Publications
OUR RESEARCH
Scientific Publications
Here you can find the comprehensive list of publications from the members of the Research Center on Computer Vision and eXtended Reality (xRAI).
Use the tag cloud to filter papers based on specific research topics, or use the menus to filter by year, type of publication, or authors.
For each paper, you have the option to view additional details such as the Abstract, Links, and BibTex record.
Research is formalized curiosity. It is poking and prying with a purpose
Zora Neale Hurston
2018
Generosi, Andrea; Ceccacci, Silvia; Mengoni, Maura
A deep learning-based system to track and analyze customer behavior in retail store Proceedings Article
In: 2018 IEEE 8th International Conference on Consumer Electronics - Berlin (ICCE-Berlin), pp. 1–6, IEEE, Berlin, 2018, ISBN: 978-1-5386-6095-9.
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Computer Vision and Pattern Recognition, Deep Learning, Emotion Recognition
@inproceedings{generosi_deep_2018,
title = {A deep learning-based system to track and analyze customer behavior in retail store},
author = {Andrea Generosi and Silvia Ceccacci and Maura Mengoni},
url = {https://ieeexplore.ieee.org/document/8576169/},
doi = {10.1109/ICCE-Berlin.2018.8576169},
isbn = {978-1-5386-6095-9},
year = {2018},
date = {2018-01-01},
urldate = {2024-12-28},
booktitle = {2018 IEEE 8th International Conference on Consumer Electronics - Berlin (ICCE-Berlin)},
pages = {1–6},
publisher = {IEEE},
address = {Berlin},
abstract = {The present work introduces an emotional tracking system to monitor Shopping Experience at different touchpoints in a retail store, based on the elaboration of the information extracted from biometric data and facial expressions. A preliminary test has been carried out to determine the system effectiveness in a real context regarding to emotion detection and customers' sex, age and ethnicity discrimination. To this end, information provided by the system have been compare with the results of a traditional video analysis. Results suggest that the proposed system can be effectively used to support the analysis of customer experience in a retail context.},
keywords = {Artificial Intelligence, Computer Vision and Pattern Recognition, Deep Learning, Emotion Recognition},
pubstate = {published},
tppubtype = {inproceedings}
}
The present work introduces an emotional tracking system to monitor Shopping Experience at different touchpoints in a retail store, based on the elaboration of the information extracted from biometric data and facial expressions. A preliminary test has been carried out to determine the system effectiveness in a real context regarding to emotion detection and customers' sex, age and ethnicity discrimination. To this end, information provided by the system have been compare with the results of a traditional video analysis. Results suggest that the proposed system can be effectively used to support the analysis of customer experience in a retail context.
Gargiulo, Francesco; Silvestri, Stefano; Fontanella, Mariarosaria; Ciampi, Mario; Pietro, Giuseppe De
A Deep Learning Approach for Scientific Paper Semantic Ranking Proceedings Article
In: Pietro, Giuseppe De; Gallo, Luigi; Howlett, Robert J.; Jain, Lakhmi C. (Ed.): Intelligent Interactive Multimedia Systems and Services 2017, pp. 471–481, Springer International Publishing, Cham, 2018, ISBN: 978-3-319-59480-4.
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Deep Learning, Word Embeddings
@inproceedings{gargiulo_deep_2018-1,
title = {A Deep Learning Approach for Scientific Paper Semantic Ranking},
author = {Francesco Gargiulo and Stefano Silvestri and Mariarosaria Fontanella and Mario Ciampi and Giuseppe De Pietro},
editor = {Giuseppe De Pietro and Luigi Gallo and Robert J. Howlett and Lakhmi C. Jain},
doi = {10.1007/978-3-319-59480-4_47},
isbn = {978-3-319-59480-4},
year = {2018},
date = {2018-01-01},
booktitle = {Intelligent Interactive Multimedia Systems and Services 2017},
pages = {471–481},
publisher = {Springer International Publishing},
address = {Cham},
abstract = {In this paper we proposed a novel Deep Learning approach to realize a Word Embeddings (WEs) similarity based search tool, considering the medical literature as case study. Using the compositional properties of the WEs we defined a methodology to aggregate the information coming from each word to obtain a vector corresponding to the abstracts of each PubMed article. Through this paradigm it is possible to capture the semantic content of the papers and, consequently, to evaluate and rank the similarity among them. The preliminary results with the proposed approach are obtained analysing a subset of the whole the PubMed collection. The results correctness has been verified by human domain experts, showing that the methodology is promising.},
keywords = {Artificial Intelligence, Deep Learning, Word Embeddings},
pubstate = {published},
tppubtype = {inproceedings}
}
In this paper we proposed a novel Deep Learning approach to realize a Word Embeddings (WEs) similarity based search tool, considering the medical literature as case study. Using the compositional properties of the WEs we defined a methodology to aggregate the information coming from each word to obtain a vector corresponding to the abstracts of each PubMed article. Through this paradigm it is possible to capture the semantic content of the papers and, consequently, to evaluate and rank the similarity among them. The preliminary results with the proposed approach are obtained analysing a subset of the whole the PubMed collection. The results correctness has been verified by human domain experts, showing that the methodology is promising.
Cristani, Marco; Cristani, Matteo; Pesarin, Anna; Tomazzoli, Claudio; Zorzi, Margherita
Making Sentiment Analysis Algorithms Scalable Proceedings Article
In: Proceedings 4th International Workshop on Knowledge Discovery on the Web, pp. 136–147, Springer, 2018.
