Search

The science of social bots seeks knowledge and solutions to one of the most debated forms of online misinformation. Yet, social bots research is plagued by widespread biases, hyped results, and misconceptions that set the stage for ambiguities, unrealistic expectations, and seemingly irreconcilable findings. Overcoming such issues is instrumental towards ensuring reliable solutions and reaffirming the validity of the scientific method. In this contribution we revise some recent results in social bots research, highlighting and correcting factual errors as well as methodological and conceptual issues. More importantly, we demystify common misconceptions, addressing fundamental points on how social bots research is discussed. Our analysis surfaces the need to discuss misinformation research in a rigorous, unbiased, and responsible way. This article bolsters such effort by identifying and refuting common fallacious arguments used by both proponents and opponents of social bots research as well as providing indications on the correct methodologies and sound directions for future research in the field.

Adversarial examples are inputs to a machine learning system that result in an incorrect output from that system. Attacks launched through this type of input can cause severe consequences: for example, in the field of image recognition, a stop signal can be misclassified as a speed limit indication.However, adversarial examples also represent the fuel for a flurry of research directions in different domains and applications. Here, we give an overview of how they can be profitably exploited as powerful tools to build stronger learning models, capable of better-withstanding attacks, for two crucial tasks: fake news and social bot detection., To appear on IEEE Internet Computing. `Accepted manuscript' version

The heavy metals transfer from the soil, where they accumulate, to the edible parts of the plants, and then, their entrance in the food chain can represent a source of concern for human health. Among heavy metals, arsenic is one of the most widespread in the soil of Lazio (central region of Italy), where the phytoavailable geogenic arsenic enters the food chain, with a dangerous exposition of the local population. In the first part of this work, plants of radish (Raphanus sativus L.) and lettuce (Lactuca sativa L.) were grown in protected culture in the experimental farm of CREA-AA, where they were daily treated with different concentrations of sodium arsenate dibasic heptahydrate in order to investigate differences in their arsenic accumulation capacities. In order to confirm the results achieved, in the second part of this study, the arsenic concentration was determined in commercial products obtained from contaminated areas of Lazio, and the potential exposition risk for human health through consumption of these widely consumed vegetables was estimated. The highest arsenic concentrations were found in the samples of lettuce. To evaluate the potential health risk from consumption of L. sativa and R. sativus, the estimated daily intake (EDI) for adults, adolescents, and elderly was calculated, finding that HRI (health risk index) index value for arsenic was low (< 1) in the case of chronic consumptions for all samples of radishes, and for the lettuces grown in the area of Viterbo. On the contrary, the lettuces obtained from Tuscania and Tarquinia presented very high concentrations of arsenic and a worrying HRI value. In order to reduce the risk of As toxicity in the people through consumption of the vegetables, the irrigation water should contain less than 0.1 mg As L−1. For this reason, the authors tested the application of red mud (RM) to remove As from groundwater before using it for the irrigation of radish and lettuce in greenhouse production.

Paclitaxel coated balloon catheters (PCB) were developed as a polymer-free non-implantable alternative to drug eluting stents, delivering similar drug payloads in a matter of minutes. While PCB have shown efficacy in treating peripheral arterial disease in certain patient groups, restenosis rates remain high and there is no class effect. To help further optimize these devices, we developed a scanning electron microscopy (SEM) imaging technique and computational modeling approach that provide insights into the coating micromorphology dependence of in vivo drug transfer and retention. PCBs coated with amorphous/flaky or microneedle coatings were inflated for 60 sec in porcine femoral arteries. Animals were euthanized at 0.5, 24 and 72 h and treated arteries processed for SEM to image endoluminal coating distribution followed by paclitaxel quantification by mass spectrometry (MS). Endoluminal surfaces exhibited sparse coating patches at 0.5 h, predominantly protruding (13.71 vs 0.59%, P 0.001), with similar micro-morphologies to nominal PCB surfaces. Microneedle coating covered a 1.5-fold endoluminal area (16.1 vs 10.7%, P = 0.0035) owing to higher proximal and distal delivery, and achieved 1.5-fold tissue concentrations by MS (1933 vs 1298 μg/g, P = 0.1745) compared to amorphous/flaky coating. Acute longitudinal coating distribution tracked computationally predicted microindentation pressure gradients (r = 0.9, P 0.001), with superior transfer of the microneedle coatings attributed to their amplification of angioplasty contact pressures. By 24 h, paclitaxel concentration and coated tissue areas both declined by93% even as nonprotruding coating levels were stable between 0.5 and 72 h, and 2.7-fold higher for microneedle vs flaky coating (0.64 vs 0.24%, P = 0.0195). Tissue retained paclitaxel concentrations at 24-72 h trended 1.7-fold higher post treatment with microneedle coating compared to the amorphous/flaky coating (69.9 vs 39.9 μg/g, P = 0.066). Thus, balloon based drug delivery is critically dependent on coating micromorphologies, with superior performance exhibited by micromorphologies that amplify angioplasty pressures.

