Luís A. Alexandre

Vasco Lopes, Bruno Degardin, Luís A. Alexandre

Neural Architecture Search (NAS) benchmarks significantly improved the capability of developing and comparing NAS methods while at the same time drastically reduced the computational overhead by providing meta-information about thousands of trained neural networks. However, tabular benchmarks have several drawbacks that can hinder fair comparisons and provide unreliable results. These usually focus on providing a small pool of operations in heavily constrained search spaces -- usually cell-based neural networks with pre-defined outer-skeletons. In this work, we conducted an empirical analysis of the widely used NAS-Bench-101, NAS-Bench-201 and TransNAS-Bench-101 benchmarks in terms of their generability and how different operations influence the performance of the generated architectures. We found that only a subset of the operation pool is required to generate architectures close to the upper-bound of the performance range. Also, the performance distribution is negatively skewed, having a higher density of architectures in the upper-bound range. We consistently found convolution layers to have the highest impact on the architecture's performance, and that specific combination of operations favors top-scoring architectures. These findings shed insights on the correct evaluation and comparison of NAS methods using NAS benchmarks, showing that directly searching on NAS-Bench-201, ImageNet16-120 and TransNAS-Bench-101 produces more reliable results than searching only on CIFAR-10. Furthermore, with this work we provide suggestions for future benchmark evaluations and design. The code used to conduct the evaluations is available at https://github.com/VascoLopes/NAS-Benchmark-Evaluation.

Vasco Lopes, Luís A. Alexandre

Networks found with Neural Architecture Search (NAS) achieve state-of-the-art performance in a variety of tasks, out-performing human-designed networks. However, most NAS methods heavily rely on human-defined assumptions that constrain the search: architecture's outer-skeletons, number of layers, parameter heuristics and search spaces. Additionally, common search spaces consist of repeatable modules (cells) instead of fully exploring the architecture's search space by designing entire architectures (macro-search). Imposing such constraints requires deep human expertise and restricts the search to pre-defined settings. In this paper, we propose LCMNAS, a method that pushes NAS to less constrained search spaces by performing macro-search without relying on pre-defined heuristics or bounded search spaces. LCMNAS introduces three components for the NAS pipeline: i) a method that leverages information about well-known architectures to autonomously generate complex search spaces based on Weighted Directed Graphs with hidden properties, ii) an evolutionary search strategy that generates complete architectures from scratch, and iii) a mixed-performance estimation approach that combines information about architectures at initialization stage and lower fidelity estimates to infer their trainability and capacity to model complex functions. We present experiments in 13 different data sets showing that LCMNAS is capable of generating both cell and macro-based architectures with minimal GPU computation and state-of-the-art results. More, we conduct extensive studies on the importance of different NAS components in both cell and macro-based settings. Code for reproducibility is public at https://github.com/VascoLopes/LCMNAS.

Seyed Jalaleddin Mousavirad, Luís A. Alexandre

Customer satisfaction is crucially affected by energy consumption in mobile devices. One of the most energy-consuming parts of an application is images. While different images with different quality consume different amounts of energy, there are no straightforward methods to calculate the energy consumption of an operation in a typical image. This paper, first, investigates that there is a correlation between energy consumption and image quality as well as image file size. Therefore, these two can be considered as a proxy for energy consumption. Then, we propose a multi-objective strategy to enhance image quality and reduce image file size based on the quantisation tables in JPEG image compression. To this end, we have used two general multi-objective metaheuristic approaches: scalarisation and Pareto-based. Scalarisation methods find a single optimal solution based on combining different objectives, while Pareto-based techniques aim to achieve a set of solutions. In this paper, we embed our strategy into five scalarisation algorithms, including energy-aware multi-objective genetic algorithm (EnMOGA), energy-aware multi-objective particle swarm optimisation (EnMOPSO), energy-aware multi-objective differential evolution (EnMODE), energy-aware multi-objective evolutionary strategy (EnMOES), and energy-aware multi-objective pattern search (EnMOPS). Also, two Pareto-based methods, including a non-dominated sorting genetic algorithm (NSGA-II) and a reference-point-based NSGA-II (NSGA-III) are used for the embedding scheme, and two Pareto-based algorithms, EnNSGAII and EnNSGAIII, are presented. Experimental studies show that the performance of the baseline algorithm is improved by embedding the proposed strategy into metaheuristic algorithms.

