Search

Artificial Intelligence and digitalization are disrupting many industries by transforming business processes and products. Automation and the new twenty-first century technologies are quickly changing business models and the organizational structures and practices in the private sector, such as manufacturing, logistics, marketing, banking and finance that actively react to changes and improvements (Klump, 2018; Chien et al., 2020). The rest of the paper is organized as follows. Section 2 offers an overview on the AI background and the main studies on AI and SMES, while Section 3 describes our research design and the methodological approach. The case study is described in Section 4 and Section 5 shows the main results of the analysis of interviews. Finally, Section 6 concludes with a discussion of the main findings and some avenues for further development of this study.

igital videos represent a fundamental informative source of those events that occur during a penal proceedings, which thanks to the technologies available nowadays, can be stored, organized and retrieved in short time and with low cost. Considering the dimension that a video source can assume with respect to a courtroom recording, various necessities have been highlighted by the main judicial actors: fast navigation of the stream, efficient access to data inside and effective representation of relevant contents. One of the possible solutions to these requirements is represented by multimedia summarization aimed at deriving a synthetic representation of audio/video contents, characterized by a limited loss of meaningful information. In this paper a multimedia summarization environment is proposed for defining a storyboard for proceedings celebrated into courtrooms.

This paper fits into the research domain aimed at inferring useful information by mining from data streams of poor quality in wireless sensor networks, with the objective to reduce at the same time the communication load and energy consumption. A set of nodes collect sensor readings from the environment and maintain local model of their evolution. The evaluation of these models is performed using them to simulate data evolution and evaluating the error w.r.t. new sensor readings: when this error for a model exceeds a threshold, model parameters are updated and sent to the sink. At the sink the values collected by sensors are known by using the parameters of local models to simulate sensor’s readings, minimizing the communication among sensors and sink and hence energy consumption. At the sink a global model, a Bayesian Network built on forecasted data, captures spatial and data dependencies among sensors, to detect single sensor and network wide anomalies missed by local error control.

The quick development and deployment of sensor technology within the general frame of the Internet of Things poses relevant opportunity and challenges. The sensor is not a pure data source, but an entity (Semantic Sensor Web) with associated metadata and it is a building block of a"worldwide distributed" real time database, to be processed through real-time queries. Important challenges are to achieve interoperability in connectivity and processing capabilities (queries) and to apply"intelligence" and processing capabilities as close as possible to the source of data. This paper presents the extension of a general architecture for data integration in which we add capabilities for processing of complex queries and discuss how they can be adapted to, and used by, an application in the Semantic Sensor Web, presenting a pilot study in environment and health domains.

Information and Communication Technologies play a fundamental role in e-justice: the traditional judicial folder is being transformed into an integrated multimedia folder, where documents, audio and video recordings can be accessed and searched via web-based judicial content management platforms. Usability of the electronic judicial folders is still hampered by traditional support toolset, allowing search only in textual information, rather than directly in audio and video recordings. Transcription of audio recordings and template filling are still largely manual activities. Thus a significant part of the information available in the trial folder is usable only through a time consuming manual search especially for audio and video recordings that describe not only what was said in the courtroom, but also the way and the specific trial context in which it was said. In this paper we present the JUMAS system, stemming from the TUMAS project started on February 2008, that takes up the challenge of using semantics towards a better usability of the multimedia judicial folders. The main aim of this paper is to show how JUMAS has provided the judicial users with a powerful toolset able to fully exploit the knowledge embedded into multimedia judicial folders

In this paper we present a software system for decision support (DSS - Decision Support System) aimed at forecasting high demand of admission on health care structures due to environmental pollution. The algorithmic kernel of the system is based on machine learning, the software architecture is such that both persistent and sensor data are integrated through a data integration infrastructure. Given the actual concentration of different pollutants, measured by a network of sensors, the DSS allows forecasting the demand of hospital admissions for acute diseases in the next 1 to 6 days. We tested our system on cardiovascular and respiratory diseases in the area of Milan