Abstract | Links | BibTeX | Tags: Machine Learning, Natural Language Processing, Sentiment analysis
@inproceedings{cristani_making_2018,
title = {Making Sentiment Analysis Algorithms Scalable},
author = {Marco Cristani and Matteo Cristani and Anna Pesarin and Claudio Tomazzoli and Margherita Zorzi},
url = {https://link.springer.com/chapter/10.1007%2F978-3-030-03056-8_12},
doi = {10.1007/978-3-030-03056-8_12},
year = {2018},
date = {2018-01-01},
booktitle = {Proceedings 4th International Workshop on Knowledge Discovery on the Web},
volume = {11153},
pages = {136–147},
publisher = {Springer},
abstract = {In this paper we introduce a simplified approach to sentiment analysis: a lexicon-driven method based upon only adjectives and adverbs. This method is compared in cross-validation with other known techniques and then compared directly to the gold standard, a sample of human subjects asked to deliver the same class of judgments computed by the method. We prove that the method is similar in accuracy and precision with the other methods. We finally argue that the approach we employ is more valid than others for it is scalable, and exportable to languages other than English.},
keywords = {Machine Learning, Natural Language Processing, Sentiment analysis},
pubstate = {published},
tppubtype = {inproceedings}
}
In this paper we introduce a simplified approach to sentiment analysis: a lexicon-driven method based upon only adjectives and adverbs. This method is compared in cross-validation with other known techniques and then compared directly to the gold standard, a sample of human subjects asked to deliver the same class of judgments computed by the method. We prove that the method is similar in accuracy and precision with the other methods. We finally argue that the approach we employ is more valid than others for it is scalable, and exportable to languages other than English.
Cristani, Matteo; Domenichini, Francesco; Olivieri, Francesco; Tomazzoli, Claudio; Zorzi, Margherita
It Could Rain: Weather Forecasting as a Reasoning Process Proceedings Article
In: Proceedings of 22nd International Conference on Knowledge-Based and Intelligent Information & Engineering Systems, pp. 850–859, Elsevier, 2018.
Abstract | Links | BibTeX | Tags: Automatic reasoning, Defeasible logic, Machine Learning
@inproceedings{cristani_it_2018,
title = {It Could Rain: Weather Forecasting as a Reasoning Process},
author = {Matteo Cristani and Francesco Domenichini and Francesco Olivieri and Claudio Tomazzoli and Margherita Zorzi},
url = {https://www.sciencedirect.com/science/article/pii/S1877050918312973},
doi = {10.1016/j.procS.2018.08.019},
year = {2018},
date = {2018-01-01},
booktitle = {Proceedings of 22nd International Conference on Knowledge-Based and Intelligent Information & Engineering Systems},
volume = {126},
pages = {850–859},
publisher = {Elsevier},
series = {PROCEDIA COMPUTER SCIENCE},
abstract = {Meteorological forecasting is the process of providing reliable prediction about the future weathear within a given interval of time. Forecasters adopt a model of reasoning that can be mapped onto an integrated conceptual framework. A forecaster essentially precesses data in advance by using some models of machine learning to extract macroscopic tendencies such as air movements, pressure, temperature, and humidity differentials measured in ways that depend upon the model, but fundamentally, as gradients. Limit values are employed to transform these tendencies in fuzzy values, and then compared to each other in order to extract indicators, and then evaluate these indicators by means of priorities based upon distance in fuzzy values. We formalise the method proposed above in a workflow of evaluation steps, and propose an architecture that implements the reasoning techniques.},
keywords = {Automatic reasoning, Defeasible logic, Machine Learning},
pubstate = {published},
tppubtype = {inproceedings}
}
Meteorological forecasting is the process of providing reliable prediction about the future weathear within a given interval of time. Forecasters adopt a model of reasoning that can be mapped onto an integrated conceptual framework. A forecaster essentially precesses data in advance by using some models of machine learning to extract macroscopic tendencies such as air movements, pressure, temperature, and humidity differentials measured in ways that depend upon the model, but fundamentally, as gradients. Limit values are employed to transform these tendencies in fuzzy values, and then compared to each other in order to extract indicators, and then evaluate these indicators by means of priorities based upon distance in fuzzy values. We formalise the method proposed above in a workflow of evaluation steps, and propose an architecture that implements the reasoning techniques.
Cristani, Matteo; Chitó, Ilaria; Tomazzoli, Claudio; Zorzi, Margherita
A Simple Algorithm for the Lexical Classification of Comparable Adjectives Proceedings Article
In: Proceedings of 22nd International Conference on Knowledge-Based and Intelligent Information & Engineering Systems (KES 2018), pp. 626–635, Elsevier, 2018.
Abstract | Links | BibTeX | Tags: Machine Learning, Natural Language Processing, Sentiment analysis
@inproceedings{cristani_simple_2018,
title = {A Simple Algorithm for the Lexical Classification of Comparable Adjectives},
author = {Matteo Cristani and Ilaria Chitó and Claudio Tomazzoli and Margherita Zorzi},
url = {https://www.sciencedirect.com/science/article/pii/S1877050918312730},
doi = {10.1016/j.procs.2018.07.297},
year = {2018},
date = {2018-01-01},
booktitle = {Proceedings of 22nd International Conference on Knowledge-Based and Intelligent Information & Engineering Systems (KES 2018)},
volume = {126},
pages = {626–635},
publisher = {Elsevier},
series = {PROCEDIA COMPUTER SCIENCE},
abstract = {Lexical classification is one of the most widely investigated fields in (computational) linguistic and Natural language Processing. Adjectives play a significant role both in classification tasks and in applications as sentiment analysis. In this paper a simple algorithm for lexical classification of comparable adjectives, called MORE (coMparable fORm dEtector), is proposed. The algorithm is efficient in time. The method is a specific unsupervised learning technique. Results are verified against a reference standard built from 80 manually annotated lists of adjective. The algorithm exhibits an accuracy of 76%},
keywords = {Machine Learning, Natural Language Processing, Sentiment analysis},
pubstate = {published},
tppubtype = {inproceedings}
}
Lexical classification is one of the most widely investigated fields in (computational) linguistic and Natural language Processing. Adjectives play a significant role both in classification tasks and in applications as sentiment analysis. In this paper a simple algorithm for lexical classification of comparable adjectives, called MORE (coMparable fORm dEtector), is proposed. The algorithm is efficient in time. The method is a specific unsupervised learning technique. Results are verified against a reference standard built from 80 manually annotated lists of adjective. The algorithm exhibits an accuracy of 76%
Cristani, Matteo; Demrozi, Florenc; Tomazzoli, Claudio
ONTO-PLC: An Ontology-Driven Methodology for Converting PLC Industrial Plants to IoT Proceedings Article
In: Proceedings of the KES Annual Conference, pp. 527–536, 2018.