Seyed Jalaleddin Mousavirad, Luís A. Alexandre

Energy consumption is a fundamental concern in mobile application development, bearing substantial significance for both developers and end-users. Main objective of this research is to propose a novel neural network-based framework, enhanced by a metaheuristic approach, to achieve robust energy prediction in the context of mobile app development. The metaheuristic approach here aims to achieve two goals: 1) identifying suitable learning algorithms and their corresponding hyperparameters, and 2) determining the optimal number of layers and neurons within each layer. Moreover, due to limitations in accessing certain aspects of a mobile phone, there might be missing data in the data set, and the proposed framework can handle this. In addition, we conducted an optimal algorithm selection strategy, employing 13 base and advanced metaheuristic algorithms, to identify the best algorithm based on accuracy and resistance to missing values. The representation in our proposed metaheuristic algorithm is variable-size, meaning that the length of the candidate solutions changes over time. We compared the algorithms based on the architecture found by each algorithm at different levels of missing values, accuracy, F-measure, and stability analysis. Additionally, we conducted a Wilcoxon signed-rank test for statistical comparison of the results. The extensive experiments show that our proposed approach significantly improves energy consumption prediction. Particularly, the JADE algorithm, a variant of Differential Evolution (DE), DE, and the Covariance Matrix Adaptation Evolution Strategy deliver superior results under various conditions and across different missing value levels.

Seyed Jalaleddin Mousavirad, Luís A. Alexandre

Energy consumption plays a vital role in mobile App development for developers and end-users, and it is considered one of the most crucial factors for purchasing a smartphone. In addition, in terms of sustainability, it is essential to find methods to reduce the energy consumption of mobile devices since the extensive use of billions of smartphones worldwide significantly impacts the environment. Despite the existence of several energy-efficient programming practices in Android, the leading mobile ecosystem, machine learning-based energy prediction algorithms for mobile App development have yet to be reported. Therefore, this paper proposes a histogram-based gradient boosting classification machine (HGBC), boosted by a metaheuristic approach, for energy prediction in mobile App development. Our metaheuristic approach is responsible for two issues. First, it finds redundant and irrelevant features without any noticeable change in performance. Second, it performs a hyper-parameter tuning for the HGBC algorithm. Since our proposed metaheuristic approach is algorithm-independent, we selected 12 algorithms for the search strategy to find the optimal search algorithm. Our finding shows that a success-history-based parameter adaption for differential evolution with linear population size (L-SHADE) offers the best performance. It can improve performance and decrease the number of features effectively. Our extensive set of experiments clearly shows that our proposed approach can provide significant results for energy consumption prediction.

António J. Abreu, Luís A. Alexandre, João A. Santos et al.

Remote sensing is the process of detecting and monitoring the physical characteristics of an area by measuring its reflected and emitted radiation at a distance. It is being broadly used to monitor ecosystems, mainly for their preservation. Ever-growing reports of invasive species have affected the natural balance of ecosystems. Exotic invasive species have a critical impact when introduced into new ecosystems and may lead to the extinction of native species. In this study, we focus on Ludwigia peploides, considered by the European Union as an aquatic invasive species. Its presence can negatively impact the surrounding ecosystem and human activities such as agriculture, fishing, and navigation. Our goal was to develop a method to identify the presence of the species. We used images collected by a drone-mounted multispectral sensor to achieve this, creating our LudVision data set. To identify the targeted species on the collected images, we propose a new method for detecting Ludwigia p. in multispectral images. The method is based on existing state-of-the-art semantic segmentation methods modified to handle multispectral data. The proposed method achieved a producer's accuracy of 79.9% and a user's accuracy of 95.5%.

