Search

The rise of programmable data plane (PDP) and in-network computing (INC) paradigms paves the way for the development of network devices (switches, network interface cards, etc.) capable of performing advanced computing tasks. This allows to execute algorithms of various nature, including machine learning ones, within the network itself to support user and network services. In particular, this paper delves into the issue of implementing in-network learning models to support distributed intrusion detection systems (IDS). It proposes a model that optimally distributes the IDS workload, resulting from the subdivision of a "Strong Learner" (SL) model into lighter distributed "Weak Learner" (WL) models, among data plane devices; the objective is to ensure complete network security without excessively burdening their normal operations. Furthermore, a meta-heuristic approach is proposed to reduce the long computational time required by the exact solution provided by the mathematical model, and its performance is evaluated. The analysis conducted and the results obtained demonstrate the enormous potential of the proposed new approach to the creation of intelligent data planes that effectively act as a first line of defense against cyber attacks, with minimal additional workload on network devices.

The problem of attacks on new generation network infrastructures is becoming increasingly relevant, given the widening of the attack surface of these networks resulting from the greater number of devices that will access them in the future (sensors, actuators, vehicles, household appliances, etc.). Approaches to the design of intrusion detection systems must evolve and go beyond the traditional concept of perimeter control to build on new paradigms that exploit the typical characteristics of future 5G and 6G networks, such as in-network computing and intelligent programmable data planes. The aim of this research is to propose a disruptive paradigm in which devices in a typical data plane of a future programmable network have %classification and anomaly detection capabilities and cooperate in a fully distributed fashion to act as an ML-enabled Active Intrusion Detection System "embedded" into the network. The reported proof-of-concept experiments demonstrate that the proposed paradigm allows working effectively and with a good level of precision while occupying overall less CPU and RAM resources of the devices involved., Comment: This work has been submitted to the IEEE for possible publication

The pursuit of targeted therapies capable of overcoming biological barriers, including the tenacious blood-brain barrier, has spurred the investigation into stimuli-responsive microagents. This approach could improve therapeutic efficacy, reduce undesirable side effects, and open avenues for treating previously incurable diseases. Intravenously-administered ultrasound-responsive microbubbles are one of the most promising agents, having demonstrated potential in several clinical trials. However, the mechanism by which microbubbles enhance drug absorption remains unclear. Here, we reveal through unprecedented time-resolved side-view visualisations that single microbubbles, upon microsecond-long ultrasound driving, puncture the cell membrane and induce drug uptake via stable cyclic microjets. Our theoretical models successfully reproduce the observed bubble and cell dynamic responses. We find that cyclic jets arise from shape instabilities, warranting recognition as a novel class of jets in bubbles, distinct from classical inertial jets driven by pressure gradients. We also establish a threshold for bubble radial expansion beyond which microjets form and facilitate cellular permeation. Remarkably, these microjets occur at ultrasound pressures below 100 kPa due to their unique formation mechanism. We show that the stress generated by microjetting surpasses all previously suggested mechanisms by at least an order of magnitude. In summary, this work elucidates the physics behind microbubble-mediated targeted drug delivery and provides criteria for its effective yet safe application.

Context: Software Quality Assurance (SQA) is a fundamental part of software engineering to ensure stakeholders that software products work as expected after release in operation. Machine Learning (ML) has proven to be able to boost SQA activities and contribute to the development of quality software systems. In this context, Causal Reasoning is gaining increasing interest as a methodology to go beyond a purely data-driven approach by exploiting the use of causality for more effective SQA strategies. Objective: Provide a broad and detailed overview of the use of causal reasoning for SQA activities, in order to support researchers to access this research field, identifying room for application, main challenges and research opportunities. Methods: A systematic review of the scientific literature on causal reasoning for SQA. The study has found, classified, and analyzed 86 articles, according to established guidelines for software engineering secondary studies. Results: Results highlight the primary areas within SQA where causal reasoning has been applied, the predominant methodologies used, and the level of maturity of the proposed solutions. Fault localization is the activity where causal reasoning is more exploited, especially in the web services/microservices domain, but other tasks like testing are rapidly gaining popularity. Both causal inference and causal discovery are exploited, with the Pearl's graphical formulation of causality being preferred, likely due to its intuitiveness. Tools to favour their application are appearing at a fast pace - most of them after 2021. Conclusions: The findings show that causal reasoning is a valuable means for SQA tasks with respect to multiple quality attributes, especially during V&V, evolution and maintenance to ensure reliability, while it is not yet fully exploited for phases like ..., Comment: Accepted to Information and Software Technology journal

