Showing papers by "fondazione bruno kessler published in 2007"

High variability design for software agents: Extending Tropos

Abstract: Many classes of distributed applications, including e-business, e-government, and ambient intelligence, consist of networking infrastructures, where the nodes (peers)—be they software components, human actors or organizational units—cooperate with each other to achieve shared goals. The multi-agent system metaphor fits very well such settings because it is founded on intentional and social concepts and mechanisms. Not surprisingly, many agent-oriented software development methods have been proposed, including GAIA, PASSI, and Tropos. This paper extends the Tropos methodology, enhancing its ability to support high variability design through the explicit modelling of alternatives, it adopts an extended notion of agent capability and proposes a refined Tropos design process. The paper also presents an implemented software development environment for Tropos, founded on the Model-Driven Architecture (MDA) framework and standards. The extended Tropos development process is illustrated through a case study involving an e-commerce application.

Efficient Handling of N-gram Language Models for Statistical Machine Translation

Abstract: Statistical machine translation, as well as other areas of human language processing, have recently pushed toward the use of large scale n-gram language models. This paper presents efficient algorithmic and architectural solutions which have been tested within the Moses decoder, an open source toolkit for statistical machine translation. Experiments are reported with a high performing baseline, trained on the Chinese-English NIST 2006 Evaluation task and running on a standard Linux 64-bit PC architecture. Comparative tests show that our representation halves the memory required by SRI LM Toolkit, at the cost of 44% slower translation speed. However, as it can take advantage of memory mapping on disk, the proposed implementation seems to scale-up much better to very large language models: decoding with a 289-million 5-gram language model runs in 2.1Gb of RAM.

Single photon timing resolution and detection efficiency of the IRST silicon photo-multipliers

Abstract: Silicon photo-multipliers (SiPM) consist in matrices of tiny, passive quenched avalanche photo-diode cells connected in parallel via integrated resistors and operated in Geiger mode. Novel types of SiPM are being developed at FBK-IRST (Trento, Italy). Despite their classical shallow junction n-on-p structure the devices are unique in their enhanced photo-detection efficiency (PDE) for short-wavelengths and in their low level of dark rate and excess noise factor. After a summary of the extensive SiPM characterization we will focus on the study of PDE and the single photon timing resolution.

A multimodal annotated corpus of consensus decision making meetings

Abstract: In this paper we present an annotated audio–video corpus of multi-party meetings. The multimodal corpus provides for each subject involved in the experimental sessions six annotation dimensions referring to group dynamics; speech activity and body activity. The corpus is based on 11 audio and video recorded sessions which took place in a lab setting appropriately equipped with cameras and microphones. Our main concern in collecting this multimodal corpus was to explore the possibility of providing feedback services to facilitate group processes and to enhance self awareness among small groups engaged in meetings. We therefore introduce a coding scheme for annotating relevant functional roles that appear in a small group interaction. We also discuss the reliability of the coding scheme and we present the first results for automatic classification.

Mining the Lexicon Used by Programmers during Sofware Evolution

Abstract: Identifiers represent an important source of information for programmers understanding and maintaining a system. Self-documenting identifiers reduce the time and effort necessary to obtain the level of understanding appropriate for the task at hand. While the role of the lexicon in program comprehension has long been recognized, only a few works have studied the quality and enhancement of the identifiers and no works have studied the evolution of the lexicon. In this paper, we characterize the evolution of program identifiers in terms of stability metrics and occurrences of renaming. We assess whether an evolution process similar to the one occurring for the program structure exists for identifiers. We report data and results about the evolution of three large systems, for which several releases are available. We have found evidence that the evolution of the lexicon is more limited and constrained than the evolution of the structure. We argue that the different evolution results from several factors including the lack of advanced tool support for lexicon construction, documentation, and evolution.

Results of the IGEC-2 search for gravitational wave bursts during 2005

Abstract: The network of resonant bar detectors of gravitational waves resumed coordinated observations within the International Gravitational Event Collaboration (IGEC-2). Four detectors are taking part in this Collaboration: ALLEGRO, AURIGA, EXPLORER and NAUTILUS. We present here the results of the search for gravitational wave bursts over 6 months during 2005, when IGEC-2 was the only gravitational wave observatory in operation. The implemented network data analysis is based on a time coincidence search among AURIGA, EXPLORER and NAUTILUS; ALLEGRO data was reserved for follow-up studies. The network amplitude sensitivity to bursts improved by a factor ≈3 over the 1997-2000 IGEC observations; the wider sensitive band also allowed the analysis to be tuned over a larger class of waveforms. Given the higher single-detector duty factors, the analysis was based on threefold coincidence, to ensure the identification of any single candidate of gravitational waves with high statistical confidence. The false detection rate was as low as 1 per century. No candidates were found.