André Correia, Luís A. Alexandre

Sequence models in reinforcement learning require task knowledge to estimate the task policy. This paper presents a hierarchical algorithm for learning a sequence model from demonstrations. The high-level mechanism guides the low-level controller through the task by selecting sub-goals for the latter to reach. This sequence replaces the returns-to-go of previous methods, improving its performance overall, especially in tasks with longer episodes and scarcer rewards. We validate our method in multiple tasks of OpenAIGym, D4RL and RoboMimic benchmarks. Our method outperforms the baselines in eight out of ten tasks of varied horizons and reward frequencies without prior task knowledge, showing the advantages of the hierarchical model approach for learning from demonstrations using a sequence model.

Vasco Lopes, Miguel Santos, Bruno Degardin et al.

Neural Architecture Search (NAS) methods have been successfully applied to image tasks with excellent results. However, NAS methods are often complex and tend to converge to local minima as soon as generated architectures seem to yield good results. This paper proposes GEA, a novel approach for guided NAS. GEA guides the evolution by exploring the search space by generating and evaluating several architectures in each generation at initialisation stage using a zero-proxy estimator, where only the highest-scoring architecture is trained and kept for the next generation. Subsequently, GEA continuously extracts knowledge about the search space without increased complexity by generating several off-springs from an existing architecture at each generation. More, GEA forces exploitation of the most performant architectures by descendant generation while simultaneously driving exploration through parent mutation and favouring younger architectures to the detriment of older ones. Experimental results demonstrate the effectiveness of the proposed method, and extensive ablation studies evaluate the importance of different parameters. Results show that GEA achieves state-of-the-art results on all data sets of NAS-Bench-101, NAS-Bench-201 and TransNAS-Bench-101 benchmarks.

André Correia, Luís A. Alexandre

With the fast improvement of machine learning, reinforcement learning (RL) has been used to automate human tasks in different areas. However, training such agents is difficult and restricted to expert users. Moreover, it is mostly limited to simulation environments due to the high cost and safety concerns of interactions in the real world. Demonstration Learning is a paradigm in which an agent learns to perform a task by imitating the behavior of an expert shown in demonstrations. It is a relatively recent area in machine learning, but it is gaining significant traction due to having tremendous potential for learning complex behaviors from demonstrations. Learning from demonstration accelerates the learning process by improving sample efficiency, while also reducing the effort of the programmer. Due to learning without interacting with the environment, demonstration learning would allow the automation of a wide range of real world applications such as robotics and healthcare. This paper provides a survey of demonstration learning, where we formally introduce the demonstration problem along with its main challenges and provide a comprehensive overview of the process of learning from demonstrations from the creation of the demonstration data set, to learning methods from demonstrations, and optimization by combining demonstration learning with different machine learning methods. We also review the existing benchmarks and identify their strengths and limitations. Additionally, we discuss the advantages and disadvantages of the paradigm as well as its main applications. Lastly, we discuss our perspective on open problems and research directions for this rapidly growing field.

André Correia, Luís A. Alexandre

Recent advancements in imitation learning have been largely fueled by the integration of sequence models, which provide a structured flow of information to effectively mimic task behaviours. Currently, Decision Transformer (DT) and subsequently, the Hierarchical Decision Transformer (HDT), presented Transformer-based approaches to learn task policies. Recently, the Mamba architecture has shown to outperform Transformers across various task domains. In this work, we introduce two novel methods, Decision Mamba (DM) and Hierarchical Decision Mamba (HDM), aimed at enhancing the performance of the Transformer models. Through extensive experimentation across diverse environments such as OpenAI Gym and D4RL, leveraging varying demonstration data sets, we demonstrate the superiority of Mamba models over their Transformer counterparts in a majority of tasks. Results show that DM outperforms other methods in most settings. The code can be found at https://github.com/meowatthemoon/DecisionMamba.

