A5064873591- Neural Networks and Applications
- Advanced Graph Neural Networks
- Face and Expression Recognition
- Computational Drug Discovery Methods
- Graph Theory and Algorithms
- Business Process Modeling and Analysis
- Fuzzy Logic and Control Systems
- Machine Learning in Materials Science
- Topic Modeling
- Text and Document Classification Technologies
- Natural Language Processing Techniques
- Service-Oriented Architecture and Web Services
- Machine Learning and Algorithms
- Domain Adaptation and Few-Shot Learning
- Complex Network Analysis Techniques
- Music and Audio Processing
- Machine Learning and Data Classification
- Data Stream Mining Techniques
- Machine Learning and ELM
- Neural Networks and Reservoir Computing
- Time Series Analysis and Forecasting
- Multimodal Machine Learning Applications
- Image Retrieval and Classification Techniques
- Bioinformatics and Genomic Networks
- Advanced Image and Video Retrieval Techniques
University of Trento
2022-2025
Fondazione Bruno Kessler
2020-2025
University of Padua
2015-2024
Civita
2020-2023
Conference Board
2021-2022
University of Pisa
1998-2020
Eindhoven University of Technology
2019
Clinical Research Consortium
2019
RWTH Aachen University
2019
Universitat Politècnica de Catalunya
2019
Standard neural networks and statistical methods are usually believed to be inadequate when dealing with complex structures because of their feature-based approach. In fact, approaches fail give satisfactory solutions the sensitivity approach a priori selection features, incapacity represent any specific information on relationships among components structures. However, we show that can, in classify structured patterns. The key idea underpinning our is use so called "generalized recursive...
A structured organization of information is typically required by symbolic processing. On the other hand, most connectionist models assume that data are organized according to relatively poor structures, like arrays or sequences. The framework described in this paper an attempt unify adaptive artificial neural nets and belief for problem processing information. In particular, relations between variables expressed directed acyclic graphs, where both numerical categorical values coexist....
Abstract On the 21 st of February 2020 a resident municipality Vo’, small town near Padua, died pneumonia due to SARS-CoV-2 infection 1 . This was first COVID-19 death detected in Italy since emergence Chinese city Wuhan, Hubei province 2 In response, regional authorities imposed lockdown whole for 14 days 3 We collected information on demography, clinical presentation, hospitalization, contact network and presence nasopharyngeal swabs 85.9% 71.5% population Vo’ at two consecutive time...
Recent developments in the area of neural networks produced models capable dealing with structured data. Here, we propose first fully unsupervised model, namely an extension traditional self-organizing maps (SOMs), for processing labeled directed acyclic graphs (DAGs). The is obtained by using unfolding procedure adopted recurrent and recursive networks, replicated neurons unfolded network comprising a full SOM. This approach enables discovery similarities among objects including vectors...
The number of malware applications targeting the Android operating system has significantly increased in recent years. Malicious pose a significant threat to platform security. We propose ANASTASIA, detect malicious through statically analyzing applications' behaviors. ANASTASIA provides more complete coverage security behaviors when compared state-of-the-art solutions. utilize large extracted features from various behavioral characteristics an application. built Machine Learning-based...
Predicting the completion time of business process instances would be a very helpful aid when managing processes under service level agreement constraints. The ability to know in advance trend running allow managers react time, order prevent delays or undesirable situations. However, making such accurate forecasts is not easy: many factors may influence required complete instance. In this paper, we propose an approach based on deep Recurrent Neural Networks (specifically LSTMs) that able...
Accurate prediction of the completion time a business process instance would constitute valuable tool when managing processes under service level agreement constraints. Such prediction, however, is very challenging task. A wide variety factors could influence trend instance, and hence just using statistics historical cases cannot be sufficient to get accurate predictions. Here we propose new approach where, in order improve quality, both control data flow perspectives are jointly used. To...
Today's business processes are often controlled and supported by information systems. These systems record real-time about during their executions. This enables the analysis at runtime of process behavior. However, many modern produce "big data", i.e., collections data sets so large complex that it becomes impossible to store all them. Moreover, few in steady-state but, due changing circumstances, they evolve need adapt continuously. In this paper, we present a novel framework for discovery...
Process Mining represents an important research field that connects Business Modeling and Data Mining. One of the most prominent task is discovery a control-flow starting from event logs. This paper focuses on problem stream data. We propose to adapt Heuristics Miner, one effective algorithms, treatment streams Two adaptations, based Lossy Counting with Budget, as well sliding window version are proposed experimentally compared against both artificial real streams. Experimental results show...
In open set recognition, a classifier has to detect unknown classes that are not known at training time. order recognize new categories, the project input samples of in very compact and separated regions features space for discriminating classes. Recently proposed Capsule Networks have shown outperform alternatives many fields, particularly image however they been fully applied yet open-set recognition. capsule networks, scalar neurons replaced by vectors or matrices, whose entries represent...
Recently, many researchers have been focusing on the definition of neural networks for graphs. The basic component these approaches remains graph convolution idea proposed almost a decade ago. In this paper, we extend component, following an intuition derived from well-known convolutional filters over multi-dimensional tensors. particular, derive simple, efficient and effective way to introduce hyper-parameter convolutions that influences filter size, i.e., its receptive field considered...
We explore the application of machine learning (ML) techniques to forecast door-to-door waste collection, addressing challenges in municipal solid (MSW) management. ML models offer a promising solution optimize collection operations, especially amid growing urban populations and evolving generation rates. Leveraging comprehensive data from northeastern Italian municipality, including various types, our study investigates algorithms' efficacy predicting household requirements. examine two key...