Biography:

In the past Beatrice Lazzerini has collaborated on articles with Mario G.C.A. Cimino and Fabio Di Francesco. One of their most recent publications is Using multilayer perceptrons as receptive fields in the design of neural networks. Which was published in journal Neurocomputing.

More information about Beatrice Lazzerini research including statistics on their citations can be found on their Copernicus Academic profile page.

Beatrice Lazzerini's Articles: (4)

Using multilayer perceptrons as receptive fields in the design of neural networks

AbstractIn this paper, we propose a new neural network architecture based on a family of referential multilayer perceptrons (RMLPs) that play a role of generalized receptive fields. In contrast to “standard” radial basis function (RBF) neural networks, the proposed topology of the network offers a considerable level of flexibility as the resulting receptive fields are highly diversified and capable of adjusting themselves to the characteristics of the locally available experimental data. We discuss in detail a design strategy of the novel architecture that fully exploits the modeling capabilities of the contributing RMLPs. The strategy comprises three phases. In the first phase, we form a “blueprint” of the network by employing a specialized version of the commonly encountered fuzzy C-means (FCM) clustering algorithm, namely the conditional (context-based) FCM. In this phase our intent is to generate a collection of information granules (fuzzy sets) in the space of input and output variables, narrowed down to some certain contexts. In the second phase, based upon a global view at the structure, we refine the input–output relationships by engaging a collection of RMLPs where each RMLP is trained by using the subset of data associated with the corresponding context fuzzy set. During training each receptive field focuses on the characteristics of these locally available data and builds a nonlinear mapping in a referential mode. Finally, the connections of the receptive fields are optimized through global minimization of the linear aggregation unit located at the output layer of the overall architecture. We also include a series of numeric experiments involving synthetic and real-world data sets which provide a thorough comparative analysis with standard RBF neural networks.

An adaptive rule-based approach for managing situation-awareness

AbstractSituation awareness is a powerful paradigm that can efficiently exploit the increasing capabilities of handheld devices, such as smart phones and PDAs. Indeed, accurate understanding of the current situation can allow the device to proactively provide information and propose services to users in mobility. Of course, to recognize the situation is a challenging task, due to such factors as the variety of possible situations, uncertain and imprecise data, and different user’s preferences and behavior.In this framework, we propose a robust and general rule-based approach to manage situation awareness. We adopt Semantic Web reasoning, fuzzy logic modeling, and genetic algorithms to handle, respectively, situational/contextual inference, uncertain input processing, and adaptation to the user’s behavior. We exploit an agent-oriented architecture so as to provide both functional and structural interoperability in an open environment. The system is evaluated by means of a real-world case study concerning resource recommendation. Experimental results show the effectiveness of the proposed approach.

Profiling risk sensibility through association rules

AbstractIn the last recent years several approaches to risk assessment and risk management have been adopted to reduce the potential for specific risks in working environments. A safety culture has also developed to let workers acquire knowledge and understanding of risks and safety. Notwithstanding, risks still exist in every workplace. One effective way to improve workers’ sensibility to risk, i.e., their ability to effectively assess and control the risks they are exposed to, is risk management training. Unfortunately, people may perceive risks in different ways depending on subjective assessment of the characteristics and severity of the considered risks, and may have tendencies to either take or avoid actions that they feel are risky. Therefore, the knowledge of how workers assess each of the risks they may be exposed to in the workplace is a key factor to conceive effective custom risk management training. In this paper we present a novel approach, based on association rules, to workers’ profiling with respect to risk perception and risk propensity in order to provide each of them with specific customized risk management training.

An electronic nose for odour annoyance assessment

AbstractAlthough in most cases annoying atmospheric emissions do not menace public health, they are less and less tolerated because of the effects on quality of life. Several approaches have been proposed to face this problem but none of them offers a completely satisfying solution. The development of electronic noses, which promise to mimic human sense of smell by means of a sensor array and a pattern recognition model, offers new interesting perspectives. In this paper, an electronic nose based on conducting polymer sensors and a fuzzy logic-based pattern recognition system is tested with waste water samples, obtaining 87% recognition rate on the test set. Current limits of this new technology are discussed and a strategy for their overcoming is proposed.

Advertisement
Join Copernicus Academic and get access to over 12 million papers authored by 7+ million academics.
Join for free!

Contact us