Nuno Pereira, Luís A. Alexandre

MaskedFusion is a framework to estimate the 6D pose of objects using RGB-D data, with an architecture that leverages multiple sub-tasks in a pipeline to achieve accurate 6D poses. 6D pose estimation is an open challenge due to complex world objects and many possible problems when capturing data from the real world, e.g., occlusions, truncations, and noise in the data. Achieving accurate 6D poses will improve results in other open problems like robot grasping or positioning objects in augmented reality. MaskedFusion improves the state-of-the-art by using object masks to eliminate non-relevant data. With the inclusion of the masks on the neural network that estimates the 6D pose of an object we also have features that represent the object shape. MaskedFusion is a modular pipeline where each sub-task can have different methods that achieve the objective. MaskedFusion achieved 97.3% on average using the ADD metric on the LineMOD dataset and 93.3% using the ADD-S AUC metric on YCB-Video Dataset, which is an improvement, compared to the state-of-the-art methods. The code is available on GitHub (https://github.com/kroglice/MaskedFusion).

Ana C. Perre, Luís A. Alexandre, Luís C. Freire

Contrast-enhanced spectral mammography (CESM) is an imaging modality that provides two types of images, commonly known as low-energy (LE) and dual-energy subtracted (DES) images. In many domains, particularly in medicine, the emergence of image-to-image translation techniques has enabled the artificial generation of images using other images as input. Within CESM, applying such techniques to generate DES images from LE images could be highly beneficial, potentially reducing patient exposure to radiation associated with high-energy image acquisition. In this study, we investigated three models for the artificial generation of DES images (virtual DES): a pre-trained U-Net model, a U-Net trained end-to-end model, and a CycleGAN model. We also performed a series of experiments to assess the impact of using virtual DES images on the classification of CESM examinations into malignant and non-malignant categories. To our knowledge, this is the first study to evaluate the impact of virtual DES images on CESM lesion classification. The results demonstrate that the best performance was achieved with the pre-trained U-Net model, yielding an F1 score of 85.59% when using the virtual DES images, compared to 90.35% with the real DES images. This discrepancy likely results from the additional diagnostic information in real DES images, which contributes to a higher classification accuracy. Nevertheless, the potential for virtual DES image generation is considerable and future advancements may narrow this performance gap to a level where exclusive reliance on virtual DES images becomes clinically viable.

André Correia, Luís A. Alexandre

This paper presents a framework for learning visual representations from unlabeled video demonstrations captured from multiple viewpoints. We show that these representations are applicable for imitating several robotic tasks, including pick and place. We optimize a recently proposed self-supervised learning algorithm by applying contrastive learning to enhance task-relevant information while suppressing irrelevant information in the feature embeddings. We validate the proposed method on the publicly available Multi-View Pouring and a custom Pick and Place data sets and compare it with the TCN triplet baseline. We evaluate the learned representations using three metrics: viewpoint alignment, stage classification and reinforcement learning, and in all cases the results improve when compared to state-of-the-art approaches, with the added benefit of reduced number of training iterations.

Nuno Pereira, Luís A. Alexandre

Object pose estimation has multiple important applications, such as robotic grasping and augmented reality. We present a new method to estimate the 6D pose of objects that improves upon the accuracy of current proposals and can still be used in real-time. Our method uses RGB-D data as input to segment objects and estimate their pose. It uses a neural network with multiple heads to identify the objects in the scene, generate the appropriate masks and estimate the values of the translation vectors and the quaternion that represents the objects' rotation. These heads leverage a pyramid architecture used during feature extraction and feature fusion. We conduct an empirical evaluation using the two most common datasets in the area, and compare against state-of-the-art approaches, illustrating the capabilities of MPF6D. Our method can be used in real-time with its low inference time and high accuracy.

Vasco Lopes, Miguel Santos, Bruno Degardin et al.

Neural Architecture Search (NAS) methods have been successfully applied to image tasks with excellent results. However, NAS methods are often complex and tend to converge to local minima as soon as generated architectures seem to yield good results. In this paper, we propose G-EA, a novel approach for guided evolutionary NAS. The rationale behind G-EA, is to explore the search space by generating and evaluating several architectures in each generation at initialization stage using a zero-proxy estimator, where only the highest-scoring network is trained and kept for the next generation. This evaluation at initialization stage allows continuous extraction of knowledge from the search space without increasing computation, thus allowing the search to be efficiently guided. Moreover, G-EA forces exploitation of the most performant networks by descendant generation while at the same time forcing exploration by parent mutation and by favouring younger architectures to the detriment of older ones. Experimental results demonstrate the effectiveness of the proposed method, showing that G-EA achieves state-of-the-art results in NAS-Bench-201 search space in CIFAR-10, CIFAR-100 and ImageNet16-120, with mean accuracies of 93.98%, 72.12% and 45.94% respectively.