In this study, we determined the concentration of six marker PCBS congeners and four trace elements, all having a maximum tolerable level of intake set by EU regulation, in samples of European hake taken from the Mediterranean Sea. The results showed that the contaminant levels were below the recommended international limits in all samples, although with higher PCBs concentrations in samples from Ligurian and Adriatic Sea. The relationships between the contaminants and the sampling sites were investigated using analysis of variance (ANOVA) and Tukey's HSD post-hoc tests. Statistical analysis revealed significant interactions especially between the PCBs, the sum of PCBs and the sites. To evaluate the potential health risk from consumption, we calculated the estimated daily intake (EDI) for adults, finding that total health risk index (HRI) value was low (< 1) in the case of chronic consumptions, despite the non-negligible mercury concentration.

The potential of Mid-infrared spectroscopy coupled with chemometric tools was evaluated for the authentication and discrimination of PDO (Protected Denomination of Origin) buffalo mozzarella produced by traditional and industrial cheese-making processes. Samples of mozzarella provided by local producers and supermarkets were analysed through both official destructive methods and Attenuated Total Reflectance-Fourier transform infrared spectroscopy (FTIR/ATR). In particular, destructive methods allowed to determine the content of fatty substances, proteins, moisture and total nitrogen. The results show that only the conjunction of MID-infrared spectroscopy with chemometric analysis can provide a satisfying solution to discriminate between the different types of mozzarella.

Urban agriculture plays an important role to provide an innovative and different connection to food. Visitors, scholars, and generally participants of community gardens activities become 'food citizens', shift from being passive food consumers to becoming co-producers. To achieve this goal, the demonstration garden of Tor Mancina with the involvement of schools and of local communities tests innovative and more sustainable agricultural practices and carries out experiments to test the phytostimulant and pest repellent effects of aromatic and officinalis plants extracts used as basic substances in plant protection management. These experiments aim at inculcating positive values on food, agriculture and environment in growing youth, providing effective solutions to increase crop performance, to enhance the tolerance of plants against stressors, to safeguard the nature, biodiversity and food quality. These activities teach sustainable agriculture practices which form basis for stable livelihood and informed consumption habits. For this reason more initiatives should be in place, nationwide, aimed at encouraging the visit of the community gardens and participation in their activities, in order to learn more about the provenance of food, agricultural processes, nutrition, safety and security, biodiversity and sustainability, and develop new skills.

Spambot detection is a must for the protection of cyberspace, in terms of both threats to sensitive information of users and trolls that may want to cheat and influence the public opinion. Unfortunately, new waves of malicious accounts are characterized by advanced features, making their detection extremely challenging. In contrast with the supervised spambot detectors largely used in recent years and inspired by biological DNA, we propose an alternative, unsupervised detection approach. Its novelty is based on the idea of modeling online user behaviors with strings of characters representing the sequence of the user's online actions. Exploiting this nature-inspired behavioral model, the proposed technique lets groups of spambots emerge from the crowd, by comparing the accounts' behaviors. Results show that the proposal outperforms the best-of-breed algorithms commonly employed for spambot detection.

