Instance hardness measures allow one to characterize and understand why some instances are harder to classify than others in a classification dataset. An instance can be hard to classify for different reasons, such as being in an overlapping region of the classes or a region of poor data representativeness. While there are many instance hardness measures in the related literature, they are mainly concerned with measuring class overlap. This paper also addresses measuring sparsity in a dataset by building a proximity graph from data and extracting some network-based measures from the nodes. Experimentally, we show that some of these measures are effective in characterizing instance hardness and complement the ones from the literature by measuring the density of the regions where the instances are located.

We increasingly integrate technology into our daily activities, and using Machine Learning (ML) algorithms in various domains has become a common practice. However, in crucial sectors where algorithmic decisions significantly impact people's lives, there is a need to scrutinize these decisions more carefully. Using these algorithms in critical areas, such as courtrooms, raises concerns about potential bias and prejudice, directly affecting the justice and partiality of these tools. There is an urge to create algorithms supporting ethical decisions. This paper proposes using data balancing techniques to mitigate the sample bias present in datasets, aiming to make subsequent ML algorithm training more impartial. A version of the ADASYN algorithm is developed, which performs data balancing at both the class level and at the level of protected attributes, enhancing the diversity and representativeness of the protected groups in the datasets. Experimental results show the technique can promote greater fairness in the predictions of different ML models while keeping a good trade-off with overall accuracy.

The optimization of machine learning models is an open research question since an optimal procedure to change the learning rate throughout training is still unknown. Manually defining a learning rate schedule involves troublesome, time-consuming try-and-error procedures to determine hyperparameters, such as learning rate decay epochs and learning rate decay rates, in order to navigate the intricate landscape of loss functions. Although adaptive learning rate optimizers automatize this process, recent studies suggest they may produce overfitting and reduce performance compared to fine-tuned learning rate schedules. Considering that the new machine learning loss function approaches present landscapes with much more saddle points than local minima, we proposed the Training Aware Sigmoidal Optimizer (TASO). This automated two-phase learning rate adaptation mechanism significantly reduces the need for manual hyperparameter tuning. The first phase uses a high learning rate to quickly traverse the numerous saddle points in the error surface, while the second phase uses a low learning rate to gradually approach the center of the local minimum previously found. We compared the proposed approach with commonly used adaptive learning rate schedules such as Adam, RMSProp, and Adagrad. The validation experiments were performed in image and text datasets and showed that TASO outperformed all competing methods in optimal (i.e., performing hyperparameter validation) and suboptimal (i.e., using default hyperparameters) scenarios. In our benchmark tests, TASO demonstrated promising performance, achieving an 8.32% increase in accuracy and a significant 46.62% decrease in training loss on average across various datasets and models. This remarkable performance positions TASO ahead of well-established adaptive optimizers, suggesting higher effectiveness and consistency in performance.

Reinforcement Learning (RL) provides a powerful framework for decision-making in complex environments. However, implementing RL in hardware-efficient and bio-inspired ways remains a challenge. This paper presents a novel Spiking Neural Network (SNN) architecture for solving RL problems with real-valued observations. The proposed model incorporates multi-layered event-based clustering, with the addition of Temporal Difference (TD)-error modulation and eligibility traces, building upon prior work. An ablation study confirms the significant impact of these components on the proposed model's performance. A tabular actor-critic algorithm with eligibility traces and a state-of-the-art Proximal Policy Optimization (PPO) algorithm are used as benchmarks. Our network consistently outperforms the tabular approach and successfully discovers stable control policies on classic RL environments: mountain car, cart-pole, and acrobot. The proposed model offers an appealing trade-off in terms of computational and hardware implementation requirements. The model does not require an external memory buffer nor a global error gradient computation, and synaptic updates occur online, driven by local learning rules and a broadcasted TD-error signal. Thus, this work contributes to the development of more hardware-efficient RL solutions.

Parameter control methods for metaheuristics with reinforcement learning put forward so far usually present the following shortcomings: (1) Their training processes are usually highly time-consuming and they are not able to benefit from parallel or distributed platforms; (2) they are usually sensitive to their hyperparameters, which means that the quality of the final results is heavily dependent on their values; (3) and limited benchmarks have been used to assess their generality. This paper addresses these issues by proposing a methodology for training out-of-the-box parameter control policies for mono-objective non-niching evolutionary and swarm-based algorithms using distributed reinforcement learning with population-based training. The proposed methodology is suitable to be used in any mono-objective optimization problem and for any mono-objective and non-niching Evolutionary and swarm-based algorithm. The results in this paper achieved through extensive experiments show that the proposed method satisfactorily improves all the aforementioned issues, overcoming constant, random and human-designed policies in several different scenarios.