Saeid Alirezazadeh, André Correia, Luís A. Alexandre

A robotic network is a system with multiple robots connected by a communication network. Certain tasks that cannot be accomplished with available robotic resources are candidates for the use of cloud robotics, which overcomes the limitations of the robot network by adding to the network, either local or remote servers or cloud infrastructure, to aid in computational demanding tasks or storage. Previous studies have mainly focused on minimizing the cost of the robots in retrieving resources by knowing the resource allocation in advance. We develop a method for a robotic network cloud system that includes robots, fog and cloud nodes, to determine where each algorithm should be allocated so that the system achieves optimal performance, regardless of which robot initiates the request. We can find the minimum required memory for the robots and the optimal way to allocate the algorithms with the shortest time to complete each task. We experimentally compare our method with a state-of-the-art method, using real-world data, showing the improvements that can be obtained.

Vasco Lopes, Saeid Alirezazadeh, Luís A. Alexandre

Neural Architecture Search (NAS) has shown excellent results in designing architectures for computer vision problems. NAS alleviates the need for human-defined settings by automating architecture design and engineering. However, NAS methods tend to be slow, as they require large amounts of GPU computation. This bottleneck is mainly due to the performance estimation strategy, which requires the evaluation of the generated architectures, mainly by training them, to update the sampler method. In this paper, we propose EPE-NAS, an efficient performance estimation strategy, that mitigates the problem of evaluating networks, by scoring untrained networks and creating a correlation with their trained performance. We perform this process by looking at intra and inter-class correlations of an untrained network. We show that EPE-NAS can produce a robust correlation and that by incorporating it into a simple random sampling strategy, we are able to search for competitive networks, without requiring any training, in a matter of seconds using a single GPU. Moreover, EPE-NAS is agnostic to the search method, since it focuses on the evaluation of untrained networks, making it easy to integrate into almost any NAS method.

Vasco Lopes, António Gaspar, Luís A. Alexandre et al.

Sentiment analysis is a research topic focused on analysing data to extract information related to the sentiment that it causes. Applications of sentiment analysis are wide, ranging from recommendation systems, and marketing to customer satisfaction. Recent approaches evaluate textual content using Machine Learning techniques that are trained over large corpora. However, as social media grown, other data types emerged in large quantities, such as images. Sentiment analysis in images has shown to be a valuable complement to textual data since it enables the inference of the underlying message polarity by creating context and connections. Multimodal sentiment analysis approaches intend to leverage information of both textual and image content to perform an evaluation. Despite recent advances, current solutions still flounder in combining both image and textual information to classify social media data, mainly due to subjectivity, inter-class homogeneity and fusion data differences. In this paper, we propose a method that combines both textual and image individual sentiment analysis into a final fused classification based on AutoML, that performs a random search to find the best model. Our method achieved state-of-the-art performance in the B-T4SA dataset, with 95.19% accuracy.

Saeid Alirezazadeh, Luís A. Alexandre

A scheduling method in a robotic network cloud system with minimal makespan is beneficial as the system can complete all the tasks assigned to it in the fastest way. Robotic network cloud systems can be translated into graphs where nodes represent hardware with independent computing power and edges represent data transmissions between nodes. Time window constraints on tasks are a natural way to order tasks. The makespan is the maximum amount of time between when the first node to receive a task starts executing its first scheduled task and when all nodes have completed their last scheduled task. Load balancing allocation and scheduling ensures that the time between when the first node completes its scheduled tasks and when all other nodes complete their scheduled tasks is as short as possible. We propose a grid of all tasks to ensure that the time window constraints for tasks are met. We propose grid of all tasks balancing algorithm for distributing and scheduling tasks with minimum makespan. We theoretically prove the correctness of the proposed algorithm and present simulations illustrating the obtained results.