Abstract | Links | BibTeX | Tags: Edge computing, Machine Learning, Semantic
@inproceedings{cristani_onto-plc_2018,
title = {ONTO-PLC: An Ontology-Driven Methodology for Converting PLC Industrial Plants to IoT},
author = {Matteo Cristani and Florenc Demrozi and Claudio Tomazzoli},
doi = {10.1016/j.procs.2018.07.287},
year = {2018},
date = {2018-01-01},
booktitle = {Proceedings of the KES Annual Conference},
volume = {126},
pages = {527–536},
series = {PROCEDIA COMPUTER SCIENCE},
abstract = {A methodology is presented that guides a user in the transition from a plant governed by PLC towards one governed by SoC.},
keywords = {Edge computing, Machine Learning, Semantic},
pubstate = {published},
tppubtype = {inproceedings}
}
A methodology is presented that guides a user in the transition from a plant governed by PLC towards one governed by SoC.
Cristani, Matteo; Olivieri, Francesco; Tomazzoli, Claudio; Zorzi, Margherita
Towards a Logical Framework for Diagnostic Reasoning Proceedings Article
In: Proceedings of 12th International Conference on Agents and Multi-Agent Systems: Technologies and Applications (KES-AMSTA-18), pp. 144–155, Springer, 2018, ISBN: 978-3-319-92030-6.
Abstract | Links | BibTeX | Tags: Automatic reasoning, Defeasible logic, Machine Learning, Non monotonic logic
@inproceedings{cristani_towards_2018,
title = {Towards a Logical Framework for Diagnostic Reasoning},
author = {Matteo Cristani and Francesco Olivieri and Claudio Tomazzoli and Margherita Zorzi},
url = {https://link.springer.com/chapter/10.1007/978-3-319-92031-3_14},
doi = {10.1007/978-3-319-92031-3},
isbn = {978-3-319-92030-6},
year = {2018},
date = {2018-01-01},
booktitle = {Proceedings of 12th International Conference on Agents and Multi-Agent Systems: Technologies and Applications (KES-AMSTA-18)},
volume = {96},
pages = {144–155},
publisher = {Springer},
abstract = {Diagnosis is widely used in many different disciplines to identify the nature and cause of a certain phenomenon. We present tL, a new logical framework able to formalise diagnostic reasoning, i.e., an hybrid learning technique based both on deduction and experiments. In this paper we introduce tL, a Labeled Modal Logic, garnishing with temporal and statistical information and a basic propositional language. After proposing examples on how tL effectively works, we sketch the main ideas about the full deduction system à la Prawitz we are currently developing.},
keywords = {Automatic reasoning, Defeasible logic, Machine Learning, Non monotonic logic},
pubstate = {published},
tppubtype = {inproceedings}
}
Diagnosis is widely used in many different disciplines to identify the nature and cause of a certain phenomenon. We present tL, a new logical framework able to formalise diagnostic reasoning, i.e., an hybrid learning technique based both on deduction and experiments. In this paper we introduce tL, a Labeled Modal Logic, garnishing with temporal and statistical information and a basic propositional language. After proposing examples on how tL effectively works, we sketch the main ideas about the full deduction system à la Prawitz we are currently developing.
2017
Galteri, Leonardo; Bazazian, Dena; Seidenari, Lorenzo; Bertini, Marco; Bagdanov, Andrew D.; Nicolau, Anguelos; Karatzas, Dimosthenis; Bimbo, Alberto Del
Reading Text in the Wild from Compressed Images Proceedings Article
In: 2017 IEEE International Conference on Computer Vision Workshops (ICCVW), pp. 2399–2407, IEEE Computer Society, Los Alamitos, CA, USA, 2017, (tex.copyright: All rights reserved).
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Compression Artifact Removal, Computer Vision and Pattern Recognition, Image Compression, Image Processing, Image Restoration, Text Recognition
@inproceedings{galteriReadingTextWild2017,
title = {Reading Text in the Wild from Compressed Images},
author = {Leonardo Galteri and Dena Bazazian and Lorenzo Seidenari and Marco Bertini and Andrew D. Bagdanov and Anguelos Nicolau and Dimosthenis Karatzas and Alberto Del Bimbo},
url = {http://openaccess.thecvf.com/content_ICCV_2017_workshops/papers/w34/degrees_leonardo.galteriunifi.it_dbazaziancvc.uab.es_lorenzo.seidenariunifi.it_ICCV_2017_paper.pdf},
doi = {10.1109/ICCVW.2017.283},
year = {2017},
date = {2017-01-01},
booktitle = {2017 IEEE International Conference on Computer Vision Workshops (ICCVW)},
pages = {2399–2407},
publisher = {IEEE Computer Society, Los Alamitos, CA, USA},
abstract = {Reading text in the wild is gaining attention in the computer vision community. Images captured in the wild are almost always compressed to varying degrees, depending on application context, and this compression introduces artifacts that distort image content into the captured images. In this paper we investigate the impact these compression artifacts have on text localization and recognition in the wild. We also propose a deep Convolutional Neural Network (CNN) that can eliminate text-specific compression artifacts and which leads to an improvement in text recognition. Experimental results on the ICDAR-Challenge4 dataset demonstrate that compression artifacts have a significant impact on text localization and recognition and that our approach yields an improvement in both–especially at high compression rates.},
note = {tex.copyright: All rights reserved},
keywords = {Artificial Intelligence, Compression Artifact Removal, Computer Vision and Pattern Recognition, Image Compression, Image Processing, Image Restoration, Text Recognition},
pubstate = {published},
tppubtype = {inproceedings}
}
Reading text in the wild is gaining attention in the computer vision community. Images captured in the wild are almost always compressed to varying degrees, depending on application context, and this compression introduces artifacts that distort image content into the captured images. In this paper we investigate the impact these compression artifacts have on text localization and recognition in the wild. We also propose a deep Convolutional Neural Network (CNN) that can eliminate text-specific compression artifacts and which leads to an improvement in text recognition. Experimental results on the ICDAR-Challenge4 dataset demonstrate that compression artifacts have a significant impact on text localization and recognition and that our approach yields an improvement in both–especially at high compression rates.
Galteri, Leonardo; Seidenari, Lorenzo; Bertini, Marco; Bimbo, Alberto Del
Deep generative adversarial compression artifact removal Proceedings Article
In: Proceedings of the IEEE international conference on computer vision, pp. 4826–4835, 2017, (tex.copyright: All rights reserved).