DNA-inspired online behavioral modeling techniques have been proposed and successfully applied to a broad range of tasks. In this paper, we investigate the fundamental laws that drive the occurrence of behavioral similarities among Twitter users, employing a DNA-inspired technique. Our findings are multifold. First, we demonstrate that, despite apparently featuring little to no similarities, the online behaviors of Twitter users are far from being uniformly random. Secondly, we benchmark different behavioral models through a number of simulations. We characterize the main properties of such models and we identify those models that better resemble human behaviors in Twitter. Then, we demonstrate that the number and the extent of behavioral similarities within a group of Twitter users obey a log-normal law, and we leverage this characterization to propose a novel bot detection system. In a nutshell, the results shed light on the fundamental properties that drive the online behaviors of groups of Twitter users, through the lenses of DNA-inspired behavioral modeling techniques. This study is based on a wealth of data gathered over several months that, for the sake of reproducibility, are publicly available for research purposes.

The ADBIS conferences provide an international forum for the presentation of research on database theory, development of advanced DBMS technologies, and their applications. The 22nd edition of ADBIS, held on September 2–5, 2018, in Budapest, Hungary, includes six thematic workshops collecting contributions from various domains representing new trends in the broad research areas of databases and information systems.

The Internet of Things (IoT) revolution has not only carried the astonishing promise to interconnect a whole generation of traditionally “dumb” devices, but also brought to the Internet the menace of billions of badly protected and easily hackable objects. Not surprisingly, this sudden flooding of fresh and insecure devices fueled older threats, such as Distributed Denial of Service (DDoS) attacks. In this paper, we first propose an updated and comprehensive taxonomy of DDoS attacks, together with a number of examples on how this classification maps to real-world attacks. Then, we outline the current situation of DDoS-enabled malwares in IoT networks, highlighting how recent data support our concerns about the growing in popularity of these malwares. Finally, we give a detailed analysis of the general framework and the operating principles of Mirai, the most disruptive DDoS-capable IoT malware seen so far.

The Internet of Things (IoT) revolution has not only carried the astonishing promise to interconnect a whole generation of traditionally “dumb” devices, but also brought to the Internet the menace of billions of badly protected and easily hackable objects. Not surprisingly, this sudden flooding of fresh and insecure devices fueled older threats, such as Distributed Denial of Service (DDoS) attacks. In this paper, we first propose an updated and comprehensive taxonomy of DDoS attacks, together with a number of examples on how this classification maps to real-world attacks. Then, we outline the current situation of DDoS-enabled malwares in IoT networks, highlighting how recent data support our concerns about the growing in popularity of these malwares. Finally, we give a detailed analysis of the general framework and the operating principles of Mirai, the most disruptive DDoS-capable IoT malware seen so far.

Android smartphones have become a vital component of the daily routine of millions of people, running a plethora of applications available in the official and alternative marketplaces. Although there are many security mechanisms to scan and filter malicious applications, malware is still able to reach the devices of many end-users. In this paper, we introduce the SafeDroid v2.0 framework, that is a flexible, robust, and versatile open-source solution for statically analysing Android applications, based on machine learning techniques. The main goal of our work, besides the automated production of fully sufficient prediction and classification models in terms of maximum accuracy scores and minimum negative errors, is to offer an out-of-the-box framework that can be employed by the Android security researchers to efficiently experiment to find effective solutions: the SafeDroid v2.0 framework makes it possible to test many different combinations of machine learning classifiers, with a high degree of freedom and flexibility in the choice of features to consider, such as dataset balance and dataset selection. The framework also provides a server, for generating experiment reports, and an Android application, for the verification of the produced models in real-life scenarios. An extensive campaign of experiments is also presented to show how it is possible to efficiently find competitive solutions: the results of our experiments confirm that SafeDroid v2.0 can reach very good performances, even with highly unbalanced dataset inputs and always with a very limited overhead.

In this paper, we analyse a dataset of hotel reviews. In details, we enrich the review dataset, by extracting additional features, consisting of information on the reviewers' profiles and the reviewed hotels. We argue that the enriched data can gain insights on the factors that most influence consumers when composing reviews (e.g., if the appreciation for a certain kind of hotel is tied to specific users' profiles). Thus, we apply statistical analyses to reveal if there are specific characteristics of reviewers (almost) always related to specific characteristics of hotels. Our experiments are carried out on a very large dataset, consisting of around 190k hotel reviews, collected from the Tripadvisor website.

We envisage a revolutionary change in the approach to spambot detection: instead of taking countermeasures only after having collected evidence of new spambot mischiefs, in a near future techniques will be able to anticipate the ever-evolving spammers.