The high glycolytic activity of multiple myeloma (MM) cells is the rationale for use of Positron Emission Tomography (PET) with18F-fluorodeoxyglucose ([18F]FDG) to detect both bone marrow (BM) and extramedullary disease. However, new tracers are actively searched because [18F]FDG-PET has some limitations and there is a portion of MM patients who are negative. Glutamine (Gln) addiction has been recently described as a typical metabolic feature of MM cells. Yet, the possible exploitation of Gln as a PET tracer in MM has never been assessed so far and is investigated in this study in preclinical models. Firstly, we have synthesized enantiopure (2S,4R)-4-fluoroglutamine (4-FGln) and validated it as a Gln transport analogue in human MM cell lines, comparing its uptake with that of3H-labelled Gln. We then radiosynthesized [18F]4-FGln, tested its uptake in two differentin vivomurine MM models, and checked the effect of Bortezomib, a proteasome inhibitor currently used in the treatment of MM. Both [18F]4-FGln and [18F]FDG clearly identified the spleen as site of MM cell colonization in C57BL/6 mice, challenged with syngeneic Vk12598 cells and assessed by PET. NOD.SCID mice, subcutaneously injected with human MM JJN3 cells, showed high values of both [18F]4-FGln and [18F]FDG uptake. Bortezomib significantly reduced the uptake of both radiopharmaceuticals in comparison with vehicle at post treatment PET. However, a reduction of glutaminolytic, but not of glycolytic, tumor volume was evident in mice showing the highest response to Bortezomib. Our data indicate that [18F](2S,4R)-4-FGln is a new PET tracer in preclinical MM models, yielding a rationale to design studies in MM patients.

Background Multiple myeloma (MM) is a clonal plasma cell malignancy associated with osteolytic bone disease. Recently, the role of MM-derived exosomes in the osteoclastogenesis has been demonstrated although the underlying mechanism is still unknown. Since exosomes-derived epidermal growth factor receptor ligands (EGFR) are involved in tumor-associated osteolysis, we hypothesize that the EGFR ligand amphiregulin (AREG) can be delivered by MM-derived exosomes and participate in MM-induced osteoclastogenesis. Methods Exosomes were isolated from the conditioned medium of MM1.S cell line and from bone marrow (BM) plasma samples of MM patients. The murine cell line RAW264.7 and primary human CD14+ cells were used as osteoclast (OC) sources. Results We found that AREG was specifically enriched in exosomes from MM samples and that exosomes-derived AREG led to the activation of EGFR in pre-OC, as showed by the increase of mRNA expression of its downstream SNAIL in both RAW264.7 and CD14+ cells. The presence of neutralizing anti-AREG monoclonal antibody (mAb) reverted this effect. Consequently, we showed that the effect of MM-derived exosomes on osteoclast differentiation was inhibited by the pre-treatment of exosomes with anti-AREG mAb. In addition, we demonstrated the ability of MM-derived AREG-enriched exosomes to be internalized into human mesenchymal stromal cells (MSCs) blocking osteoblast (OB) differentiation, increasing MM cell adhesion and the release of the pro-osteoclastogenic cytokine interleukin-8 (IL8). Accordingly, anti-AREG mAb inhibited the release of IL8 by MSCs suggesting that both direct and indirect effects are responsible for AREG-enriched exosomes involvement on MM-induced osteoclastogenesis. Conclusions In conclusion, our data indicate that AREG is packed into MM-derived exosomes and implicated in OC differentiation through an indirect mechanism mediated by OBs. Electronic supplementary material The online version of this article (10.1186/s13045-018-0689-y) contains supplementary material, which is available to authorized users.

