Papers by Leonardo Reyneri
This article presents PAPRICA-3, a high-speed 32-neurons slice for real-time neural processing of... more This article presents PAPRICA-3, a high-speed 32-neurons slice for real-time neural processing of images. The system is a programmable par-allel processor array with the instruction set tailored to the emulation of neural networks and to image processing. Dedicated hardware features allow simultaneous image acquisition, image processing and neural network emula-tion. The 32-neurons slice is cascadable to match the required size. The system is under fabrication and has been simulated at a clock frequency of 100 MHz typical. Estimated performance is up to 130 MCPS for percep-trons. The system has been tested on the recog-nition of handwritten check amounts, by inte-grating neural network algorithms with context analysis techniques.
Lecture Notes in Computer Science, 2001
In the present article are described the results obtained by the application of neuro-fuzzy metho... more In the present article are described the results obtained by the application of neuro-fuzzy methodologies in the study of Bactrocera Oleae (olive fly) infestation in Liguria region olive grows. The main aim of this project is create an informatic decisional support for experts in the applications of Integrated Pest Management strategies against the Bactrocera Oleae infestation. This system will suggest
Lecture Notes in Computer Science, 2001
ABSTRACT In this paper we consider the successful hybridation of a two modern computational schem... more ABSTRACT In this paper we consider the successful hybridation of a two modern computational schemes, Clustering and Neural Networks, for the Predictive Classification of the future value of insect infestation levels for Integrated Pest Management (IPM) of olive groves. The predictive classification techniques employed allow managers to improve their work in two ways: first, by reducing sampling demands of the variables involved, which is a costly process; and second, by recognizing potential infestation problems a up to two weeks beforehand, in order to optimize the use of pesticide chemical products and thus reduce financial costs.
Aim of the present article is to show the results obtained from the application of neuro-fuzzy me... more Aim of the present article is to show the results obtained from the application of neuro-fuzzy methodology in the solution of agriculture problems like the Bactrocera Oleae (olive fly) infestation in the Liguria region olive grows. The research is focused to create an informatic decisional instrument to support experts in the applications of Integrated Pest Management strategies against the Bactrocera Oleae infestation. The system will suggest types of treatments for each monitored farm in order to optimize the quality of the olive oil and improve the economic and environmental impact of these treatments. Statistical and forecast analyses on data sets referred to agronomic case studies, like the growth of olive fly, are actually made using standard and model approaches like analytical; these dates instead present characteristics (big variability and non-linearity) which make them complex to be treat mathematically. Agronomic research needs to introduce new analysis techniques of tak...
This article presents PAPRICA-3, a VLSI-oriented architecture for real-time processing of images ... more This article presents PAPRICA-3, a VLSI-oriented architecture for real-time processing of images and its implementation on HACRE, a high-speed, cascadable, 32-processors VLSI slice. The architecture is based on an array of programmable processing elements with the instruction set tailored to image processing, mathematical morphology, and neural networks emulation. Dedicated hardware features allow simultaneous image acquisition, processing, neural network emulation, and a straightforward interface with a hosting PC. HACRE has been fabricated and successfully tested at a clock frequency of 50 MHz. A board hosting up to four chips and providing a 33 MHz PCI interface has been manufactured and used to build BEATRIX, a system for the recognition of handwritten check amounts, by integrating image processing and neural network algorithms (on the board) with context analysis techniques (on the hosting PC).
Intelligent Data Analysis, 2015
ABSTRACT The estimate of the probability density function or probability mass function of an unkn... more ABSTRACT The estimate of the probability density function or probability mass function of an unknown stochastic process is a very important preliminary step for any further elaboration. Most of the traditional approaches to this problem perform a preliminary choice of a parametric mathematical model of the function to estimate and a subsequent fitting on its parameters. To this aim some a-priori knowledge and/or assumptions on the phenomenon under consideration are needed. In this paper an alternative approach is proposed, which does not require any assumption on the available data, as it extracts the probability density function from the output of a neural network, that is trained with a suitable database including the original data and some ad hoc created data with known distribution. The results of the tests performed on synthetic and industrial databases are described and discussed in the paper.
Advances in Neural Networks, 2016
The paper analyses the issues related to the use of neuro-fuzzy techniques in the industrial fiel... more The paper analyses the issues related to the use of neuro-fuzzy techniques in the industrial field focusing on the characteristics that influence the acceptance of the various paradigms. The advantages provided by these techniques and the limits that prevent their wide acceptance in the industrial framework are depicted. Exemplar case study are presented and future perspective and guidelines for the successful integration of soft computing techniques within industry are outlined.
International Journal of Hybrid Intelligent Systems, 2010
ABSTRACT The paper analyses the issues behind allocation and reordering strategies optimization f... more ABSTRACT The paper analyses the issues behind allocation and reordering strategies optimization for an existing automated warehouse for the steelmaking industry. Genetic Algorithms are employed to this purpose by deriving custom chromosome structures as well as ad-hoc crossover and mutation operators. A comparison between three different solutions capable to deal with multi-objective optimization are presented: the first approach is based on a common linear weighting function that combines different objectives; in the second one, a fuzzy system is used to aggregate objective functions, while in the last one the Strength Pareto Evolutionary Algorithm is applied in order to exploit a real multi-objective optimization. These three approaches are described and results are presented in order to highlight benefits and pitfalls of each technique.
An efficient procedure is proposed for initializing two-layer perceptrons and for determining the... more An efficient procedure is proposed for initializing two-layer perceptrons and for determining the optimal number of hidden neurons. This is based on the Orthogonal Least Squares method, which is typical of RBF as well as Wavelet networks. Some experiments are discussed, in which the proposed method is coupled with standard backpropagation training and compared with random initialization. 1 Introduction A suitable and efficient initialization procedure allows to set the initial values of the weights of a network not far from the optimal values determined by training: so doing, the training procedure takes a shorter time to reach the optimal values and therefore considerable saving of computation time is achieved. In networks design, great importance must be attributed also to a correct choice of the number of hidden neurons, which helps avoiding problems of overfitting and contributes to reduce the time required for the training without significantly affecting the network perform...