The paper explores the Biased Random-Key Genetic Algorithm (BRKGA) in the domain of logistics and vehicle routing. Specifically, the application of the algorithm is contextualized within the framework of the Vehicle Routing Problem with Occasional Drivers and Time Window (VRPODTW) that represents a critical challenge in contemporary delivery systems. Within this context, BRKGA emerges as an innovative solution approach to optimize routing plans, balancing cost-efficiency with operational constraints. This research introduces a new BRKGA, characterized by a variable mutant population which can vary from generation to generation, named BRKGA-VM. This novel variant was tested to solve a VRPODTW. For this purpose, an innovative specific decoder procedure was proposed and implemented. Furthermore, a hybridization of the algorithm with a Variable Neighborhood Descent (VND) algorithm has also been considered, showing an improvement of problem-solving capabilities. Computational results show a better performances in term of effectiveness over a previous version of BRKGA, denoted as MP. The improved performance of BRKGA-VM is evident from its ability to optimize solutions across a wide range of scenarios, with significant improvements observed for each type of instance considered. The analysis also reveals that VM achieves preset goals more quickly compared to MP, thanks to the increased variability induced in the mutant population which facilitates the exploration of new regions of the solution space. Furthermore, the integration of VND has shown an additional positive impact on the quality of the solutions found., Comment: 25 pages, 9 figures

Deep Neural Networks (DNN) are core components for classification and regression tasks of many software systems. Companies incur in high costs for testing DNN with datasets representative of the inputs expected in operation, as these need to be manually labelled. The challenge is to select a representative set of test inputs as small as possible to reduce the labelling cost, while sufficing to yield unbiased high-confidence estimates of the expected DNN accuracy. At the same time, testers are interested in exposing as many DNN mispredictions as possible to improve the DNN, ending up in the need for techniques pursuing a threefold aim: small dataset size, trustworthy estimates, mispredictions exposure. This study presents DeepSample, a family of DNN testing techniques for cost-effective accuracy assessment based on probabilistic sampling. We investigate whether, to what extent, and under which conditions probabilistic sampling can help to tackle the outlined challenge. We implement five new sampling-based testing techniques, and perform a comprehensive comparison of such techniques and of three further state-of-the-art techniques for both DNN classification and regression tasks. Results serve as guidance for best use of sampling-based testing for faithful and high-confidence estimates of DNN accuracy in operation at low cost., Comment: Accepted for publication at ICSE 2024, Lisbon, Portugal

We briefly report and elaborate on some conditions allowing a hydrodynamic description of the impact of a very short and arbitrarily intense laser pulse onto a cold plasma, as well as the localization of the first wave-breaking due to the plasma inhomogeneity. We use a recently developed fully relativistic plane model whereby we reduce the system of the Lorentz-Maxwell and continuity PDEs into a 1-parameter family of decoupled systems of non-autonomous Hamilton equations in dimension 1, with the light-like coordinate $\xi=ct\!-\!z$ replacing time $t$ as an independent variable. Apriori estimates on the Jacobian $\hat J$ of the change from Lagrangian to Eulerian coordinates in terms of the input data (initial density and pulse profile) are obtained applying Liapunov direct method to an associated family of pairs of ODEs; wave-breaking is pinpointed by the inequality $\hat J\le 0$. These results may help in drastically simplifying the study of extreme acceleration mechanisms of electrons, which have very important applications., Comment: 15 pages, 3 figures. arXiv admin note: substantial text overlap with arXiv:2207.11188