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Compression Artifact Removal, Computer Vision and Pattern Recognition, Generative Adversarial Networks, Image Compression, Image Processing, Image Synthesis and Enhancement
@inproceedings{galteriDeepGenerativeAdversarial2017,
title = {Deep generative adversarial compression artifact removal},
author = {Leonardo Galteri and Lorenzo Seidenari and Marco Bertini and Alberto Del Bimbo},
url = {https://openaccess.thecvf.com/content_ICCV_2017/papers/Galteri_Deep_Generative_Adversarial_ICCV_2017_paper.pdf},
doi = {10.1109/ICCV.2017.517},
year = {2017},
date = {2017-01-01},
booktitle = {Proceedings of the IEEE international conference on computer vision},
pages = {4826–4835},
abstract = {Compression artifacts arise in images whenever a lossy compression algorithm is applied. These artifacts eliminate details present in the original image, or add noise and small structures; because of these effects they make images less pleasant for the human eye, and may also lead to decreased performance of computer vision algorithms such as object detectors. To eliminate such artifacts, when decompressing an image, it is required to recover the original image from a disturbed version. To this end, we present a feed-forward fully convolutional residual network model trained using a generative adversarial framework. To provide a baseline, we show that our model can be also trained optimizing the Structural Similarity (SSIM), which is a better loss with respect to the simpler Mean Squared Error (MSE). Our GAN is able to produce images with more photorealistic details than MSE or SSIM based networks. Moreover we show that our approach can be used as a pre-processing step for object detection in case images are degraded by compression to a point that state-of-the art detectors fail. In this task, our GAN method obtains better performance than MSE or SSIM trained networks.},
note = {tex.copyright: All rights reserved},
keywords = {Artificial Intelligence, Compression Artifact Removal, Computer Vision and Pattern Recognition, Generative Adversarial Networks, Image Compression, Image Processing, Image Synthesis and Enhancement},
pubstate = {published},
tppubtype = {inproceedings}
}
Compression artifacts arise in images whenever a lossy compression algorithm is applied. These artifacts eliminate details present in the original image, or add noise and small structures; because of these effects they make images less pleasant for the human eye, and may also lead to decreased performance of computer vision algorithms such as object detectors. To eliminate such artifacts, when decompressing an image, it is required to recover the original image from a disturbed version. To this end, we present a feed-forward fully convolutional residual network model trained using a generative adversarial framework. To provide a baseline, we show that our model can be also trained optimizing the Structural Similarity (SSIM), which is a better loss with respect to the simpler Mean Squared Error (MSE). Our GAN is able to produce images with more photorealistic details than MSE or SSIM based networks. Moreover we show that our approach can be used as a pre-processing step for object detection in case images are degraded by compression to a point that state-of-the art detectors fail. In this task, our GAN method obtains better performance than MSE or SSIM trained networks.
Tomazzoli, Claudio; Storti, Silvia Francesca; Galazzo, Ilaria Boscolo; Cristani, Matteo; Menegaz, Gloria
The Brain Is a Social Network Proceedings Article
In: CEUR Workshop Proceedings, pp. 1–6, 2017.
Abstract | Links | BibTeX | Tags: brain connectomics, Machine Learning, semantic social network analysis
@inproceedings{tomazzoli_brain_2017,
title = {The Brain Is a Social Network},
author = {Claudio Tomazzoli and Silvia Francesca Storti and Ilaria Boscolo Galazzo and Matteo Cristani and Gloria Menegaz},
url = {https://ceur-ws.org/Vol-1959/paper-10.pdf},
year = {2017},
date = {2017-01-01},
booktitle = {CEUR Workshop Proceedings},
pages = {1–6},
abstract = {Social Network Analysis is employed widely as a means to compute the probability that a given message flows through a social net- work. This approach is mainly grounded upon the correct usage of three basic graph-theoretic measures: degree centrality, closeness centrality and betweeness centrality. We developed a model, using Semantic Social Net- work Analysis, that overcomes the drawbacks of general indices and we found that this model can be applied, after appropriate adaptations, to a very different domain such as brain connectivity.},
keywords = {brain connectomics, Machine Learning, semantic social network analysis},
pubstate = {published},
tppubtype = {inproceedings}
}
Social Network Analysis is employed widely as a means to compute the probability that a given message flows through a social net- work. This approach is mainly grounded upon the correct usage of three basic graph-theoretic measures: degree centrality, closeness centrality and betweeness centrality. We developed a model, using Semantic Social Net- work Analysis, that overcomes the drawbacks of general indices and we found that this model can be applied, after appropriate adaptations, to a very different domain such as brain connectivity.
Tomazzoli, Claudio; Scannapieco, Simone
Machine Learning for Energy Efficiency - Automatic Detection of Electric Loads from Power Consumption Proceedings Article
In: IEEE Xplore, pp. 1–6, 2017.
Abstract | Links | BibTeX | Tags: Edge computing, Energy management, Machine Learning
@inproceedings{tomazzoli_machine_2017,
title = {Machine Learning for Energy Efficiency - Automatic Detection of Electric Loads from Power Consumption},
author = {Claudio Tomazzoli and Simone Scannapieco},
doi = {10.23919/AEIT.2017.8240544},
year = {2017},
date = {2017-01-01},
booktitle = {IEEE Xplore},
pages = {1–6},
abstract = {This work deals with the problem of energy efficiency and saving: we present a method to automatically extract behavioral rules from consumption data, so that these rules can be applied or fed to an automatic control system. To extract behavioral rules we shall be able to both (i) define power plants similarity techniques and (ii) analyze and gather rules from data, making the correct assumptions.},
keywords = {Edge computing, Energy management, Machine Learning},
pubstate = {published},
tppubtype = {inproceedings}
}
This work deals with the problem of energy efficiency and saving: we present a method to automatically extract behavioral rules from consumption data, so that these rules can be applied or fed to an automatic control system. To extract behavioral rules we shall be able to both (i) define power plants similarity techniques and (ii) analyze and gather rules from data, making the correct assumptions.
Scannapieco, Simone; Tomazzoli, Claudio
Ubiquitous and Pervasive Computing for Real-Time Energy Management and Saving Proceedings Article
In: Advances in Intelligent Systems and Computing, pp. 3–15, 2017.