ISDA 2009 - 9th International Conference on Intelligent Systems Design and Applications, 2009
The paper analyses the issues behind strategies optimization of an existing automated warehouse f... more The paper analyses the issues behind strategies optimization of an existing automated warehouse for the steelmaking industry. Genetic algorithms are employed to this purpose by deriving a custom chromosome structure as well as ad-hoc crossover and mutation operators. A comparison between three different solutions able to deal with multiobjective optimization are presented: the first approach is based on a common
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2009
ABSTRACT This paper presents a method to estimate the reliability of the output of a (possibly ne... more ABSTRACT This paper presents a method to estimate the reliability of the output of a (possibly neuro-fuzzy) model by means of an additional neural network. The proposed technique is most effective when the reliability of the model significantly varies in different areas of input space, as it often happens in many real-world problems, allowing the user to predict how reliable is a given model for each specific situation. Alternatively, the proposed technique can be used to analyze particular anomalies of input data set such as the outliers.
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2011
A correct estimate of the probability density function of an unknown stochatic process is a preli... more A correct estimate of the probability density function of an unknown stochatic process is a preliminary step of utmost importance for any subsequent elaboration stages, such as modelling and classification. Traditional approaches are based on the preliminary choice of a mathematical model of the function and subsequent fitting on its parameters. Therefore some a-priori knowledge and/or assumptions on the phenomenon under consideration are required. Here an alternative approach is presented, which does not require any assumption on the available data, but extracts the probability density function from the output of a neural network, that is trained with a suitable database including the original data and some ad hoc created data with known distribution. This approach has been tested on a synthetic and on an industrial dataset and the obtained results are presented and discussed.
Communications in Computer and Information Science, 2012
In binary classification problems, in presence of unbalanced datasets, the detection of rare patt... more In binary classification problems, in presence of unbalanced datasets, the detection of rare patterns is a difficult task due to several interacting factors which affect the performance of standard classifiers. In this paper a novel approach to this problem is presented. The described method tries to overcome the criticalities encountered by standard methods and by some systems expressly developed to face this problem by means of a dynamic resampling technique, which suitably resamples the training dataset by means of a feed–forward neural network counterbalancing the natural distribution of the dataset. The proposed method has been tested on literature and industrial datasets: the achieved encouraging results are presented and discussed in the paper.
2008 IEEE Aerospace Conference, 2008
Many universities are now involved in projects related to design, assembly, and operation of smal... more Many universities are now involved in projects related to design, assembly, and operation of small satellites. These projects, with participation of researchers and students, and support of external companies, do not aim to compete with commercial satellites; the main goal is to increase the experience level which contributes to make space applications affordable also to small organizations. For students, participation in the complete process of satellite design, assembly, and testing, offers a unique experience within an interdisciplinary complex project. External companies are involved in creating a community of researchers focused on space applications, thus creating new markets and opportunities. The paper describes the architecture and design solutions of a small satellite developed at Politecnico di Torino in the above mentioned context. The main design goal was to combine the usually conflicting cost and reliability constraints; cost has been limited by using properly selected COTS (Commercial Off The Shelf) devices. Reliability has been achieved through redundancy and design diversity. Focus of the paper is on overall satellite design and architecture, with details of solutions to enhance reliability down to the hardware level. The experience led to the development of a new course (Master level) and to several new projects currently under way. 1 2 TABLE OF CONTENTS 1
ISDA 2009 - 9th International Conference on Intelligent Systems Design and Applications, 2009
ABSTRACT When designing a neural or fuzzy system, a careful preprocessing of the database is of u... more ABSTRACT When designing a neural or fuzzy system, a careful preprocessing of the database is of utmost importance in order to produce a trustable system. In function approximation applications, when a functional relationship between input and output variables is supposed to exist, the presence of data where the similar set of input variables is associated to very different values of the output is not always beneficial for the final system to design. A method is presented which can be used to detect anomalous data, namely non-coherent associations between input and output patterns. This technique, by mean of a comparison between two distance matrix associated to the input and output patterns, is able to detect elements in a dataset, where similar values of input variables are associated to quite different output values. A numerical example and a more complex application in the pre-processing of data coming from an industrial database were presented.
This paper describes a silicon implementation of an Arti cial Neural Networks based on Coherent P... more This paper describes a silicon implementation of an Arti cial Neural Networks based on Coherent Pulse Width modulation techniques. Synapses use current generators controlled by an input Pulse Stream. Net charge generated is the product of synaptic current by pulse width. Neurons accumulate synaptic contributions and convert internal activation into an output Pulse Stream. A system optimized for lowest computation energy and highest recon gurability has been designed, manufactured and tested.
IEEE Transactions on Dependable and Secure Computing, 2014
Integrity assurance of configuration data has a significant impact on microcontroller-based syste... more Integrity assurance of configuration data has a significant impact on microcontroller-based systems reliability. This is especially true when running applications driven by events which behavior is tightly coupled to this kind of data. This work proposes a new hybrid technique that combines hardware and software resources for detecting and recovering soft-errors in system configuration data. Our approach is based on the utilization of a common built-in microcontroller resource (timer) that works jointly with a software-based technique, which is responsible to periodically refresh the configuration data. The experiments demonstrate that non-destructive single event effects can be effectively mitigated with reduced overheads. Results show an important increase in fault coverage for SEUs and SETs, about one order of magnitude.
Uploads
Papers by Leonardo Reyneri