224 research outputs found
A Modified SVM Classification Algorithm for Data of Variable Quality
No full textWe propose a modified SVM algorithm for the classification of data augmented with explicit quality quantification for each example in the training set. As the extension to nonlinear decision functions through the use of kernels brings to a non-convex optimization problem, we develop an approximate solution. Finally, the proposed approach is applied to a set of benchmarks and contrasted with analogous methodologies in the literature
Collaboration at the basis of sharing focused information: the opportunistic networks
No full textThere is no doubt that the sharing of information lies at the basis of any collaborative framework. While this is the keen contrivance of social computation paradigms such as ant colonies and neural networks, it also represented the Achilles’ heel of many parallel computation protocols of the eighties. In addition to computational overhead due to the transfer of the information in these protocols, a modern drawback is constituted by intrusions in the communication channels, e.g. spamming in the e-mails, injection of malicious programming codes, or in general attacks on the data communication.While swarm intelligence and connectionist paradigms overcome these drawbacks with a fault tolerant broadcasting of data - any agent has access massively to any message reaching him - in this chapter we discuss within the paradigm of opportunistic networks an automatically selective communication protocol particularly suited to set up a robust collaboration within a very local community of agents. Like medieval monks who escaped world chaos and violence by taking refuge in small and protected communities, modern people may escape the information avalanche by forming virtual communities that do not in any case relinquish most ITC (Information Technology Community) benefits. A communication middleware to obtain this result is represented by opportunistic networks
SVM with Random Labels
No full textWe devise an SVM for partitioning a sample space affected by random binary labels. In the hypothesis that a smooth, possibly symmetric, conditional label distribution graduates the passage from the all 0-label domain to the all 1-label domain and under other regularity conditions, the algorithm supplies an estimate of the above probabilities. Within the Algorithmic Inference framework, the randomness of the labels maintains the main features of the binary classification problem, yet adding a further dimension to the search space. Namely the new dimension of each point in the original space hosts the uniform seeds accounting for the randomness of the labels, so that the problem becomes that of separating the points in the augmented space. We solve it with a new kind of bootstrap technique. As for error bounds of the proposed algorithm, we obtain confidence intervals that are up to an order narrower than those supplied in the literature. This benefit comes from the fact that: (i) we devise a special algorithm to take into account the random profile of the labels; (ii) we know the number of support vectors really employed, as an ancillary output of the learning procedure; and (iii) we can appreciate confidence intervals of misclassifying probability exactly in function of the cardinality of these vectors. We numerically check these results by measuring the coverage of the confidence intervals
Toward a space-time mobility model for social communities
No full textWe present a sensitivity study of a wait and chase scheme introduced in a previous work to model the contact times between people belonging to a social community. The membership presupposes that, besides purely occasional encounters, people are motivated to meet other members of the community, while the social character of the latter makes each person met an equivalent target. This calls
for a mobility in the family of Lévy jumps alternating a wandering period within a limited environment – waiting phase – with jumping to a new site constituting the
target of a chase phase. In this paper we aim to connect specific features of single individual dynamics to the overall evolution of the social community in the true
thread of the Palm calculus philosophy. We base this study on a large mobility track dataset expressly collected with this objective
Discrete Production Control through SCADA Technology: An application case in automotive industry
No full textThe development of the hardware technologies has made cheaper microprocessors and, in general, electronic components. At the same time, a continuous improvement of the performances of these components, has made possible a rapid development of their applications to production’s problems and an increase of their functionalities.As disadvantage the complexity of these components is increased because the software dedicated to their management, supervision and control has to be expanded so a lot of problems are rise up both for the maintenance that for the expandability of the systems.This is the main reason because actually automation technologies are oriented versus distributed systems characterised by autonomous systems able to communicate each other (through, i.e., using SCADA – Supervisory Control and Data Acquisition –technologies) resolving the problem of the complexity.In such context we have implemented a SCADA (based on RSView 32) able to collect – real time – the information coming from each part of a given production line and, subsequently, to process them in order to follow the system’s evolution; at the same time we have defined a lot of synoptical tables to make easier the evaluation of the main production parameters. Using these tables we are able, in the end, to single out an optimal balancing solution.The developed system is currently employed, successfully, in a well known factory of engines allowing us to identify a solution able to rationalise the monitoring phase of the whole process and, at the same time, to maximise the efficiency of the management of the technical and the human resources.Especially our system is able, given an overbooking situation of the production cycle analysed, to identify on time the machine that have caused such situation and, consequently, to adapt the productive rhythms of the machines upstream and downstream of it
Relevance regression learning with support vector machines
No full textWe propose a variant of two SVM regression algorithms expressly tailored in order to exploit additional information summarizing the relevance of each data item, as a measure of its relative importance w.r.t. the remaining examples. These variants, enclosing the original formulations when all data items have the same relevance, are preliminary tested on synthetic and real-world data sets. The obtained results outperform standard SVM approaches to regression if evaluated in light of the above mentioned additional information about data quality
Training a network of mobile neurons
No full textWe introduce a new paradigm of neural networks
where neurons autonomously search for the best reciprocal position in a topological space so as to exchange information more profitably. The idea that elementary processors move within a network to get a proper position is borne out by biological neurons in brain morphogenesis. The basic rule we state for this dynamics is that a neuron is attracted by the mates which are most informative and repelled by ones which aremost similar to it. By embedding this rule into a Newtonian dynamics, we obtain a network which autonomously organizes its layout. Thanks to this further adaptation, the network proves to be robustly trainable through an extended version of the back- propagation algorithm even in the case of deep architectures. We test this network on two classic benchmarks and thereby get many insights on how the network behaves, and when and why it succeeds
A comparison between the use of ESNN on Long Stereo-EEG Recordings and their largest Lyapunov exponent profiles for epileptic brain analysis
No full textIn the last 25 years many works in literature about the capability to detect or predict the occurrence of epileptic seizures, starting from the electroencephalogram (EEG) signal analysis, have often hypothesized that the epileptogenic activity is the result of an abnormal electrical activity hyper-synchronization of different points in an epileptic brain. We already proposed our method to integrate Neural Networks (NN) and the largest Lyapunov exponent (Lmax) for capturing brain dynamics through long stereo-EEG (sEEG) data recorded. In this paper we want to compare the use of a classical Evolving Spiking NN (ESNN) on long sEEG recordings with the integrated method previously proposed. Results are interesting and encourage us to develop, in the next future, a framework for EEG signal analysis
Wireless domotic: an enabling platform for granular intelligence
No full textWe describe a wireless platform for implementing domotic applications drawn from a fuzzy rule system. The design of its architecture hits three targets: i) complete scalability with the addition of new devices into the domotic system, ii) full operability by the user who is the arbiter of any operational option, and iii) robotic controllability based on a series of environmental sensors and an autonomous rule system under the total supervision of the user. The hardware counterpart is based on a series of micro-transmitters connecting the house devices to the platform through Z-Wave or wired protocols and smartphones as end-user terminals. In the paper we describe how the entire platform has been implemented in our lab and provide some considerations deriving from its early operations
Histamine beyond its effects on allergy: Potential therapeutic benefits for the treatment of Amyotrophic Lateral Sclerosis (ALS)
No full textALS currently remains a challenge despite many efforts in performing successful clinical trials and formulating therapeutic solutions. By learning from current failures and striving for success, scientists and clinicians are checking every possibility to search for missing hints and efficacious treatments. Because the disease is very complex and heterogeneous and, moreover, targeting not only motor neurons but also several different cell types including muscle, glial, and immune cells, the right answer to ALS is conceivably a multidrug strategy or the use of broad-spectrum molecules. The aim of the present work is to gather evidence about novel perspectives on ALS pathogenesis and to present recent and innovative paradigms for therapy. In particular, we describe how an old molecule possessing immunomodulatory and neuroprotective functions beyond its recognized effects on allergy, histamine, might have a renewed and far-reaching momentum in ALS. (C) 2019 The Author(s). Published by Elsevier Inc
- …
