Manno, Switzerland


Manno, Switzerland
Time filter
Source Type

Harou J.J.,University of Manchester | Garrone P.,Polytechnic of Milan | Rizzoli A.E.,IDSIA USI SUPSI | Maziotis A.,University of Manchester | And 5 more authors.
Procedia Engineering | Year: 2014

The SmartH2O project aims to provide water utilities, municipalities and citizens with an ICT enabled platform to design, develop and implement better water management policies using innovative metering, social media and pricing mechanisms. This project has as a working hypothesis that high data quality obtained from smart meters and communicable through social media and other forms of interaction could be used to design and implement innovative and effective water pricing policies. Planned case studies in the UK and Switzerland are introduced. We anticipate that SmartH20 research outcomes will be of use to those interested in linking smart metering, social media and smart pricing approaches to achieve more sustainable water management outcomes. © 2014 The Authors.

Rizzoli A.E.,IDSIA USI SUPSI | Castelletti A.,Polytechnic of Milan | Cominola A.,Polytechnic of Milan | Fraternali P.,Polytechnic of Milan | And 6 more authors.
Proceedings - 7th International Congress on Environmental Modelling and Software: Bold Visions for Environmental Modeling, iEMSs 2014 | Year: 2014

SmartH2O is an EU funded project which aims at creating a virtuous feedback cycle between water users and the utilities, providing users information on their consumption in quasi real time, and thus enabling water utilities to plan and implement strategies to reduce/reallocate water consumption. Traditional metering data, usually gathered twice a year, can be used to model consumers' behaviour at an aggregate level, but the motivations and individual attitudes of consumers are hidden. The advent of smart water meters allows gathering high frequency consumption data that can be used to provide instantaneous information to water utilities on the state of the network. At the same time, the consumption information can be fed back to the user to stimulate increased awareness on water use. The SmartH2O project aims at developing methodologies to involve consumers and promote water savings by increasing their awareness, using a social computing approach, and also exploring their sensitivity to water prices, e.g., to penalise water waste during droughts. In this paper, first we review similar experiences that exploit consumer awareness to reduce consumption, then we review the role of persuasive games for sustainability, and finally we present the SmartH2O approach, sketching the architecture of its modelling and social computing components.

Rizzoli A.E.,IDSIA USI SUPSI | Rudel R.,University of Applied Sciences and Arts Southern Switzerland | Forster A.,University of Applied Sciences and Arts Southern Switzerland | Corani G.,IDSIA USI SUPSI | And 4 more authors.
Proceedings - 7th International Congress on Environmental Modelling and Software: Bold Visions for Environmental Modeling, iEMSs 2014 | Year: 2014

Diffusion of smart mobile devices offers unprecedented opportunities to monitor travel behaviour, by means of the GPS devices they are equipped with: using a suitable application, potentially every smartphone owner can produce huge, inexpensive quantities of data suitable to profile her mobility patterns. We take advantage of this opportunity within the e-mobiliTI project, which aims at analysing the main psychological and behavioural barriers affecting the transition to new mobility solutions. The project sets up a "living lab" made up of around twenty families and gives them the opportunity to test electric cars and bikes, public transport season tickets and car and bike sharing. Beside traditional social research tools (questionnaires, interviews, focus groups), to analyse their mobility behaviour we use a specifically developed smartphone application. In this paper we present the results of the first phase of our field trial and discuss the major challenges faced so far in the automatic gathering of mobility data: high battery consumption, limited performances of the GPS smartphone devices, problems in the Internet connectivity, limited reliability of the information the application asks to the users and risk that the users quit using the application, due to the lack of immediate compensation for the nuisance of being always monitored and for the daily effort of actively using the application.

Cuff W.,Link Forecasting Inc. | Sands P.,39 Oakleigh Ave. | Benyon M.,61 27 Rangers Road | Rizzoli A.E.,IDSIA USI SUPSI
Environmental Modelling and Software | Year: 2011

We present an electronic book titled " Simulation and Modelling of System Dynamics" , just recently published online by the International Environmental Modelling & Software Society (iEMSs), and authored by the late Peter Benyon.Simulation allows exploration of more complex models than can be analysed theoretically; it can be used where trials on the real world system would be too costly, time consuming or dangerous. This book provides for the needs of those undertaking simulation.The book aims at a wide readership using an informal style. The material varies from elementary to advanced and encourages selective browsing and reading, with a wide range of examples and chapter summaries to aid this choice.The material focuses on the need to formulate the models with the right structure and equations - in other words, it concentrates on system description rather than model solution. The importance of subsequent validation is indicated.The book is freely accessible and it is available online as a PDF file at: © 2011 Elsevier Ltd.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2012

We show that Incremental Slow Feature Analysis (IncSFA) provides a low complexity method for learning Proto-Value Functions (PVFs). It has been shown that a small number of PVFs provide a good basis set for linear approximation of value functions in reinforcement environments. Our method learns PVFs from a high-dimensional sensory input stream, as the agent explores its world, without building a transition model, adjacency matrix, or covariance matrix. A temporal-difference based reinforcement learner improves a value function approximation upon the features, and the agent uses the value function to achieve rewards successfully. The algorithm is local in space and time, furthering the biological plausibility and applicability of PVFs. © 2012 Springer-Verlag.

Neural Networks | Year: 2012

We describe the approach that won the final phase of the German traffic sign recognition benchmark. Our method is the only one that achieved a better-than-human recognition rate of 99.46%. We use a fast, fully parameterizable GPU implementation of a Deep Neural Network (DNN) that does not require careful design of pre-wired feature extractors, which are rather learned in a supervised way. Combining various DNNs trained on differently preprocessed data into a Multi-Column DNN (MCDNN) further boosts recognition performance, making the system insensitive also to variations in contrast and illumination. © 2012 Elsevier Ltd.

Masci J.,IDSIA USI SUPSI | Angulo J.,MINES ParisTech | Schmidhuber J.,IDSIA USI SUPSI
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2013

We present a novel framework for learning morphological operators using counter-harmonic mean. It combines concepts from morphology and convolutional neural networks. A thorough experimental validation analyzes basic morphological operators dilation and erosion, opening and closing, as well as the much more complex top-hat transform, for which we report a real-world application from the steel industry. Using online learning and stochastic gradient descent, our system learns both the structuring element and the composition of operators. It scales well to large datasets and online settings. © 2013 Springer-Verlag.

Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition | Year: 2012

Traditional methods of computer vision and machine learning cannot match human performance on tasks such as the recognition of handwritten digits or traffic signs. Our biologically plausible, wide and deep artificial neural network architectures can. Small (often minimal) receptive fields of convolutional winner-take-all neurons yield large network depth, resulting in roughly as many sparsely connected neural layers as found in mammals between retina and visual cortex. Only winner neurons are trained. Several deep neural columns become experts on inputs preprocessed in different ways; their predictions are averaged. Graphics cards allow for fast training. On the very competitive MNIST handwriting benchmark, our method is the first to achieve near-human performance. On a traffic sign recognition benchmark it outperforms humans by a factor of two. We also improve the state-of-the-art on a plethora of common image classification benchmarks. © 2012 IEEE.

Loading IDSIA USI SUPSI collaborators
Loading IDSIA USI SUPSI collaborators