Analisi delle preferenze per le diverse categorie di pomodoro trasformato nel food service

Il pomodoro è una delle colture principali del panorama agro-alimentare italiano e rappresenta un ingrediente base della tradizione culinaria nazionale. Il pomodoro lavorato dall’industria conserviera può essere trasformato in diverse tipologie merceologiche, che si differenziano in base alla tecniche di lavorazione impiegate ed alle caratteristiche del prodotto finito. la percentuale di spesa totale destinata all’acquisto di cibo fuori casa è in aumento a livello globale e l’interesse dell’industria alimentare nei confronti di questo canale di vendita è quindi crescente. Mentre sono numerose le indagine in letteratura che studiano i processi di acquisto dei consumatori finali, non ci sono evidenze di studi simili condotti sugli operatori del Food Service. Obiettivo principale della ricerca è quello di valutare le preferenze dei responsabili acquisti del settore Food Service per diverse tipologie di pomodoro trasformato, in relazione ad una gamma di attributi rilevanti del prodotto e di caratteristiche del cliente. La raccolta dei dati è avvenuta attraverso un esperimento di scelta ipotetico realizzato in Italia e alcuni mercati esteri. Dai risultati ottenuti dall’indagine emerge che i Pelati sono la categoria di pomodoro trasformato preferita dai responsabili degli acquisti del settore Food Service intervistati, con il 35% delle preferenze dichiarate nell'insieme dei contesti di scelta proposti, seguita dalla Polpa (25%), dalla Passata (20%) e dal Concentrato (15%). Dai risultati ottenuti dalla stima del modello econometrico Logit a parametri randomizzati è emerso che alcuni attributi qualitativi di fiducia (credence), spesso impiegati nelle strategie di differenziazione e posizionamento da parte dell’industria alimentare nel mercato Retail, possono rivestire un ruolo importante anche nell’influenzare le preferenze degli operatori del Food Service. Questo potrebbe quindi essere un interessante filone di ricerca da sviluppare nel futuro, possibilmente con l'impiego congiunto di metodologie di analisi basate su esperimenti di scelta ipotetici e non ipotetici.

Machine learning as a service for high energy physics (MLaaS4HEP): a service for ML-based data analyses

With the CERN LHC program underway, there has been an acceleration of data growth in the High Energy Physics (HEP) field and the usage of Machine Learning (ML) in HEP will be critical during the HL-LHC program when the data that will be produced will reach the exascale. ML techniques have been successfully used in many areas of HEP nevertheless, the development of a ML project and its implementation for production use is a highly time-consuming task and requires specific skills. Complicating this scenario is the fact that HEP data is stored in ROOT data format, which is mostly unknown outside of the HEP community. The work presented in this thesis is focused on the development of a ML as a Service (MLaaS) solution for HEP, aiming to provide a cloud service that allows HEP users to run ML pipelines via HTTP calls. These pipelines are executed by using the MLaaS4HEP framework, which allows reading data, processing data, and training ML models directly using ROOT files of arbitrary size from local or distributed data sources. Such a solution provides HEP users non-expert in ML with a tool that allows them to apply ML techniques in their analyses in a streamlined manner. Over the years the MLaaS4HEP framework has been developed, validated, and tested and new features have been added. A first MLaaS solution has been developed by automatizing the deployment of a platform equipped with the MLaaS4HEP framework. Then, a service with APIs has been developed, so that a user after being authenticated and authorized can submit MLaaS4HEP workflows producing trained ML models ready for the inference phase. A working prototype of this service is currently running on a virtual machine of INFN-Cloud and is compliant to be added to the INFN Cloud portfolio of services.

User scheduling for low earth orbit multi-user multiple-input-multiple-output non-terrestrial network systems

In rural and isolated areas without cellular coverage, Satellite Communication (SatCom) is the best candidate to complement terrestrial coverage. However, the main challenge for future generations of wireless networks will be to meet the growing demand for new services while dealing with the scarcity of frequency spectrum. As a result, it is critical to investigate more efficient methods of utilizing the limited bandwidth; and resource sharing is likely the only choice. The research community’s focus has recently shifted towards the interference management and exploitation paradigm to meet the increasing data traffic demands. In the Downlink (DL) and Feedspace (FS), LEO satellites with an on-board antenna array can offer service to numerous User Terminals (UTs) (VSAT or Handhelds) on-ground in FFR schemes by using cutting-edge digital beamforming techniques. Considering this setup, the adoption of an effective user scheduling approach is a critical aspect given the unusually high density of User terminals on the ground as compared to the on-board available satellite antennas. In this context, one possibility is that of exploiting clustering algorithms for scheduling in LEO MU-MIMO systems in which several users within the same group are simultaneously served by the satellite via Space Division Multiplexing (SDM), and then these different user groups are served in different time slots via Time Division Multiplexing (TDM). This thesis addresses this problem by defining a user scheduling problem as an optimization problem and discusses several algorithms to solve it. In particular, focusing on the FS and user service link (i.e., DL) of a single MB-LEO satellite operating below 6 GHz, the user scheduling problem in the Frequency Division Duplex (FDD) mode is addressed. The proposed State-of-the-Art scheduling approaches are based on graph theory. The proposed solution offers high performance in terms of per-user capacity, Sum-rate capacity, SINR, and Spectral Efficiency.

An architecture for network acceleration as a service in the cloud continuum

The pervasive availability of connected devices in any industrial and societal sector is pushing for an evolution of the well-established cloud computing model. The emerging paradigm of the cloud continuum embraces this decentralization trend and envisions virtualized computing resources physically located between traditional datacenters and data sources. By totally or partially executing closer to the network edge, applications can have quicker reactions to events, thus enabling advanced forms of automation and intelligence. However, these applications also induce new data-intensive workloads with low-latency constraints that require the adoption of specialized resources, such as high-performance communication options (e.g., RDMA, DPDK, XDP, etc.). Unfortunately, cloud providers still struggle to integrate these options into their infrastructures. That risks undermining the principle of generality that underlies the cloud computing scale economy by forcing developers to tailor their code to low-level APIs, non-standard programming models, and static execution environments. This thesis proposes a novel system architecture to empower cloud platforms across the whole cloud continuum with Network Acceleration as a Service (NAaaS). To provide commodity yet efficient access to acceleration, this architecture defines a layer of agnostic high-performance I/O APIs, exposed to applications and clearly separated from the heterogeneous protocols, interfaces, and hardware devices that implement it. A novel system component embodies this decoupling by offering a set of agnostic OS features to applications: memory management for zero-copy transfers, asynchronous I/O processing, and efficient packet scheduling. This thesis also explores the design space of the possible implementations of this architecture by proposing two reference middleware systems and by adopting them to support interactive use cases in the cloud continuum: a serverless platform and an Industry 4.0 scenario. A detailed discussion and a thorough performance evaluation demonstrate that the proposed architecture is suitable to enable the easy-to-use, flexible integration of modern network acceleration into next-generation cloud platforms.