Tool-supported development with Tropos: the conference management system case study

Abstract: The agent-oriented software engineering methodology Troposoffers a structured development process and supporting tools fordeveloping complex, distributed systems. The objective of this paper is twofold: first, to illustrate the use ofTropos to develop a Multi-Agent System, performing basic analysis anddesign activities, code generation and testing, with the support of a setof tools; second, to enable the comparison with other, tool-supported,agent-oriented software engineering methodologies through a descriptionof the main steps of these activities and of excerpts of the resultingartefacts, with reference to a common case study, namely, the ConferenceManagement System case study.

Multimodal corpus of multi-party meetings for automatic social behavior analysis and personality traits detection

Abstract: This paper describes an automatically annotated multimodal corpus of multi-party meetings. The corpus provides for each subject involved in the experimental sessions information on her/his social behavior and personality traits, as well as audiovisual cues (speech rate, pitch and energy, head orientation, head, hand and body fidgeting). The corpus is based on the audio and video recordings of thirteen sessions, which took place in a lab setting equipped with cameras and microphones. Our main concern in collecting this corpus was to investigate the possibility of creating a system capable of automatically analyzing social behaviors and predicting personality traits using audio-visual cues.

An Empirical Study to Compare the Accuracy of AHP and CBRanking Techniques for Requirements Prioritization

Abstract: Requirements prioritization aims at identifying the most important requirements for a system (or a release). A large number of approaches have been proposed so far, to help decision makers in performing this activity. Some of them provide supporting tools. Questions on when a prioritization technique should be preferred to another one as well as on how to characterize and measure their properties arise. Several empirical studies have been conducted to analyze characteristics of the available approaches, but their results are often difficult to compare. In this paper we discuss an empirical study aiming at evaluating two state-of-the art, tool-supported requirements prioritization techniques, AHP and CBRanking. The experiment has been conducted with 18 experienced subjects on a set of 20 requirements from a real project. We focus on a crucial variable, namely the ranking accuracy. We discuss different ways to measure it and analyze the data collected in the experimental study with reference to this variable. Results indicate that AHP gives more accurate rankings than CBRanking, but the ranks produced by the two methods are similar for all the involved subjects.

Silicon photomultiplier performance tests in magnetic resonance pulsed fields

Abstract: We are investigating the performance of silicon photomultipliers (SiPM) in a magnetic resonance imaging (MRI) environment with a view to exploiting the potential advantages in a combined positron emission tomography / magnetic resonance imaging (PET/MR) system SiPMs are attractive for PET because they are compact, provide high gain at low bias voltage, and fast response times Importantly for combined PET/MR applications, they offer performance comparable to that of the photomultiplier tube (PMT) and are insensitive to magnetic fields To study the effect of MR static magnetic, pulsed gradient and radio-frequency fields on SiPM performance, we have acquired data whilst switching magnetic fields with the SiPM positioned in the MR system The wide 'dynamic range' of SiPMs facilitates their use in both low light level and PET applications; we have acquired 'single photoelectron' and 22Na energy spectra to demonstrate SiPM performance for this application As expected, no degradation in inherent SiPM performance in an MR environment is observed

Empirical Validation of a Web Fault Taxonomy and its usage for Fault Seeding

Abstract: The increasing demand for reliable Web applications gives a central role to Web testing. Most of the existing works are focused on the definition of novel testing techniques, specifically tailored to the Web. However, no attempt was carried out so far to understand the specific nature of Web faults. This is of fundamental importance to assess the effectiveness of the proposed Web testing techniques. In this paper, we describe the process followed in the construction of a Web fault taxonomy. After the initial, top- down construction, the taxonomy was subjected to four iterations of empirical validation aimed at refining it and at understanding its effectiveness in bug classification. The final taxonomy is publicly available for consultation and editing on a Wiki page. Testers can use it in the definition of test cases that target specific classes of Web faults. Researchers can use it to build fault seeding tools that inject artificial faults which resemble the real ones.