High glycolitic activity of multiple myeloma (MM) cells is the rationale for the use of Positron Emission Tomography (PET) with 18F-fluorodeoxyglucose ([18F]FDG) to detect both medullary and extramedullary disease. However, FDG-PET has some limitations, since there is a good portion of MM patients who are false-negative. Besides enhanced glycolysis, glutamine (Gln) addiction has been recently described as a metabolic feature of MM by our group. To sustain high Gln demand, MM cells increase the expression of several Gln transporters (ASCT2, SNAT1, LAT1) and are endowed with fast Gln uptake. Yet, at variance with other Gln-addicted cancers, the possible exploitation of Gln as a PET tracer in MM has never been assessed and was investigated in this study. To this purpose, we have firstly synthesized enantiopure (2S,4R)-4-Fluoroglutamine (4-FGln) and validated it as a Gln analogue in human MM cell lines (RPMI8226 and JJN3) comparing its uptake with that of 3H-labelled Gln. The intracellular levels of 4-FGln were determined by HPLC-MS/MS employing a HILIC gradient separation and multiple reaction monitoring (MRM) detection. Both Gln and 4-FGln were actively accumulated by MM cells and exhibited a strong reciprocal competition, pointing to shared transporters. Inhibition analysis revealed that ASCT2 was the major entry route of both compounds, with minor contributions from the other transporters. However, compared with Gln, 4-FGln exhibited higher affinity for both ASCT2 and LAT1 transporters. On the basis of these results, we then tested [18F]4-FGln uptake for MM detection by Positron Emission Tomography (PET) in two different in vivo murine models. Firstly, to investigate sensitivity of human MM to [18F]4-FGln in vivo, JJN3 cells were subcutaneously injected in immunodeficient NSG mice In this xenograft model, [18F]4-FGln- and[18F]FDG-PET scans were performed after plasmacytomas became palpable and repeated after one week. All the tumours were positive for [18F]FDG and displayed [18F]4-FGln uptake with Standard Uptake Values (SUV) of 1.21±1.9 and 0.99±0.07 after 2 weeks, respectively. Thereafter, the effect of bortezomib (BOR) was investigated to evaluate the potential use of [18F]4-FGln to monitor anti-MM treatment. Ten NGS mice were injected with JJN3 cells and, after 14 days, treated twice weekly with BOR, 1mg/kg, or vehicle for two weeks. PET scans were performed before and after 5 and 12 days of BOR treatment. As expected, BOR reduced tumour size as compared to vehicle. At the first post-BOR PET scan, [18F]4-FGln (SUV mean: pre 0.85±0.31; post 0.45±0.10, P Thereafter, to mimic BOR-resistant MM in a syngeneic mouse model, C57BL/6 mice were injected intravenously with Vk12598 cells obtained from transgenic Vk*MYC mice repeatedly treated with sub-optimal doses of BOR. Upon injection into C57BL/6 mice, Vk12598 cells colonize the BM without lytic lesions and extensively colonize the spleen generating an aggressive MM that brings animals to death within five weeks. PET scans were performed with [18F]4-FGln and [18F]FDG before Vk*MYC MM cells injection and after three, four and five weeks. Blood samples for M-spike evaluation were obtained in parallel. Four weeks after MM cells injection a significant increase of both [18F]4-FGln and [18F]FDG uptake was detected in spleens (SUV mean: 1.14±0.23, P=0.018; 0.94±0.24, P= 0.005). In both MM models, the volume of distribution of [18F]4-F-Gln did not overlap that of [18F]FDG. In conclusion, our data indicate that [18F]-(2S,4R)-4-Fluoroglutamine is a new potential PET tracer in pre-clinical MM models especially of extramedullary disease, either in a BOR-sensitive or in a BOR-resistant context, supporting the exploitation of Gln addiction for diagnostic purposes in MM patients. Disclosures Giuliani: Janssen: Research Funding.

In this paper we present a Decision Support System (DSS) aimed at forecasting high demand of admission on health care structures due to environmental pollution. The computational engine of the DSS is based on Autoregressive Hidden Markov Models (AHMM). We estimate the forecasting model by analyzing the historical daily average concentrations of pollutants and the number of hospital admissions, collected from multiple data sources. Given the actual concentration of different pollutants, measured by a network of sensors, the DSS allows to forecast the demand of hospital admissions for acute diseases in the following 1 to 6 days. We tested our system on cardiovascular and respiratory diseases in the area of Milan. The performances of our system, compared with multiple linear regression, show that AHMM are a robust approach to capture the connections among health and environmental indicators.

This paper presents a framework for managing data from sensor of poor quality, withthe objective to reduce at the same time the communication load and hence energy consumption.Each node in a wireless sensor network maintains a simple local model of the data it is collectingand sends its parameters to a central location (sink), where it is executed the global monitoring.Local models are used to simulate sensor's readings, minimising the need of communication withsensors and hence the consumption of their battery; they are updated locally, when sensorreadings differ excessively from simulated data. At the sink the global model (a BayesianNetwork) is learnt on the simulated data. It is used to identify and replace anomalous readings(outliers) that a sensor should have produced and to detect anomalies missed by any single node(when communication with a sensor is interrupted). © 2010 Inderscience Enterprises Ltd.