1,720,961 research outputs found
Fuzzy-Rough Cognitive Networks: Theoretical Analysis and Simpler Models
No full textFuzzy-Rough Cognitive Networks (FRCNs) are recurrent neural networks intended for structured classification purposes in which the problem is described by an explicit set of features. The advantage of this granular neural system relies on its transparency and simplicity while being competitive to state-of-the-art classifiers. Despite of their relative empirical success in terms of prediction rates, there are limited studies on FRCNs' dynamic properties and how their building blocks contribute to algorithm's performance. In this paper, we theoretically study these issues and conclude that boundary and negative neurons always converge to a unique fixed-point attractor. Moreover, we demonstrate that negative neurons have no impact on algorithm's performance and that the ranking of positive neurons is invariant. Moved by our theoretical findings, we propose two simpler fuzzy-rough classifiers that overcome the detected issues and maintain the competitive prediction rates of this classifier. Toward the end, we present a case study concerned with image classification in which a Convolutional Neural Network is coupled with one of the simpler models derived from the theoretical analysis of the FRCN model. The numerical simulations suggest that, once the features have been extracted, our granular neural system performs as well as other recurrent neural networks.The authors would like to thank the anonymous reviewers for their valuable and constructive feedback
On the Performance of the Nonsynaptic Backpropagation for Training Long-term Cognitive Networks
No full textLong-term Cognitive Networks (LTCNs) are recurrent neural networks for modeling and simulation. Such networks can be trained in a synaptic or non-synaptic mode according to their goal. Non-synaptic learning refers to adjusting the transfer function parameters while preserving the weights connecting the neurons. In that regard, the Non-synaptic Backpropagation (NSBP) algorithm has proven successful in training LTCN-based 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 non-synaptic 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 non-synaptic learnable parameter that does not contribute significantly to the LTCNs’ performance
Theoretical Analysis of the State Space of Fuzzy Cognitive Maps using Shrink Functions
No full textWe proposed definitions and theorems regarding Fuzzy Cognitive Maps (FCMs), which allow estimating bounds for the activation value of each neuron and analyzing the covering and proximity of feasible activation spaces. The main theoretical findings suggest that the state space of any FCM model equipped with transfer F - functions shrinks infinitely with no guarantee for the FCM to converge to a fixed point but to its limit state space. This result, in conjunction with the covering and proximity values of FCM-based models, helps to
understand their poor performance when solving complex simulation problems
Theoretical Analysis of the State Space of Fuzzy Cognitive Maps using Shrink Functions
No full textWe proposed definitions and theorems regarding Fuzzy Cognitive Maps (FCMs), which allow estimating bounds for the activation value of each neuron and analyzing the covering and proximity of feasible activation spaces. The main theoretical findings suggest that the state space of any FCM model equipped with transfer F - functions shrinks infinitely with no guarantee for the FCM to converge to a fixed point but to its limit state space. This result, in conjunction with the covering and proximity values of FCM-based models, helps to
understand their poor performance when solving complex simulation problems
Framework for Author Name Disambiguation in Scientific Papers Using an Ontological Approach and Deep Learning
No full textThe aim of this paper is to solve the problem of disambiguation of authors’ names in scientific papers. In particular, it focuses on the problem of synonyms and homonyms. Thus, we often find two or more names written in different forms denoting the same person. Moreover, there may be several authors using the same name. To address both the synonym and homonym problems in scientific papers, we propose a framework that uses a hybrid approach of an ontological model and a deep learning model. First, we describe the design of the ontology model, the automatic ontology creation process, and the construction of a weighted co-author network through a set of semantic rules and queries. Second, the selected features are preprocessed during the attribute engineering process to measure the similarity indicator for each feature. Third, the similarity indicators are reduced to a vector space model and used as input to the Deep Learning-based author name disambiguation method to model different types of features. Fourth, the proposed framework is tested on smaller groups of the gold standard large dataset of scientific papers from several international databases named LAGOSAND and achieves promising results compared to other similar solutions proposed in the literature
