can be trained in a synaptic or nonsynaptic mode according to their goal. Nonsynaptic learning refers to adjusting the transfer function parameters while preserving the weights connecting the neurons. In that regard, the Nonsynaptic Backpropagation (NSBP) algorithm has proven successful in training LTCNbased models. Despite NSBP's success, a question worthy of investigation is whether the backpropagation process is necessary when training these recurrent neural networks. This paper investigates this issue and presents three nonsynaptic learning methods that modify the original algorithm. In addition, we perform a sensitivity analysis of both the NSBP's hyperparameters and the LTCNs' learnable parameters. The main conclusions of our study are i) the backward process attached to the NSBP algorithm is not necessary to train these recurrent neural systems, and ii) there is a nonsynaptic learnable parameter that does not contribute significantly to the LTCNs' performance.

show abstract

Section: Network Constructionmentioning

confidence: 99%

Section: Nonsynaptic Learningmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

On the Performance of the Nonsynaptic Backpropagation for Training Long-term Cognitive Networks

et al. 2021

View full text Add to dashboard Cite

show abstract

“…The LTCN model was introduced in [14] to overcome these issues. In short, LTCNs are neither causal nor fuzzy, and their weights can take values in the real domain.…”

Section: Long-term Cognitive Networkmentioning

confidence: 99%

“…To tackle the last two issues, Nápoles et al [14] introduced the Long-term Cognitive Networks (LTCNs). In this FCM-like model, the weights are not constrained to any specific interval, and the learnable parameters are computed using a non-synaptic backpropagation algorithm.…”

Section: Introductionmentioning

confidence: 99%

Recurrence-Aware Long-Term Cognitive Network for Explainable Pattern Classification

Nápoles¹,

Salgueiro²,

Grau³

et al. 2021

Preprint

View full text Add to dashboard Cite

Machine learning solutions for pattern classification problems are nowadays widely deployed in society and industry. However, the lack of transparency and accountability of most accurate models often hinders their meaningful and safe use. Thus, there is a clear need for developing explainable artificial intelligence mechanisms. There exist model-agnostic methods that summarize feature contributions, but their interpretability is limited to specific predictions made by black-box models. An open challenge is to develop models that have intrinsic interpretability and produce their own explanations, even for classes of models that are traditionally considered black boxes like (recurrent) neural networks. In this paper, we propose an LTCNbased model for interpretable pattern classification of structured data. Our method brings its own mechanism for providing explanations by quantifying the relevance of each feature in the decision process. For supporting the interpretability without affecting the performance, the model incorporates more flexibility through a quasi-nonlinear reasoning rule that allows controlling nonlinearity. Besides, we propose a recurrence-aware decision model that evades the issues posed by unique fixed points while introducing a deterministic learning method to compute the learnable parameters. The simulations show that our interpretable model obtains competitive performance when compared to the stateof-the-art white and black boxes.

show abstract

Evolving Recurrent Neural Networks for Pattern Classification

Nápoles

2020

Advances in Intelligent Systems and Computing

View full text Add to dashboard Cite

Nonsynaptic Error Backpropagation in Long-Term Cognitive Networks

Cited by 18 publications

References 20 publications

On the Performance of the Nonsynaptic Backpropagation for Training Long-term Cognitive Networks

On the Performance of the Nonsynaptic Backpropagation for Training Long-term Cognitive Networks

Recurrence-Aware Long-Term Cognitive Network for Explainable Pattern Classification

Evolving Recurrent Neural Networks for Pattern Classification

Contact Info

Product

Resources

About