This paper presents an overview of research in Weightless Neural Models. Weightless Neural Networks (WNNs) do not have weighted connections between nodes. They use a different kind of neuron, usually based on RAM memory devices. The weightless (or RAM-based) neural networks are based on variations of the RAM node proposed by Aleksander. Recent works in the literature, such as the quantum weightless neuron, provide a novel perspective to this area. The paper describes classical and quantum weightless models and important recent works found in the literature, pointing out the challenges and future directions in the area.

In this article, we argue that the unsatisfactory out-of-distribution (OOD) detection performance of neural networks is mainly due to the SoftMax loss anisotropy and propensity to produce low entropy probability distributions in disagreement with the principle of maximum entropy. On the one hand, current OOD detection approaches usually do not directly fix the SoftMax loss drawbacks, but rather build techniques to circumvent it. Unfortunately, those methods usually produce undesired side effects (e.g., classification accuracy drop, additional hyperparameters, slower inferences, and collecting extra data). On the other hand, we propose replacing SoftMax loss with a novel loss function that does not suffer from the mentioned weaknesses. The proposed IsoMax loss is isotropic (exclusively distance-based) and provides high entropy posterior probability distributions. Replacing the SoftMax loss by IsoMax loss requires no model or training changes. Additionally, the models trained with IsoMax loss produce as fast and energy-efficient inferences as those trained using SoftMax loss. Moreover, no classification accuracy drop is observed. The proposed method does not rely on outlier/background data, hyperparameter tuning, temperature calibration, feature extraction, metric learning, adversarial training, ensemble procedures, or generative models. Our experiments showed that IsoMax loss works as a seamless SoftMax loss drop-in replacement that significantly improves neural networks' OOD detection performance. Hence, it may be used as a baseline OOD detection approach to be combined with current or future OOD detection techniques to achieve even higher results.

The model selection procedure is usually a single-criterion decision making in which we select the model that maximizes a specific metric in a specific set, such as the Validation set performance. We claim this is very naive and can perform poor selections of over-fitted models due to the over-searching phenomenon, which over-estimates the performance on that specific set. Futhermore, real world data contains noise that should not be ignored by the model selection procedure and must be taken into account when performing model selection. Also, we have defined four theoretical optimality conditions that we can pursue to better select the models and analyze them by using a multi-criteria decision-making algorithm (TOPSIS) that considers proxies to the optimality conditions to select reasonable models.

As solar generation gains an increased importance in a smart grid, an efficient control of a photovoltaic (PV) array has to be considered. However, energy efficiency of a PV array greatly depends on environmental conditions, such as uneven shading, solar irradiance and temperature. Maximum power point tracking (MPPT) algorithms aim to dynamically find an optimal operation voltage in order to compensate for changes in the environment, as well as degradation of solar panels. Besides generating a maximum amount of power, an MPPT controller aims for stability in order to avoid additional losses. Reinforcement Learning (RL) is a flexible training method that can produce a controller for a complex problem without a detailed prior knowledge of the environment. In this work we propose an approach that combines an optimized neural network, trained through deep reinforcement learning (DRL), with a classical closed-loop control. Experimental results suggest that the proposed approach outperforms a recently proposed DRL network both in terms of efficiency and stability. A compact and discrete version of the proposed controller is also evaluated and shown to further increase performance. The implemented algorithms and the RL simulation environment are made available in an open-source repository.

In this paper we propose a procedure to enable the training of several independent Multilayer Perceptron Neural Networks with a different number of neurons and activation functions in parallel (ParallelMLPs) by exploring the principle of locality and parallelization capabilities of modern CPUs and GPUs. The core idea of this technique is to represent several sub-networks as a single large network and use a Modified Matrix Multiplication that replaces an ordinal matrix multiplication with two simple matrix operations that allow separate and independent paths for gradient flowing. We have assessed our algorithm in simulated datasets varying the number of samples, features and batches using 10,000 different models as well as in the MNIST dataset. We achieved a training speedup from 1 to 4 orders of magnitude if compared to the sequential approach. The code is available online.