Abstract | Links | BibTeX | Tags: Edge computing, Energy Efficiency, Energy management, Machine Learning
@inproceedings{scannapieco_ubiquitous_2017,
title = {Ubiquitous and Pervasive Computing for Real-Time Energy Management and Saving},
author = {Simone Scannapieco and Claudio Tomazzoli},
doi = {10.1007/978-3-319-61542-4_1},
year = {2017},
date = {2017-01-01},
booktitle = {Advances in Intelligent Systems and Computing},
volume = {612},
pages = {3–15},
abstract = {In the present paper we investigate one of the emerging applicability fields of pervasive computing, that is, energy management and saving. We exploit innovative technologies to define a brand new system architecture for (i) centralized monitoring and (ii) real-time energy saving in distributed sub-networks of power consuming electric appliances. The architecture allows the definition of reactive and intelligent systems that take autonomous courses of action on electric devices. We also introduce a prototype which is an embodiment of such architecture, called .MyElettra},
keywords = {Edge computing, Energy Efficiency, Energy management, Machine Learning},
pubstate = {published},
tppubtype = {inproceedings}
}
In the present paper we investigate one of the emerging applicability fields of pervasive computing, that is, energy management and saving. We exploit innovative technologies to define a brand new system architecture for (i) centralized monitoring and (ii) real-time energy saving in distributed sub-networks of power consuming electric appliances. The architecture allows the definition of reactive and intelligent systems that take autonomous courses of action on electric devices. We also introduce a prototype which is an embodiment of such architecture, called .MyElettra
Scannapieco, Simone; Tomazzoli, Claudio
Shoo the Spectre of Ignorance with QAASPR - an Open Domain Question Answering Architecture with Semantic Prioritisation of Roles Proceedings Article
In: CEUR Workshop Proceedings, pp. 1–6, 2017.
Links | BibTeX | Tags: Machine Learning, Natural Language Processing, Semantic
@inproceedings{scannapieco_shoo_2017,
title = {Shoo the Spectre of Ignorance with QAASPR - an Open Domain Question Answering Architecture with Semantic Prioritisation of Roles},
author = {Simone Scannapieco and Claudio Tomazzoli},
url = {https://ceur-ws.org/Vol-1959/paper-05.pdf},
year = {2017},
date = {2017-01-01},
booktitle = {CEUR Workshop Proceedings},
pages = {1–6},
keywords = {Machine Learning, Natural Language Processing, Semantic},
pubstate = {published},
tppubtype = {inproceedings}
}
2016
Bonci, Andrea; Pirani, Massimiliano; Longhi, Sauro
Embedded solutions for a class of highly unstable, underactuated and self-balancing robotic systems Journal Article
In: EURASIP Journal on Embedded Systems, vol. 2017, no. 1, pp. 9, 2016, ISSN: 1687-3963.
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Collaborative Robotics, Cyber-Physical Systems, Microcontrollers
@article{bonciEmbeddedSolutionsClass2016,
title = {Embedded solutions for a class of highly unstable, underactuated and self-balancing robotic systems},
author = {Andrea Bonci and Massimiliano Pirani and Sauro Longhi},
url = {https://doi.org/10.1186/s13639-016-0046-6},
doi = {10.1186/s13639-016-0046-6},
issn = {1687-3963},
year = {2016},
date = {2016-08-01},
urldate = {2024-10-09},
journal = {EURASIP Journal on Embedded Systems},
volume = {2017},
number = {1},
pages = {9},
abstract = {This paper presents a didactic framework in embedded electronics systems that is used to elicit awareness into students and engineers on the design issues arising in the realization of a class of underactuated robots and aerial vehicles that needs be robustly controlled due to their intrinsic unstability. The applications prototyped on the embedded platform presented here are conceived, by design, to be compliant with tiny collaborative robotics applications in order to adhere to the needs of the complex cyber-physical systems problem. The proposed platform is self-contained with on-board sensing and computation. Its engineering uses only off-the-shelf and mass production components. The system is based on a general purpose embedded board equipped with a 32-bit microcontroller which is able to manage all the basic tasks of this robotic platform: sensing, actuation, control and communication. The framework is described, and initial experimental results are introduced. Three applications are presented in this work as a validation of the methodology: a ballbot robot, a legged robot and a quadrotor aerial vehicle. The chosen case studies are robotics applications that are specialized in performing maneuvers when operating in tight spaces as in the human living environments.},
keywords = {Artificial Intelligence, Collaborative Robotics, Cyber-Physical Systems, Microcontrollers},
pubstate = {published},
tppubtype = {article}
}
This paper presents a didactic framework in embedded electronics systems that is used to elicit awareness into students and engineers on the design issues arising in the realization of a class of underactuated robots and aerial vehicles that needs be robustly controlled due to their intrinsic unstability. The applications prototyped on the embedded platform presented here are conceived, by design, to be compliant with tiny collaborative robotics applications in order to adhere to the needs of the complex cyber-physical systems problem. The proposed platform is self-contained with on-board sensing and computation. Its engineering uses only off-the-shelf and mass production components. The system is based on a general purpose embedded board equipped with a 32-bit microcontroller which is able to manage all the basic tasks of this robotic platform: sensing, actuation, control and communication. The framework is described, and initial experimental results are introduced. Three applications are presented in this work as a validation of the methodology: a ballbot robot, a legged robot and a quadrotor aerial vehicle. The chosen case studies are robotics applications that are specialized in performing maneuvers when operating in tight spaces as in the human living environments.
Cristani, Matteo; Tomazzoli, Claudio
A Multimodal Approach to Relevance and Pertinence of Documents Proceedings Article
In: Trends in Applied Knowledge-Based Systems and Data Science, pp. 157–168, 2016.
Links | BibTeX | Tags: Classification, Computational Methods, Document automation, Machine Learning
@inproceedings{cristani_multimodal_2016,
title = {A Multimodal Approach to Relevance and Pertinence of Documents},
author = {Matteo Cristani and Claudio Tomazzoli},
doi = {10.1007/978-3-319-42007-3_14},
year = {2016},
date = {2016-01-01},
booktitle = {Trends in Applied Knowledge-Based Systems and Data Science},
volume = {9799},
pages = {157–168},
keywords = {Classification, Computational Methods, Document automation, Machine Learning},
pubstate = {published},
tppubtype = {inproceedings}
}
Tomazzoli, Claudio; Cristani, Matteo; Olivieri, Francesco
Automatic Synthesis of Best Practices for Energy Consumptions Proceedings Article
In: Proceedings of the Tenth International Conference on Innovative Mobile and Internet Services in Ubiquitous Computing, pp. 1–8, IEEE CPS, 2016.