RAT: a tool for the formal analysis of requirements

Abstract: Formal languages are increasingly used to describe the functional requirements of circuits. Although formal requirements can be hard to understand and subtle, they are seldom the object of verification. In this paper we present our requirement analysis tool, RAT. Our tool supports quality assurance of formal specifications. A designer can interactively explore the requirements' semantics and automatically check the specification against assertions (which must be satisfied) and possibilities (which describe allowed corner-case behavior). Using RAT, a designer can also investigate the realizability of a specification. RAT was successfully examined in several industrial projects.

Preliminary results from a current mode CMOS front-end circuit for silicon photomultiplier detectors

Abstract: We propose a CMOS front-end circuit suitable for Silicon Photomultiplier detectors (SiPM) based on a current buffer, as input stage, which features small input impedance and large bandwidth, thanks to the application of current feedback techniques The current mode approach enhances the dynamic range of the front-end and does not suffer from possible voltage limitations due to deep-submicron CMOS implementation We report the first measurement results obtained by coupling the circuit prototype to a SiPM detector excited by a blue LED light source The measurements confirm the effectiveness of the proposed front-end approach and demonstrate its capability of managing large current signals with good linearity

A Modular Framework for Ontology-based Representation of Patent Information

Abstract: In this paper, we present a new ontology-based formalism for representing patent information. The framework defines concepts and relations for the major aspects of patent information pertaining to patent metadata, patent content structure, patent semantics, and patent classification. Each aspect is logically covered by an individual ontology module. The paper further discusses aspects of ontology learning and integration of non-textual information for the patent domain.

Photoluminescence of silicon nanocrystals in silicon oxide

Abstract: Recent results on the photoluminescence properties of silicon nanocrystals embedded in silicon oxide are reviewed and discussed. The attention is focused on Si nanocrystals produced by high-temperature annealing of silicon rich oxide layers deposited by plasma-enhanced chemical vapor deposition. The influence of deposition parameters and layer thickness is analyzed in detail. The nanocrystal size can be roughly controlled by means of Si content and annealing temperature and time. Unfortunately, a technique for independently fine tuning the emission efficiency and the size is still lacking; thus, only middle size nanocrystals have high emission efficiency. Interestingly, the layer thickness affects the nucleation and growth kinetics so changing the luminescence efficiency.

Boron deactivation in preamorphized silicon on insulator: Efficiency of the buried oxide as an interstitial sink

Abstract: Preamorphization of ultrashallow implanted boron in silicon on insulator is optimized to produce an abrupt boxlike doping profile with negligible electrical deactivation and significantly reduced transient enhanced diffusion. The effect is achieved by positioning the as-implanted amorphous/crystalline interface close to the buried oxide interface to minimize interstitials while leaving a single-crystal seed to support solid-phase epitaxy. Results support the idea that the interface between the Si overlayer and the buried oxide is an efficient interstitial sink.

Tactile sensing arrays for humanoid robots

Abstract: Development of robots capable of operating in unstructured environments or intended to substitute for man in hazardous or inaccessible environments, demands the implementation of sophisticated sensory capabilities, far beyond those available today. In this regard, the development of tactile sensors is one of the key technical challenges in advanced robotics and minimal access surgery. In this work we present arrays of 'taxels' (tactile elements) which will be placed on the distal phalange of the humanoid robot in our lab. We present two different designs and implementations. In the first one, microelectrode arrays(MEAs) of 32 elements, with 1 mm center to center distance, have been designed. The taxel is implemented by epoxy-adhering the sensing material (piezoelectric polymer film of PVDF-TrFE) on a microelectrode. Each taxel is intended to be used as an extended gate of an FET (external to the chip); the taxel collects the charge/voltage generated, as consequence of the applied stress, on the deposited piezoelectric polymer film (i.e. the extended gate itself). The second design and implementation integrates both the taxels array and the FET devices, on the same silicon die.

Periodic oscillations in transmission decay of anderson localized one-dimensional dielectric systems.