Deep Neural Networks (DNN) are nowadays largely adopted in many application domains thanks to their human-like, or even superhuman, performance in specific tasks. However, due to unpredictable/unconsidered operating conditions, unexpected failures show up on field, making the performance of a DNN in operation very different from the one estimated prior to release. In the life cycle of DNN systems, the assessment of accuracy is typically addressed in two ways: offline, via sampling of operational inputs, or online, via pseudo-oracles. The former is considered more expensive due to the need for manual labeling of the sampled inputs. The latter is automatic but less accurate. We believe that emerging iterative industrial-strength life cycle models for Machine Learning systems, like MLOps, offer the possibility to leverage inputs observed in operation not only to provide faithful estimates of a DNN accuracy, but also to improve it through remodeling/retraining actions. We propose DAIC (DNN Assessment and Improvement Cycle), an approach which combines ''low-cost'' online pseudo-oracles and ''high-cost'' offline sampling techniques to estimate and improve the operational accuracy of a DNN in the iterations of its life cycle. Preliminary results show the benefits of combining the two approaches and integrating them in the DNN life cycle., Comment: Paper accepted at 45th International Conference on Software Engineering (ICSE'23 NIER), May 2023

Conflict-Free Vehicle Routing Problems (CF-VRPs) arise in manufacturing, transportation and logistics facilities where Automated Guided Vehicles (AGVs) are utilized to move loads. Unlike \textit{Vehicle Routing Problems} arising in distribution management, CF-VRPs explicitly consider the limited capacity of the arcs of the guide path network to avoid collisions among vehicles. AGV applications have two peculiar features. First, the uncertainty affecting both travel times and machine ready times may result in vehicle delays or anticipations with respect to the fleet nominal plan. Second, the relatively high vehicle speed (in the order of one or two meters per second) requires vehicle plans to be revised in a very short amount of time (usually few milliseconds) in order to avoid collisions. In this paper we present fast exact algorithms to recover plan feasibility in real-time. In particular, we identify two corrective actions that can be implemented in real-time and formulate the problem as a linear program with the aim to optimize four common performance measures (total vehicle delay, total weighted delay, maximum route duration and total lateness). Moreover, we develop tailored algorithms which, tested on randomly generated instances of various sizes, prove to be three orders of magnitude faster than using off-the-shelf solvers.

COVID-19 has caused considerable disruption to education around the world. Disadvantaged and marginalised learners are being particularly hard-hit. Naturally, earlier in the pandemic, the focus of much attention was on how to safely reopen schools, often featuring a preoccupation with hygiene and social distancing considerations. A shift is now noticeable and welcome. With schools in many jurisdictions reopening partially or fully, there is now a growing interest in the immensely important area of recovering the learning lost while pupils have been away from face-to-face education. This document summarises the key findings from two larger reports conducted as part of the first phase of this research collaboration. These two reports are a review of policy and grey literature and a rapid evidence assessment (REA). [The report is written with assistance from Susy Ndaruhutse, Ella Page and Rosie Leonard-Kane. For "Learning Loss, Learning Gains and Wellbeing: A Review of Policy and Grey Literature," see ED615067. For "Learning Loss, Learning Gains and Wellbeing: A Rapid Evidence Assessment," see ED615066.]

As schools begin reopening around the world, education systems are faced with the challenge of mitigating learning lost during the pandemic, designing and implementing remedial or accelerated learning programmes, remote learning and next year's curriculum. Effective teachers are a crucial part of supporting learning recovery. School closure during COVID-19 has left many teachers uncertain about their role, worried about their working conditions and their health, safety and wellbeing, unable to use technology effectively, and unprepared for classroom challenges when schools reopen. Teachers will need to respond to students' academic losses (and gains), but also to their socio-emotional wellbeing. United Nations Educational, Scientific and Cultural Organisation (UNESCO) and Education Development Trust have undertaken this study to provide information to help teachers, schools and governments understand how to support teachers to best support students as they return to school. This research is designed to allow to focus on the available existing evidence, and focuses on the most marginalised students and include research from previous crisis and disaster contexts to examine broadly what is know about actions to mitigate and recover learning loss, not just related to the current global pandemic but in a broader sense. The key research questions for this Rapid Evidence Assessment (REA) are grouped into two sections. Firstly, questions about learning loss, gain and impact on student wellbeing. The second group of questions focus on the experience of teachers and the literature around the skills and support required for teachers to recover learning and wellbeing themselves and in students as they return to school. The REA is a companion to a review of policy and grey literature which focuses more on commentary and guidance literature. These papers are the first stage in a wider research project which will look at how teachers support learning recovery post COVID-19. [For "Learning Loss, Learning Gains and Wellbeing: A Review of Policy and Grey Literature," see ED615067.]