Abstract | Links | BibTeX | Tags: Defeasible logic, Energy savings, Machine Learning
@inproceedings{tomazzoli_automatic_2016,
title = {Automatic Synthesis of Best Practices for Energy Consumptions},
author = {Claudio Tomazzoli and Matteo Cristani and Francesco Olivieri},
doi = {10.1109/IMIS.2016.79},
year = {2016},
date = {2016-01-01},
booktitle = {Proceedings of the Tenth International Conference on Innovative Mobile and Internet Services in Ubiquitous Computing},
pages = {1–8},
publisher = {IEEE CPS},
abstract = {Conflicting rules and rules with exceptions are very common in natural language specification to describe the behaviour of devices operating in a real-world context. This is common exactly because those specifications are processed by humans, and humans apply common sense and strategic reasoning about those rules. In this paper, we deal with the challenge of providing, step by step, a model of energy saving rule specification and processing methods that are used to reduce the consumptions of a system of devices. We argue that a very promising non-monotonic approach to such a problem can lie upon Defeasible Logic. Starting with rules specified at an abstract level, but compatibly with the natural aspects of such a specification (including temporal and power absorption constraints), we provide a formalism that generates the extension of a basic defeasible logic, which corresponds to turned on or off devices.We define a procedure to achieve automatic synthesis of best practices, to be used as rules to obtain savings in electric consumptions.},
keywords = {Defeasible logic, Energy savings, Machine Learning},
pubstate = {published},
tppubtype = {inproceedings}
}
Conflicting rules and rules with exceptions are very common in natural language specification to describe the behaviour of devices operating in a real-world context. This is common exactly because those specifications are processed by humans, and humans apply common sense and strategic reasoning about those rules. In this paper, we deal with the challenge of providing, step by step, a model of energy saving rule specification and processing methods that are used to reduce the consumptions of a system of devices. We argue that a very promising non-monotonic approach to such a problem can lie upon Defeasible Logic. Starting with rules specified at an abstract level, but compatibly with the natural aspects of such a specification (including temporal and power absorption constraints), we provide a formalism that generates the extension of a basic defeasible logic, which corresponds to turned on or off devices.We define a procedure to achieve automatic synthesis of best practices, to be used as rules to obtain savings in electric consumptions.
Cristani, Matteo; Tomazzoli, Claudio; Olivieri, Francesco
Semantic Social Network Analysis Foresees Message Flows Proceedings Article
In: Proceedings of the 8th International Conference on Agents and Artificial Intelligence, pp. 296–303, 2016.
Abstract | Links | BibTeX | Tags: Closeness centralities, Machine Learning, semantic social network analysis, Social Network
@inproceedings{cristani_semantic_2016,
title = {Semantic Social Network Analysis Foresees Message Flows},
author = {Matteo Cristani and Claudio Tomazzoli and Francesco Olivieri},
doi = {10.5220/0005832902960303},
year = {2016},
date = {2016-01-01},
booktitle = {Proceedings of the 8th International Conference on Agents and Artificial Intelligence},
volume = {1},
pages = {296–303},
abstract = {Social Network Analysis is employed widely as a means to compute the probability that a given message flows through a social network. This approach is mainly grounded upon the correct usage of three basic graph-theoretic measures: Degree centrality, Closeness centrality and Betweenness centrality. We show that, in general, those indices are not adapt to foresee the flow of a given message, which depends upon indices based on the sharing of interests and the trust about depth in knowledge of a topic. We provide an extended model, that is a simplified version of a more general model already documented in the literature, the Semantic Social Network Analysis, and we show that by means of this model it is possible to exceed the drawbacks of general indices discussed above.},
keywords = {Closeness centralities, Machine Learning, semantic social network analysis, Social Network},
pubstate = {published},
tppubtype = {inproceedings}
}
Social Network Analysis is employed widely as a means to compute the probability that a given message flows through a social network. This approach is mainly grounded upon the correct usage of three basic graph-theoretic measures: Degree centrality, Closeness centrality and Betweenness centrality. We show that, in general, those indices are not adapt to foresee the flow of a given message, which depends upon indices based on the sharing of interests and the trust about depth in knowledge of a topic. We provide an extended model, that is a simplified version of a more general model already documented in the literature, the Semantic Social Network Analysis, and we show that by means of this model it is possible to exceed the drawbacks of general indices discussed above.
Tomazzoli, Claudio; Cristani, Matteo; Fogoroasi, Diana
Measuring Homophily Proceedings Article
In: CEUR Workshop Proceedings, pp. 1–12, 2016.
Abstract | Links | BibTeX | Tags: Machine Learning, semantic social network analysis, Social Network
@inproceedings{tomazzoli_measuring_2016,
title = {Measuring Homophily},
author = {Claudio Tomazzoli and Matteo Cristani and Diana Fogoroasi},
url = {https://ceur-ws.org/Vol-1748/paper-09.pdf},
year = {2016},
date = {2016-01-01},
booktitle = {CEUR Workshop Proceedings},
pages = {1–12},
abstract = {Social Network Analysis is employed widely as a means to compute the probability that a given message flows through a social network. This approach is mainly grounded upon the correct usage of three basic graphtheoretic measures: degree centrality, closeness centrality and betweeness centrality. We show that, in general, those indices are not adapt to foresee the flow of a given message, that depends upon indices based on the sharing of interests and the trust about depth in knowledge of a topic. We provide new definitions for measures that overcome the drawbacks of general indices discussed above, using Semantic Social Network Analysis, and show experimental results that show that with these measures we have a different understanding of a social network compared to standard measures.},
keywords = {Machine Learning, semantic social network analysis, Social Network},
pubstate = {published},
tppubtype = {inproceedings}
}
Social Network Analysis is employed widely as a means to compute the probability that a given message flows through a social network. This approach is mainly grounded upon the correct usage of three basic graphtheoretic measures: degree centrality, closeness centrality and betweeness centrality. We show that, in general, those indices are not adapt to foresee the flow of a given message, that depends upon indices based on the sharing of interests and the trust about depth in knowledge of a topic. We provide new definitions for measures that overcome the drawbacks of general indices discussed above, using Semantic Social Network Analysis, and show experimental results that show that with these measures we have a different understanding of a social network compared to standard measures.