Abstract: It is well recognized that the transmittance of Anderson localized systems decays exponentially on average with sample size, showing large fluctuations brought up by extremely rare occurrences of necklaces of resonantly coupled states, possessing almost unity transmission. We show here that in a one-dimensional (1D) random photonic system with resonant layers these fluctuations appear to be very regular and have a period defined by the localization length xi of the system. We stress that necklace states are the origin of these well-defined oscillations. We predict that in such a random system efficient transmission channels form regularly each time the increasing sample length fits so-called optimal-order necklaces and demonstrate the phenomenon through numerical experiments. Our results provide new insight into the physics of Anderson localization in random systems with resonant units.

Assessing procedural risks and threats in e-voting: challenges and an approach

Abstract: Performing a good security analysis on the design of a system is an essential step in order to guarantee a reasonable level of protection. However, different attacks and threats may be carried out depending on the operational environment in which the system is used, i.e. the procedures that define how to operate the systems. We are interested in reasoning about the security of e-Voting procedures, namely on the risks and attacks that can be carried out during an election. Our focus is more on people and organizations than on systems and technologies. In this paper we describe some ongoing work that we are carrying out within the ProVotE project (a project sponsored by the Autonomous Province of Trento to switch to e-Voting for local elections) to analyze and (possibly) improve procedural security of electronic elections. To do so, we are providing models of the Italian electoral laws using the UML and we are developing a custom methodology for analyzing threats from the models. Our reasoning approach is based on asset mobility, asset values and existence of multiple instances.

Symbolic Implementation of Alternating Automata

Abstract: This paper addresses the challenges of symbolic model checking and language emptiness checking where the specification is given as an alternating Buchi automaton. We introduce a novel version of Miyano and Hayashi's construction that allows us to directly convert an alternating automaton to a polynomially-sized symbolic structure. We thus avoid building an exponentially-sized explicit representation of the corresponding nondeterministic automaton. For one-weak automata, Gastin and Oddoux' construction produces smaller automata than Miyano and Hayashi's construction. We present a (symbolic) hybrid approach that combines the benefits of both: while retaining full generality, it uses the cheaper construction for those parts of the automaton that are one-weak. We performed a thorough experimental comparison of the explicit and symbolic approaches and several variants of Miyano and Hayashi's construction, using both BDD-based and SAT-based model checking techniques. The symbolic approaches clearly outperform the explicit one.

Novel RF-MEMS widely-reconfigurable directional coupler

Abstract: This paper presents an innovative widely reconfigurable directional coupler based on concentric selectable squares. The device can be reconfigured by simply addressing the RF signal to the selected square through independent sets of MEMS switches. A very wide tuning range of the operating frequency can be attained, still maintaining good performance in terms of both return loss and isolation in the all states. A 2-state MEMS prototype has been designed on a 525 mum high resistivity silicon substrate, employing a well-established eight-mask surface micro-machining process developed at Fondazione Bruno Kessler, Trento, Italy. A 41% tuning range has been achieved ranging from C to X band with a return loss and isolation better than 25 dB in both states. Two test circuits have been realized in microstrip technology confirming the feasibility of the proposed architecture.

Formation of optimal-order necklace modes in one-dimensional random photonic superlattices

Abstract: We study the appearance of resonantly coupled optical modes, optical necklaces, in Anderson-localized one-dimensional random systems with resonant layers. It has been stated that the increasingly low probability of their appearance in thick samples should reduce the theoretically predicted number of coupled resonances, leading to the formation of optimal-order necklaces (OONs). Here we examine through transmission phase studies how the optimal order ${m}^{*}$, the number of resonances in a necklace, shifts gradually toward higher orders with increasing sample size. We present a model for OON formation in random samples with resonant layers and derive an empirical formula that predicts ${m}^{*}$. We discuss the situation when in a sample length $L$ the number of resonances degenerate in energy exceeds the optimal one, and show how the extra resonances are pushed out to the miniband edges of the necklace, reducing the order of the latter always by multiples of two.