We present an innovative step towards a parameterless out-of-the-box population size control for evolutionary and swarm-based algorithms for single objective bound constrained real-parameter numerical optimization. To the best of our knowledge, our approach is the first parameterless out-of-the-box parameter control for such a kind of technique. It is easy to implement and to use, since it does not require the adjustment of any parameter. The general idea is to increment the velocity of the population change if the best fitness stagnates, and decrement it otherwise. Then, in order to effectively change the population size, a mechanism of removal/addition of individuals inspired by the selection methods of evolutionary algorithms is executed. Our experimental results provide evidence that our controller is not only compatible with any evolutionary or swarm-based algorithm for single objective bound constrained real-parameter numerical optimization, but that it also performs well in many scenarios. • This paper proposes a parameterless out-of-the-box population size control method for monoobjective evolutionary and swarm-based algorithms. • The absence of parameters makes such a method straightforward to be used. Besides, its simple rationale makes it easy to implement. • The proposed method is lightweight, which means that it usually does not add any significant overhead to the optimization process. • The proposed controller is suitable to be implemented with any evolutionary and swarm-based algorithm.

With the advance of Information Technologies, recipe sharing websites have become very common, and people can use such systems in an attempt to find a recipe that fits both their desires and nutritional needs. But sometimes ingredients from a given recipe list may not be available, what may require some adaptations by replacing the missing ingredients. In this work, a data-driven approach is employed to develop a new recipe generation system, that recommends substitute ingredients to adapt recipes into a target domain. The proposed system, which is based on Text Mining and Data Clustering techniques, is evaluated by means of a qualitative analysis, showing promising results.

Neste trabalho, três models de Agrupamento Automático de Dados, baseados na meta-heurística de Otimização por Busca em Grupo (GSO), são introduzidos, chamados RHGSO, ADHGSO e BDHGSO. Nos modelos propostos, a busca global do GSO é melhorada através de heurísticas de busca local adaptadas ao contexto de Agrupamento Automático de Dados, onde operações de ativação, desativação e substituição de centroides de agrupamentos são executadas, objetivando a realização de perturbações que visam o aumento da velocidade de exploração do grupo do GSO. Os algoritmos propostos são comparados a outros Algoritmos Evolucionários e de Inteligência de Enxames da literatura, apresentando resultados promissores.

Building robust deterministic neural networks remains a challenge. On the one hand, some approaches improve out-of-distribution detection at the cost of reducing classification accuracy in some situations. On the other hand, some methods simultaneously increase classification accuracy, uncertainty estimation, and out-of-distribution detection at the expense of reducing the inference efficiency. In this paper, we propose training deterministic neural networks using our DisMax loss, which works as a drop-in replacement for the usual SoftMax loss (i.e., the combination of the linear output layer, the SoftMax activation, and the cross-entropy loss). Starting from the IsoMax+ loss, we create each logit based on the distances to all prototypes, rather than just the one associated with the correct class. We also introduce a mechanism to combine images to construct what we call fractional probability regularization. Moreover, we present a fast way to calibrate the network after training. Finally, we propose a composite score to perform out-of-distribution detection. Our experiments show that DisMax usually outperforms current approaches simultaneously in classification accuracy, uncertainty estimation, and out-of-distribution detection while maintaining deterministic neural network inference efficiency. The code to reproduce the results is available at https://github.com/dlmacedo/distinction-maximization-loss.

Nowadays, any individual interested in learning how to prepare a new dish or complete meal can consult specialized cooking recipe websites and video sharing platforms. Such online repositories are able to keep hundreds of thousands of entries in their databases, being rich sources of information for users with the most vast degrees of expertise. But sometimes it is hard for a user to find an adequate recipe that fits, simultaneously, his nutritional needs, tastes, dietary restrictions and the set of ingredients at hand, and, for non-expert users, it may be too much complicated, or even impossible, to adapt the recipes returned by such systems into his current state of needs. In this work, we propose a new cooking recipe recommendation and generation system, based on improved alternative Average Support Value (ASV) filters and a data-driven text mining approach, for single ingredient substitution. Three new ASV variants are proposed as mechanisms to aggregate recipe context and ingredient relevance into standard ASV, in an attempt to promote better ingredient substitute recommendations to adapt recipes into new culinary domains. The proposed ASV-based recipe generation systems are tested and evaluated by means of qualitative analysis, and the proposed filters performances are compared with standard ASV, when adapting recipes with no restrictions into different dietary restriction domains, showing promising results.