2015
Cristani, Matteo; Karafili, Erisa; Tomazzoli, Claudio
Improving Energy Saving Techniques by Ambient Intelligence Scheduling Proceedings Article
In: Proceedings of the 2015 IEEE 29th International Conference on Advanced Information Networking and Applications (AINA 2015), pp. 324–331, Conference Publishing Services (CPS) – IEEE Computer Society, Los Alamitos, California, 2015.
Abstract | Links | BibTeX | Tags: Energy Efficiency, Machine Learning, Scheduling
@inproceedings{cristani_improving_2015,
title = {Improving Energy Saving Techniques by Ambient Intelligence Scheduling},
author = {Matteo Cristani and Erisa Karafili and Claudio Tomazzoli},
doi = {10.1109/AINA.2015.202},
year = {2015},
date = {2015-01-01},
booktitle = {Proceedings of the 2015 IEEE 29th International Conference on Advanced Information Networking and Applications (AINA 2015)},
volume = {1},
pages = {324–331},
publisher = {Conference Publishing Services (CPS) – IEEE Computer Society},
address = {Los Alamitos, California},
abstract = {Energy saving is one of the most challenging aspects of modern ambient intelligence technologies, for both domestic and business usages. In this paper we show how to combine Ambient Intelligence and Artificial Intelligence techniques to solve the problem of scheduling a set of devices under a given set of constraints, like limits to the maximal energy usage (Energy Span) and maximal energy absorption (Energy Peak). We provide a method that can be used to schedule the usage of devices in a given environment in a way that respects the input constraints. We adapt an existent approach to scheduling for Ambient Intelligence to aspecific framework and exhibit a sample usage for a real life system, Elettra, that is in use in an industrial context.},
keywords = {Energy Efficiency, Machine Learning, Scheduling},
pubstate = {published},
tppubtype = {inproceedings}
}
Energy saving is one of the most challenging aspects of modern ambient intelligence technologies, for both domestic and business usages. In this paper we show how to combine Ambient Intelligence and Artificial Intelligence techniques to solve the problem of scheduling a set of devices under a given set of constraints, like limits to the maximal energy usage (Energy Span) and maximal energy absorption (Energy Peak). We provide a method that can be used to schedule the usage of devices in a given environment in a way that respects the input constraints. We adapt an existent approach to scheduling for Ambient Intelligence to aspecific framework and exhibit a sample usage for a real life system, Elettra, that is in use in an industrial context.
2014
Cristani, Matteo; Tomazzoli, Claudio
A Multimodal Approach to Exploit Similarity in Documents Proceedings Article
In: IEA-AIE-2014proceedings, pp. 490–499, Springer, 2014.
Abstract | Links | BibTeX | Tags: Classification, Document automation, Machine Learning, Natural Language Processing, Taxonomy
@inproceedings{cristani_multimodal_2014,
title = {A Multimodal Approach to Exploit Similarity in Documents},
author = {Matteo Cristani and Claudio Tomazzoli},
doi = {10.1007/978-3-319-07455-9_51},
year = {2014},
date = {2014-01-01},
booktitle = {IEA-AIE-2014proceedings},
pages = {490–499},
publisher = {Springer},
abstract = {Automated document classification process extracts information with a systematic analysis of the content of documents. This is an active research field of growing importance due to the large amount of electronic documents produced in the world wide web and available thanks to diffused technologies including mobile ones. Several application areas benefit from automated document classification, including document archiving, invoice processing in business environments, press releases and research engines. Current tools classify or ”tag” either text or images separately.In this paper we show how, by linking image and text-based contents together, a technology improves fundamental document management tasks like retrieving information from a database or automated documents. We present an investigation of a model of conceptual spaces for investigation using joint information sources from the text and the images forming complex documents. We present a formal model and the computable algorithms and the dataset from which we took a subset to make experiments and relative tests and results.},
keywords = {Classification, Document automation, Machine Learning, Natural Language Processing, Taxonomy},
pubstate = {published},
tppubtype = {inproceedings}
}
Automated document classification process extracts information with a systematic analysis of the content of documents. This is an active research field of growing importance due to the large amount of electronic documents produced in the world wide web and available thanks to diffused technologies including mobile ones. Several application areas benefit from automated document classification, including document archiving, invoice processing in business environments, press releases and research engines. Current tools classify or ”tag” either text or images separately.In this paper we show how, by linking image and text-based contents together, a technology improves fundamental document management tasks like retrieving information from a database or automated documents. We present an investigation of a model of conceptual spaces for investigation using joint information sources from the text and the images forming complex documents. We present a formal model and the computable algorithms and the dataset from which we took a subset to make experiments and relative tests and results.
Cristani, Matteo; Karafili, Erisa; Tomazzoli, Claudio
Energy Saving by Ambient Intelligence Techniques Proceedings Article
In: Network-Based Information Systems (NBiS), 2014 17th International Conference On, pp. 157–164, IEEE, 2014.
Abstract | Links | BibTeX | Tags: Energy Efficiency, Intelligent Systems, Machine Learning
@inproceedings{cristani_energy_2014,
title = {Energy Saving by Ambient Intelligence Techniques},
author = {Matteo Cristani and Erisa Karafili and Claudio Tomazzoli},
doi = {10.1109/NBiS.2014.39},
year = {2014},
date = {2014-01-01},
booktitle = {Network-Based Information Systems (NBiS), 2014 17th International Conference On},
pages = {157–164},
publisher = {IEEE},
abstract = {Nowadays the problem of energy consumption is becoming a pressing problem. We present an innovative system named Elettra able to allow people to monitor and control energy consumption in one or more buildings. For improving Elettra we introduce different methods taken from ambient intelligence. Through these methods we can infer energy consumption, construct a plan for decreasing energy consumption, improve this plan and adopt it to the system. The implementation of these methods to Elettra helps its automation and increases its efficiency.},
keywords = {Energy Efficiency, Intelligent Systems, Machine Learning},
pubstate = {published},
tppubtype = {inproceedings}
}
Nowadays the problem of energy consumption is becoming a pressing problem. We present an innovative system named Elettra able to allow people to monitor and control energy consumption in one or more buildings. For improving Elettra we introduce different methods taken from ambient intelligence. Through these methods we can infer energy consumption, construct a plan for decreasing energy consumption, improve this plan and adopt it to the system. The implementation of these methods to Elettra helps its automation and increases its efficiency.