We determine conditions allowing to simplify the description of the impact of a short and arbitrarily intense laser pulse onto a cold plasma at rest. If both the initial plasma density and pulse profile have plane simmetry, then suitable matched upper bounds on the maximum and the relative variations of the initial density, as well as the intensity and duration of the pulse, ensure a strictly hydrodynamic evolution of the electron fluid (without wave-breaking or vacuum-heating) during its whole interaction with the pulse, while ions can be regarded as immobile. We use a recently developed fully relativistic plane model whereby the system of the (Lorentz-Maxwell and continuity) PDEs is reduced into a family of highly nonlinear but decoupled systems of non-autonomous Hamilton equations with one degree of freedom, with the light-like coordinate $\xi=ct\!-\!z$ instead of time $t$ as an independent variable, and new apriori estimates (eased by use of a Liapunov function) of the solutions in terms of the input data (initial density and pulse profile). If the laser spot radius $R$ is finite but not too small the same conclusions hold for the part of the plasma close to the axis $\vec{z}$ of cylindrical symmetry. These results may help in drastically simplifying the study of extreme acceleration mechanisms of electrons., Comment: 29 pages, 8 figures. To appear in the Journal "Mathematics"

The Tourist Trip Design Problem aims to prescribe a sightseeing plan that maximizes tourist satisfaction while taking into account a multitude of parameters and constraints, such as the distances among points of interest, the expected duration of each visit, the opening hours of each attraction, the time available daily. In this article we deal with a variant of the problem in which the mobility environment consists of a pedestrian network and a road network. So, one plan includes a car tour with a number of stops from which pedestrian subtours to attractions (each with its own time windows) depart. We study the problem and develop a method to evaluate the feasibility of solutions in constant time, to speed up the search. This result is used to devise an ad-hoc iterated local search. Experimental results show that our approach can handle realistic instances with up to 3643 points of interest (over a seven day planning horizon) in few seconds.

Given a graph whose arc traversal times vary over time, the Time-Dependent Travelling Salesman Problem consists in finding a Hamiltonian tour of least total duration covering the vertices of the graph. The main goal of this work is to define tight upper bounds for this problem by reusing the information gained when solving instances with similar features. This is customary in distribution management, where vehicle routes have to be generated over and over again with similar input data. To this aim, we devise an upper bounding technique based on the solution of a classical (and simpler) time-independent Asymmetric Travelling Salesman Problem, where the constant arc costs are suitably defined by the combined use of a Linear Program and a mix of unsupervised and supervised Machine Learning techniques. The effectiveness of this approach has been assessed through a computational campaign on the real travel time functions of two European cities: Paris and London. The overall average gap between our heuristic and the best-known solutions is about 0.001\%. For 31 instances, new best solutions have been obtained., Comment: arXiv admin note: text overlap with arXiv:2009.07588

Deep Neural Networks (DNN) are typically tested for accuracy relying on a set of unlabelled real world data (operational dataset), from which a subset is selected, manually labelled and used as test suite. This subset is required to be small (due to manual labelling cost) yet to faithfully represent the operational context, with the resulting test suite containing roughly the same proportion of examples causing misprediction (i.e., failing test cases) as the operational dataset. However, while testing to estimate accuracy, it is desirable to also learn as much as possible from the failing tests in the operational dataset, since they inform about possible bugs of the DNN. A smart sampling strategy may allow to intentionally include in the test suite many examples causing misprediction, thus providing this way more valuable inputs for DNN improvement while preserving the ability to get trustworthy unbiased estimates. This paper presents a test selection technique (DeepEST) that actively looks for failing test cases in the operational dataset of a DNN, with the goal of assessing the DNN expected accuracy by a small and ''informative'' test suite (namely with a high number of mispredictions) for subsequent DNN improvement. Experiments with five subjects, combining four DNN models and three datasets, are described. The results show that DeepEST provides DNN accuracy estimates with precision close to (and often better than) those of existing sampling-based DNN testing techniques, while detecting from 5 to 30 times more mispredictions, with the same test suite size., Comment: Paper accept at 43rd ACM/IEEE International Conference on Software Engineering, Madrid, Spain. May 2021