This works proposes a methodology to searching for automatically Artificial Neural Networks (ANN) by using Cellular Genetic Algorithm (CGA). The goal of this methodology is to find compact networks whit good performance for classification problems. The main reason for developing this work is centered at the difficulties of configuring compact ANNs with good performance rating. The use of CGAs aims at seeking the components of the RNA in the same way that a common Genetic Algorithm (GA), but it has the differential of incorporating a Cellular Automaton (CA) to give location for the GA individuals. The location imposed by the CA aims to control the spread of solutions in the populations to maintain the genetic diversity for longer time. This genetic diversity is important for obtain good results with the GAs.

The optimization of Artificial Neural Networks (ANNs) is an important task to the success of using these models in real-world applications. The solutions adopted to this task are expensive in general, involving trial-and-error procedures or expert knowledge which are not always available. In this work, we investigated the use of meta-learning to the optimization of ANNs. Meta-learning is a research field aiming to automatically acquiring knowledge which relates features of the learning problems to the performance of the learning algorithms. The meta-learning techniques were originally proposed and evaluated to the algorithm selection problem and after to the optimization of parameters for Support Vector Machines. However, meta-learning can be adopted as a more general strategy to optimize ANN parameters, which motivates new efforts in this research direction. In the current work, we performed a case study using meta-learning to choose the number of hidden nodes for MLP networks, which is an important parameter to be defined aiming a good networks performance. In our work, we generated a base of meta-examples associated to 93 regression problems. Each meta-example was generated from a regression problem and stored: 16 features describing the problem (e.g., number of attributes and correlation among the problem attributes) and the best number of nodes for this problem, empirically chosen from a range of possible values. This set of meta-examples was given as input to a meta-learner which was able to predict the best number of nodes for new problems based on their features. The experiments performed in this case study revealed satisfactory results.

RESUMO A área de Inteligência Artificial demonstrou avanços extraordinários nos últimos anos e, atualmente, é utilizada para solucionar inúmeros problemas tecnológicos e econômicos. Como boa parte do sucesso atual da Inteligência Artificial se deve às técnicas de Aprendizado de Máquina, particularmente às Redes Neurais Artificiais, neste artigo falamos dessas áreas, estado atual, desafios e oportunidades de pesquisas. Vamos também mencionar preocupações com impactos sociais e questões éticas.

Recipe sharing websites have become even more popular in the past few decades, and such repositories are able to keep hundreds of thousands of cooking recipes at the same time. Many recipe websites are developed with the participation of their community of users, which are allowed to upload new recipes and to provide evaluations and comments on the available recipes. However, in such repositories, the amount of recipes that are safe for users with special needs, such as food restrictions or allergies, is much smaller than ordinary food recipes, what may restrict the access and usability provided by such websites to that public. In this work, we propose a new recipe recommendation and generation system, based on a data-driven approach for single ingredient substitution, in such a way that recipes containing forbidden ingredients, according to a category of user food restrictions, are adapted by replacing such ingredients by safe ingredients. The proposed ingredient substitute recommendation system is based on a filtering process that takes into consideration the original recipe context, the relationship among sets of ingredients and the user preferences, towards the generation of recipes that are safe, and at the same time contemplate both user needs and tastes. The proposed system is evaluated by means of a qualitative analysis, showing promising results.

Spiking neural networks (SNNs) aim to replicate energy efficiency, learning speed and temporal processing of biological brains. However, accuracy and learning speed of such networks is still behind reinforcement learning (RL) models based on traditional neural models. This work combines a pre-trained binary convolutional neural network with an SNN trained online through reward-modulated STDP in order to leverage advantages of both models. The spiking network is an extension of its previous version, with improvements in architecture and dynamics to address a more challenging task. We focus on extensive experimental evaluation of the proposed model with optimized state-of-the-art baselines, namely proximal policy optimization (PPO) and deep Q network (DQN). The models are compared on a grid-world environment with high dimensional observations, consisting of RGB images with up to 256 × 256 pixels. The experimental results show that the proposed architecture can be a competitive alternative to deep reinforcement learning (DRL) in the evaluated environment and provide a foundation for more complex future applications of spiking networks.