2008
Turchetti, Claudio; Crippa, Paolo; Pirani, Massimiliano; Biagetti, Giorgio
Representation of Nonlinear Random Transformations by Non-Gaussian Stochastic Neural Networks Journal Article
In: IEEE Transactions on Neural Networks, vol. 19, no. 6, pp. 1033–1060, 2008, ISSN: 1941-0093.
Abstract | Links | BibTeX | Tags: Approximation, Artificial Intelligence, Biology Computing, Computer Networks, Lee-Schetzen Method, Neural Computation, Neural networks, Nonlinear systems, Signal Processing, Stochastic Processes, Stochastic Resonance, Stochastic Systems
@article{turchettiRepresentationNonlinearRandom2008,
title = {Representation of Nonlinear Random Transformations by Non-Gaussian Stochastic Neural Networks},
author = {Claudio Turchetti and Paolo Crippa and Massimiliano Pirani and Giorgio Biagetti},
url = {https://ieeexplore.ieee.org/document/4460850},
doi = {10.1109/TNN.2007.2000055},
issn = {1941-0093},
year = {2008},
date = {2008-06-01},
urldate = {2024-10-09},
journal = {IEEE Transactions on Neural Networks},
volume = {19},
number = {6},
pages = {1033–1060},
abstract = {The learning capability of neural networks is equivalent to modeling physical events that occur in the real environment. Several early works have demonstrated that neural networks belonging to some classes are universal approximators of input-output deterministic functions. Recent works extend the ability of neural networks in approximating random functions using a class of networks named stochastic neural networks (SNN). In the language of system theory, the approximation of both deterministic and stochastic functions falls within the identification of nonlinear no-memory systems. However, all the results presented so far are restricted to the case of Gaussian stochastic processes (SPs) only, or to linear transformations that guarantee this property. This paper aims at investigating the ability of stochastic neural networks to approximate nonlinear input-output random transformations, thus widening the range of applicability of these networks to nonlinear systems with memory. In particular, this study shows that networks belonging to a class named non-Gaussian stochastic approximate identity neural networks (SAINNs) are capable of approximating the solutions of large classes of nonlinear random ordinary differential transformations. The effectiveness of this approach is demonstrated and discussed by some application examples.},
keywords = {Approximation, Artificial Intelligence, Biology Computing, Computer Networks, Lee-Schetzen Method, Neural Computation, Neural networks, Nonlinear systems, Signal Processing, Stochastic Processes, Stochastic Resonance, Stochastic Systems},
pubstate = {published},
tppubtype = {article}
}
The learning capability of neural networks is equivalent to modeling physical events that occur in the real environment. Several early works have demonstrated that neural networks belonging to some classes are universal approximators of input-output deterministic functions. Recent works extend the ability of neural networks in approximating random functions using a class of networks named stochastic neural networks (SNN). In the language of system theory, the approximation of both deterministic and stochastic functions falls within the identification of nonlinear no-memory systems. However, all the results presented so far are restricted to the case of Gaussian stochastic processes (SPs) only, or to linear transformations that guarantee this property. This paper aims at investigating the ability of stochastic neural networks to approximate nonlinear input-output random transformations, thus widening the range of applicability of these networks to nonlinear systems with memory. In particular, this study shows that networks belonging to a class named non-Gaussian stochastic approximate identity neural networks (SAINNs) are capable of approximating the solutions of large classes of nonlinear random ordinary differential transformations. The effectiveness of this approach is demonstrated and discussed by some application examples.
2005
Orcioni, Simone; Pirani, Massimiliano; Turchetti, Claudio
Advances in Lee–Schetzen Method for Volterra Filter Identification Journal Article
In: Multidimensional Systems and Signal Processing, vol. 16, no. 3, pp. 265–284, 2005, ISSN: 1573-0824.
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Lee-Schetzen Method, Nonlinear System Identification, Volterra Filters, Wiener Kernels
@article{orcioniAdvancesLeeSchetzen2005,
title = {Advances in Lee–Schetzen Method for Volterra Filter Identification},
author = {Simone Orcioni and Massimiliano Pirani and Claudio Turchetti},
url = {https://doi.org/10.1007/s11045-004-1677-7},
doi = {10.1007/s11045-004-1677-7},
issn = {1573-0824},
year = {2005},
date = {2005-07-01},
urldate = {2024-10-09},
journal = {Multidimensional Systems and Signal Processing},
volume = {16},
number = {3},
pages = {265–284},
abstract = {This paper concerns the identification of nonlinear discrete causal systems that can be approximated with the Wiener–Volterra series. Some advances in the efficient use of Lee–Schetzen (L–S) method are presented, which make practical the estimate of long memory and high order models. Major problems in L–S method occur in the identification of diagonal kernel elements. Two approaches have been considered: approximation of gridded data, with interpolation or smoothing, and improved techniques for diagonal elements estimation. A comparison of diagonal elements estimated, with different methods has been shown with extended tests on fifth order Volterra systems.},
keywords = {Artificial Intelligence, Lee-Schetzen Method, Nonlinear System Identification, Volterra Filters, Wiener Kernels},
pubstate = {published},
tppubtype = {article}
}
This paper concerns the identification of nonlinear discrete causal systems that can be approximated with the Wiener–Volterra series. Some advances in the efficient use of Lee–Schetzen (L–S) method are presented, which make practical the estimate of long memory and high order models. Major problems in L–S method occur in the identification of diagonal kernel elements. Two approaches have been considered: approximation of gridded data, with interpolation or smoothing, and improved techniques for diagonal elements estimation. A comparison of diagonal elements estimated, with different methods has been shown with extended tests on fifth order Volterra systems.