Coal plants can be a major source of mutagenic pollutants. In this study we used the common land snail Helix aspersa, to detect the mutagenic effect of pollution from a coal plant in central Italy applying the micronucleus test (MN) on snail's haemocytes and evaluating trace elements concentration (As Cd, Pb, Hg, and Zn) in soil and snails. Snails from a biological farm were exposed for 13 days in five locations at different distances from the plant. Wild snails collected in the same locations were also analysed. MN frequency in exposed snails was significantly higher in four locations within 10 km from to the plant, with respect to the control and the farthest location. Comparing the MN frequency between farmed and wild snails, a significantly higher frequency emerged for the exposed snails in all locations except the farthest, likely indicating adaptation or selection of the wild organisms due to chronic exposure to pollutants. In natural snails significantly higher MN frequencies with near the plant emerged as well. Trace elements analysis showed significant correlations between MN frequencies and both Zn and As concentrations in soil, for both exposed and wild snails, and Zn and Pb concentrations in exposed snails. Our results were consistent with those previously obtained when evaluating primary DNA damage in natural snails from the same area and show that the snails near the plant were affected by a permanent cytogenetic damage. Moreover, they confirm the suitability of snails for biomonitoring the presence of pollutants with mutagenic effect.

Recently, DNA-inspired online behavioral modeling and analysis techniques have been proposed and successfully applied to a broad range of tasks. In this paper, we employ a DNA-inspired technique to investigate the fundamental laws that drive the occurrence of similarities among Twitter users. The achieved results are multifold. First, we demonstrate that, despite apparently showing little to no similarities, the online behaviors of Twitter users are far from being uniformly random. Then, we perform a set of simulations to benchmark different behavioral models and to identify the models that better resemble human behaviors in Twitter. Finally, we demonstrate that the number and the extent of behavioral similarities within a group of Twitter users obey a log-normal distribution. Our results shed light on the fundamental properties that drive behaviors of groups of Twitter users, through the lenses of DNA-inspired behavioral modeling techniques. Our datasets are publicly available to the scientific community to further explore analytics of online behaviors.

More than a personal microblogging site, Twitter has been transformed by common use to an information publishing venue, which public characters, media channels and common people daily rely on for, e.g., news reporting and consumption, marketing, and social messaging. The use of Twitter in a cooperative and interactive setting calls for the precise awareness of the dynamics regulating message spreading. In this paper, we describe Twitlang, a language for modelling the interactions among Twitter accounts. The associated operational semantics allows users to precisely determine the effects of their actions on Twitter, such as post, reply-to or delete tweets. The language is implemented in the form of a Maude interpreter, Twitlanger, which takes a language term as an input and explores the computations arising from the term. By combining the strength of Twitlanger and the Maude model checker, it is possible to automatically verify communication properties of Twitter accounts. We illustrate the benefits of our executable formalisation by means of an application scenario inspired from real life. While the scenario highlights the benefits of adopting Twitter for a cooperative use in the everyday life, our analysis shows that appropriate settings are essential for a proper usage of the platform, in respect of fulfilling those communication properties expected within collaborative and interactive contexts.

Spambot detection in online social networks is a long-lasting challenge involving the study and design of detection techniques capable of efficiently identifying ever-evolving spammers. Recently, a new wave of social spambots has emerged, with advanced human-like characteristics that allow them to go undetected even by current state-of-the-art algorithms. In this paper, we show that efficient spambots detection can be achieved via an in-depth analysis of their collective behaviors exploiting the digital DNA technique for modeling the behaviors of social network users. Inspired by its biological counterpart, in the digital DNA representation the behavioral lifetime of a digital account is encoded in a sequence of characters. Then, we define a similarity measure for such digital DNA sequences. We build upon digital DNA and the similarity between groups of users to characterize both genuine accounts and spambots. Leveraging such characterization, we design the Social Fingerprinting technique, which is able to discriminate among spambots and genuine accounts in both a supervised and an unsupervised fashion. We finally evaluate the effectiveness of Social Fingerprinting and we compare it with three state-of-the-art detection algorithms. Among the peculiarities of our approach is the possibility to apply off-the-shelf DNA analysis techniques to study online users behaviors and to efficiently rely on a limited number of lightweight account characteristics.