New developments on 3D detectors at IRST

Abstract: We report on the latest results from the development of 3D silicon radiation detectors at IRST (Trento, Italy) A new detector concept has been defined, namely 3D-DDTC (Double-side Double-Type Column): it involves alternate etching of columnar electrodes of both doping types from both wafer sides, and stopping at a short distance (d) from the opposite surface Simulations prove that, if d is kept small with respect to the wafer thickness, this approach can yield charge collection properties comparable to those of standard 3D detectors, with the advantage of a simpler fabrication process Two wafer layouts (including strip detectors, pixel detectors, and test structures) have been designed with reference to this technology, and two fabrication runs have been fabricated Simulation results and preliminary experimental results from the characterization of test structures from the first 3D-DDTC batch are reported

Uphill diffusion of ultralow-energy boron implants in preamorphized silicon and silicon-on-insulator

Abstract: Redistribution during annealing of low-energy boron (B) implants in silicon on insulator (SOI) structures and in bulk Si has been investigated by comparing secondary ion mass spectrometry (SIMS) and simulated profiles. All the samples have been preamorphized with Ge at different implantation energies in order to investigate the effects of the position of the damage on B diffusion. Different B doses in the range between 2×1013 and 2×1015 cm−2 and annealing temperatures between 700 and 1100 °C have been investigated. All SIMS profiles show a B pileup in the first few nanometers of the Si matrix in proximity of the Si surface. The results of our simulations, performed on samples implanted at different doses (below and above the solid solubility), indicate that the B redistribution upon annealing can be explained with a simple model which considers the presence of traps in the surface region, without considering any asymmetric behavior of the dopant diffusion. The sink region is a few monolayers (1–2 nm) for ...

Application of the BJT detector for simple, low-cost, and low-power alpha-particle detection systems

Abstract: We present a simple, low-cost, and low-power alpha-particle detection system for environmental radioactivity monitoring. The system exploits a previuosly-developed high- resistivity-silicon detector with internal amplification capability based on the bipolar-transistor (BJT) effect and readout electronics based on commercial IC's. Two-dimensional numerical device simulations are adopted to assess the feasibility of the BJT detector as an alpha-particle detector that can be operated, without losing its internal signal amplification capability, with floating base and low collector voltages, so that device technology can be kept simple, very small DC power consumption can be achieved, and a single 5-V power-supply voltage can be used for readout electronics and detector biasing. The charge amplification accomplished by the BJT detector allows a single, commercial chip to be adopted, to perform charge preamplification and 20-bit A/D conversion. The digital output is sent to a low-cost microcontroller that can be periodically interrogated through the IR port. The cost of the readout electronics is in the order of 60$ and it can operate with standard Li-ion battery for about 60 hours.

Semantic domains and supersense tagging for domain-specific ontology learning

Abstract: In this paper we propose a novel unsupervised approach to learning domain-specific ontologies from large open-domain text collections. The method is based on the joint exploitation of Semantic Domains and Super Sense Tagging for Information Retrieval tasks. Our approach is able to retrieve domain specific terms and concepts while associating them with a set of high level ontological types, named supersenses, providing flat ontologies characterized by very high accuracy and pertinence to the domain.

A collision-free time-to-first spike camera architecture based on a winner-take-all network

Abstract: A new type of asynchronous time-based digital camera is here presented. The sensor architecture is characterized by a new analog low-power winner take all (WTA) network shared among all the pixels of the array. This circuit detects the most illuminated pixel and transmits it through the channel by queuing the other competitors. An iterative algorithm allows to read-out the entire array, temporally dispatching the information by means of an asynchronous read-out circuit placed at column and row level. The pixel, designed with a 0.35 mum CMOS technology and consisting of 28 transistors, performs a simulated average power consumption less than 10 nW @ 1.8 V for an impinging light power of 100 muW/cm2.

Refining goal models by evaluating system behaviour

Abstract: Nowadays, information systems have to perform in complex, heterogeneousenvironments, considering a variety of system users with different needsand preferences. Software engineering methodologies need to cope with the complexityof requirements specification in such scenarios, where new requirementsmay emerge also at run-time and the system's goals are expected to evolve tomeet new stakeholder needs. Following an agent-oriented approach, we are studying methods and techniquesto design adaptive and evolvable information systems able to fulfill stakeholders'objectives. In a previous work we defined an Agent-Oriented framework to design andcode system specifications in terms of goal models and we instantiated it in atool supported process which exploits the Agent-Oriented Software Engineeringmethodology Tropos and the Multi-Agent Platform JADE/Jadex [11]. In this paper, we show how to use this framework to develop a system followingan iterative process, where the system execution allows enriching the systemspecification given in terms of goal models. Experimental evaluation has been performed on a simple example and leadto the refinement of the designed goal model upon the analysis of the system'srun-time behaviour.