Vasco Lopes, Luís A. Alexandre

The dominant approach for surface defect detection is the use of hand-crafted feature-based methods. However, this falls short when conditions vary that affect extracted images. So, in this paper, we sought to determine how well several state-of-the-art Convolutional Neural Networks perform in the task of surface defect detection. Moreover, we propose two methods: CNN-Fusion, that fuses the prediction of all the networks into a final one, and Auto-Classifier, which is a novel proposal that improves a Convolutional Neural Network by modifying its classification component using AutoML. We carried out experiments to evaluate the proposed methods in the task of surface defect detection using different datasets from DAGM2007. We show that the use of Convolutional Neural Networks achieves better results than traditional methods, and also, that Auto-Classifier out-performs all other methods, by achieving 100% accuracy and 100% AUC results throughout all the datasets.

Vasco Lopes, Luís A. Alexandre

Neural Architecture Search has achieved state-of-the-art performance in a variety of tasks, out-performing human-designed networks. However, many assumptions, that require human definition, related with the problems being solved or the models generated are still needed: final model architectures, number of layers to be sampled, forced operations, small search spaces, which ultimately contributes to having models with higher performances at the cost of inducing bias into the system. In this paper, we propose HMCNAS, which is composed of two novel components: i) a method that leverages information about human-designed models to autonomously generate a complex search space, and ii) an Evolutionary Algorithm with Bayesian Optimization that is capable of generating competitive CNNs from scratch, without relying on human-defined parameters or small search spaces. The experimental results show that the proposed approach results in competitive architectures obtained in a very short time. HMCNAS provides a step towards generalizing NAS, by providing a way to create competitive models, without requiring any human knowledge about the specific task.

Saeid Alirezazadeh, Luís A. Alexandre

Every robotic network cloud system can be seen as a graph with nodes as hardware with independent computational processing powers and edges as data transmissions between nodes. When assigning a task to a node we may change several values corresponding to the node such as distance to other nodes, the time to complete all of its tasks, the energy level of the node, energy consumed while performing all of its tasks, geometrical position, communication with other nodes, and so on. These values can be seen as fingerprints for the current state of the node which can be evaluated as a subspace of a hyperspace. We proposed a theoretical model describing how assigning tasks to a node will change the subspace of the hyperspace, and from that, we show how to obtain the optimal task allocation. We described the communication instability between nodes and the capability of nodes as subspaces of a hyperspace. We translate task scheduling to nodes as finding the maximum volume of the hyperspace.

Saeid Alirezazadeh, Luís A. Alexandre

In order for a robot to perform a task, several algorithms need to be executed, sometimes, simultaneously. Algorithms can be run either on the robot itself or, upon request, be performed on cloud infrastructure. The term cloud infrastructure is used to describe hardware, storage, abstracted resources, and network resources related to cloud computing. Depending on the decisions on where to execute the algorithms, the overall execution time and necessary memory space for the robot will change accordingly. The price of a robot depends, among other things, on its memory capacity and computational power. We answer the question of how to keep a given performance and use a cheaper robot (lower resources) by assigning computational tasks to the cloud infrastructure, depending on memory, computational power, and communication constraints. Also, for a fixed robot, our model provides a way to have optimal overall performance. We provide a general model for the optimal decision of algorithm allocation under certain constraints. We exemplify the model with simulation results. The main advantage of our model is that it provides an optimal task allocation simultaneously for memory and time.

Miguel Fernandes, Luís A. Alexandre

A blockchain, during its lifetime, records large amounts of data, that in a common usage its kept on its entirety. In a robotics environment, the old information is useful for human evaluation, or oracles interfacing with the blockchain but it is not useful for the robots that require only current information in order to continue their work. This causes a storage problem in blockchain nodes that have limited storage capacity, such as in the case of nodes attached to robots that are usually built around embedded solutions. This paper presents a time-segmentation solution for devices with limited storage capacity, integrated in a particular robot-directed blockchain called RobotChain. Results are presented regarding the proposed solution that show that the goal of restricting each node's capacity is reached without compromising all the benefits that arise from the use of blockchains in these contexts, and on the contrary, it allows for cheap nodes to use this blockchain, reduces storage costs and allows faster deployment of new nodes.