REPROT: Explaining the predictions of complex deep learning architectures for object detection through reducts of an image
No full textAlthough deep learning models can solve complex prediction problems, they have been criticized for being 'black boxes'. This implies that their decisions are difficult, if not impossible, to explain by simply inspecting their internal knowledge structures. Explainable Artificial Intelligence has attempted to open the black-box through model-specific and agnostic post-hoc methods that generate visualizations or derive associations between the problem features and the model predictions. This paper proposes a new method, termed REPROT, that explains the decisions of complex deep learning architectures based on local reducts of an image. A 'reduct' is a set of sufficiently descriptive features that can fully characterize the acquired knowledge. The created reducts are used to build a 'prototype image' that visually explains the inference obtained by a black-box model for an image. We focus on deep learning architectures whose complexity and internal particularities demand adapting existing model-specific explanation methods, making the explanation process more difficult. Experimental results show that the black-box model can detect an object using the prototype image generated from the reduct. Hence, the explanations will be given by "the minimum set of features sufficient for the neural model to detect an object". The confidence scores obtained by architectures such as Inception, Yolo, and Mask R-CNN are higher for prototype images built from the reduct than those built from the most important superpixels according to the LIME method. Moreover, the target object is not detected on several occasions through the LIME output, thus supporting the superiority of the proposed explanation method.Research funded by MCIN/AEI/10.13039/501100011033/ and FEDER “Una manera de hacer Europa” under grant CONFIA (PID2021-122916NB-I00)
Estimating the limit state space of quasi-nonlinear Fuzzy Cognitive Maps
Full text linkQuasi-Nonlinear Fuzzy Cognitive Maps (q-FCMs) generalize the classic Fuzzy Cognitive Maps (FCMs) by incorporating a nonlinearity coefficient that is related to the model's convergence. While q-FCMs can be configured to avoid unique fixed-point attractors, there is still limited knowledge of their dynamic behavior. In this paper, we propose two iterative, mathematically-driven algorithms that allow estimating the limit state space of any q-FCM model. These algorithms produce accurate lower and upper bounds for the activation values of neural concepts in each iteration without using any information about the initial conditions. As a result, we can determine which activation values will never be produced by a neural concept regardless of the initial conditions used to perform the simulations. In addition, these algorithms could help determine whether a classic FCM model will converge to a unique fixed-point attractor. As a second contribution, we demonstrate that the covering of neural concepts decreases as the nonlinearity coefficient approaches its maximal value. However, large covering values do not necessarily translate into better approximation capabilities, especially in the case of nonlinear problems. This finding points to a trade-off between the model's nonlinearity and the number of reachable states.</p
Learning of Fuzzy Cognitive Map models without training data
No full textThis paper introduces a novel zero-data learning algorithm tailored for Fuzzy Cognitive Map (FCM) models utilized in control applications where we must maintain concepts’ activation values within predefined intervals. Our approach allows domain experts to specify these intervals and optionally impose weight constraints, ensuring the algorithm produces feasible models. At the core of our approach lies a mathematical formalism that approximates the smallest feasible activation space for each neural concept, which translates into lower and upper bounds for concepts’ activation values. Moreover, a parameterized quasi-nonlinear reasoning rule allows controlling whether or not the network converges to a unique fixed point. The learning goal of our algorithm narrows down to computing a weight matrix minimizing the error between the analytical bounds and the target intervals specified by domain experts. To address such a constrained minimization problem, we employ numerical methods operating with approximate gradients, which provide highly accurate solutions with short execution times. The main contribution of our learning algorithm is that it does not require any training data to compute the network structure. Therefore, by accurately approximating the specified activation intervals, our learning algorithm guarantees that the outputs produced by the FCM model will remain within these intervals regardless of the initial conditions used to start the recurrent reasoning process
- …
