Convolutional Neural Networks can be effectively used only when data are endowed with an intrinsic concept of neighbourhood in the input space, as is case pixels images. We introduce here Ph-CNN, a novel deep learning architecture for classification metagenomics based on Networks, patristic distance defined phylogenetic tree being proximity measure. The between variables together sparsified version MultiDimensional Scaling to embed Euclidean space. Ph-CNN tested domain adaptation approach...

Developers have a lot of freedom in writing comments as well choosing identifiers and method names. These are intentional nature provide different relevance information to understand what software system implements, particular the role each source file. In this paper we investigate effectiveness exploiting lexical for clustering. explore contribution combined use six dictionaries, corresponding parts code where programmers introduce information, namely: class, attribute, parameter names,...

Artificial Intelligence is exponentially increasing its impact on healthcare. As deep learning mastering computer vision tasks, application to digital pathology natural, with the promise of aiding in routine reporting and standardizing results across trials. Deep features inferred from scans can improve validity robustness current clinico-pathological features, up identifying novel histological patterns, e.g., tumor infiltrating lymphocytes. In this study, we examine issue evaluating...

Information Retrieval (IR) techniques are being exploited by an increasing number of tools supporting Software Maintenance activities. Indeed the lexical information embedded in source code can be valuable for tasks such as concept location, clustering or recovery traceability links. The application IR-based relies on consistency lexicon available different artifacts, and their effectiveness worsen if programmers introduce abbreviations (e.g: rect) and/or do not strictly follow naming...

Insect pests are often associated with food contamination and public health risks. Accurate timely species-specific identification of is a key step to scale impacts, trace back the process promptly set intervention measures, which usually have serious economic impact. The current procedure involves visual inspection by human analysts pest fragments recovered from samples, time-consuming error-prone process. Deep Learning models been widely applied for image recognition, outperforming other...

Almost every clinical specialty will use artificial intelligence in the future. The first area of practical impact is expected to be rapid and accurate interpretation image streams such as radiology scans, histo-pathology slides, ophthalmic imaging, any other bioimaging diagnostic systems, enriched by phenotypes used outcome labels or additional descriptors. In this study, we introduce a machine learning framework for automatic that combines current pattern recognition approach ("radiomics")...

Abstract We introduce , a high-resolution radar reflectivity dataset collected by the Civil Protection weather of Trentino South Tyrol Region, in Italian Alps. The includes 894,916 timesteps precipitation from more than 9 years data, offering novel resource to develop and benchmark analog ensemble models machine learning solutions for nowcasting. Data are expressed as 2D images, considering maximum on vertical section at 5 min sampling rate, covering an area 240 km diameter 500 m horizontal...

Reusing software by copying and pasting is a common practice in development. This phenomenon widely known as code cloning. Problems with clones are mainly due to the need of managing each duplication, thus increasing effort maintain systems. Clone detection approaches generally take into account either syntactic structure (e.g., Abstract Syntax Tree) or lexical elements signature function). In this paper we propose an approach detect clones, based on information enriched elements. To end,...

We introduce the CDRP (Concatenated Diagnostic-Relapse Prognostic) architecture for multi-task deep learning that incorporates a clinical algorithm, e.g., risk stratification schema to improve prognostic profiling. present first application survival prediction in High-Risk (HR) Neuroblastoma from transcriptomics data, task studies MAQC consortium have shown remain hardest among multiple diagnostic and endpoints predictable same dataset. To obtain more accurate needed appropriate treatment...

Source code comments provide useful information on the implementation of a software and intent behind design decisions goals. Writing informative is far from being trivial task. Moreover, source tend to remain mostly unchanged during maintenance activities. As consequence, provided in comment method its corresponding may be not coherent with each other (i.e., The does properly describe implementation). In this paper, we present results manual assessment coherence between implementations 3636...

Abstract Almost every clinical specialty will use artificial intelligence in the future. The first area of practical impact is expected to be rapid and accurate interpretation image streams such as radiology scans, histo-pathology slides, ophthalmic imaging, any other bioimaging diagnostic systems, enriched by phenotypes used outcome labels or additional descriptors. In this study, we introduce a machine learning framework for automatic that combines current pattern recognition approach...

The use of spreadsheets to implement Information Systems is widespread in industry. Scripting languages and ad-hoc frameworks (e.g., Visual Basic for Applications) Rapid Application Development are often exploited by organizations quickly develop Spreadsheets-based supporting the information management their business processes. Maintenance tasks on these systems can be very difficult cause a remarkable worsening overall system quality. To prevent issues, migration such new architectures may...

Convolutional Neural Networks (CNNs) are a popular deep learning architecture widely applied in different domains, particular classifying over images, for which the concept of convolution with filter comes naturally. Unfortunately, requirement distance (or, at least, neighbourhood function) input feature space has so far prevented its direct use on data types such as omics data. However, number metrizable, i.e., they can be endowed metric structure, enabling to adopt convolutional based...

Social media represent an unrivalled opportunity for epidemiological cohorts to collect large amounts of high-resolution time course data on mental health. Equally, the high-quality held by could greatly benefit social research as a source ground truth validating digital phenotyping algorithms. However, there is currently lack software doing this in secure and acceptable manner. We worked with cohort leaders participants co-design open-source, robust expandable framework gathering cohorts.

Abstract Artificial Intelligence is exponentially increasing its impact on healthcare. As deep learning mastering computer vision tasks, application to digital pathology natural, with the promise of aiding in routine reporting and standardizing results across trials. Deep features inferred from scans can improve validity robustness current clinico-pathological features, up identifying novel histological patterns, e.g. tumor infiltrating lymphocytes. In this study, we examine issue evaluating...