Vasco Lopes, Luís A. Alexandre, Nuno Pereira

Blockchain is a disruptive technology that is normally used within financial applications, however it can be very beneficial also in certain robotic contexts, such as when an immutable register of events is required. Among the several properties of Blockchain that can be useful within robotic environments, we find not just immutability but also decentralization of the data, irreversibility, accessibility and non-repudiation. In this paper, we propose an architecture that uses blockchain as a ledger and smart-contract technology for robotic control by using external parties, Oracles, to process data. We show how to register events in a secure way, how it is possible to use smart-contracts to control robots and how to interface with external Artificial Intelligence algorithms for image analysis. The proposed architecture is modular and can be used in multiple contexts such as in manufacturing, network control, robot control, and others, since it is easy to integrate, adapt, maintain and extend to new domains.

Vasco Lopes, Luís A. Alexandre

Blockchain technology is growing everyday at a fast-passed rhythm and it's possible to integrate it with many systems, namely Robotics with AI services. However, this is still a recent field and there isn't yet a clear understanding of what it could potentially become. In this paper, we conduct an overview of many different methods and platforms that try to leverage the power of blockchain into robotic systems, to improve AI services or to solve problems that are present in the major blockchains, which can lead to the ability of creating robotic systems with increased capabilities and security. We present an overview, discuss the methods and conclude the paper with our view on the future of the integration of these technologies.

Abel S. Zacarias, Luís A. Alexandre

Lifelong learning aims to develop machine learning systems that can learn new tasks while preserving the performance on previous learned tasks. In this paper we present a method to overcome catastrophic forgetting on convolutional neural networks, that learns new tasks and preserves the performance on old tasks without accessing the data of the original model, by selective network augmentation. The experiment results showed that SeNA-CNN, in some scenarios, outperforms the state-of-art Learning without Forgetting algorithm. Results also showed that in some situations it is better to use SeNA-CNN instead of training a neural network using isolated learning.

Ricardo Gamelas Sousa, Jorge M. Santos, Luís M. Silva et al.

In this paper we present a system for the detection of immunogold particles and a Transfer Learning (TL) framework for the recognition of these immunogold particles. Immunogold particles are part of a high-magnification method for the selective localization of biological molecules at the subcellular level only visible through Electron Microscopy. The number of immunogold particles in the cell walls allows the assessment of the differences in their compositions providing a tool to analise the quality of different plants. For its quantization one requires a laborious manual labeling (or annotation) of images containing hundreds of particles. The system that is proposed in this paper can leverage significantly the burden of this manual task. For particle detection we use a LoG filter coupled with a SDA. In order to improve the recognition, we also study the applicability of TL settings for immunogold recognition. TL reuses the learning model of a source problem on other datasets (target problems) containing particles of different sizes. The proposed system was developed to solve a particular problem on maize cells, namely to determine the composition of cell wall ingrowths in endosperm transfer cells. This novel dataset as well as the code for reproducing our experiments is made publicly available. We determined that the LoG detector alone attained more than 84\% of accuracy with the F-measure. Developing immunogold recognition with TL also provided superior performance when compared with the baseline models augmenting the accuracy rates by 10\%.

Ricardo Gamelas Sousa, Luís A. Alexandre, Jorge M. Santos et al.

Transfer Learning (TL) aims to transfer knowledge acquired in one problem, the source problem, onto another problem, the target problem, dispensing with the bottom-up construction of the target model. Due to its relevance, TL has gained significant interest in the Machine Learning community since it paves the way to devise intelligent learning models that can easily be tailored to many different applications. As it is natural in a fast evolving area, a wide variety of TL methods, settings and nomenclature have been proposed so far. However, a wide range of works have been reporting different names for the same concepts. This concept and terminology mixture contribute however to obscure the TL field, hindering its proper consideration. In this paper we present a review of the literature on the majority of classification TL methods, and also a distribution-based categorization of TL with a common nomenclature suitable to classification problems. Under this perspective three main TL categories are presented, discussed and illustrated with examples.