Search

Cybersecurity is one of the great challenges of today's world. Rapid technological development has allowed society to prosper and improve the quality of life and the world is more dependent on new technologies. Managing security risks quickly and effectively, preventing, identifying, or mitigating them is a great challenge. The appearance of new attacks, and with more frequency, requires a constant update of threat detection methods. Traditional signature-based techniques are effective for known attacks, but they are not able to detect a new attack. For this reason, intrusion detection systems (IDS) that apply machine learning (ML) techniques represent an alternative that is gaining importance today. In this work, we have analyzed different machine learning techniques to determine which ones permit to obtain the best traffic classification results based on classification performance measurements and execution times, which is decisive for further real-time deployments. The CICIDS2017 dataset was selected in this work since it contains bidirectional traffic flows (derived from traffic captures) that include benign traffic and different types of up-to-date attacks. Each traffic flow is characterized by a set of connection-related attributes that can be used to model the traffic and distinguish between attacks and normal flows. The CICIDS2017 also contains the raw network traffic captures collected during the dataset creation in a packet-based format, thus permitting to extract the traffic flows from them. Various classification techniques have been evaluated using the Weka software: naive Bayes, logistic, multilayer perceptron, sequential minimal optimization, k-nearest neighbors, adaptive boosting, OneR, J48, PART, and random forest. As a general result, methods based on decision trees (PART, J48, and random forest) have turned out to be the most efficient with F1 values above 0.999 (average obtained in the complete dataset). Moreover, multiclass classification (distinguishing between different types of attack) and binary classification (distinguishing only between normal traffic and attack) have been compared, and the effect of reducing the number of attributes using the correlation-based feature selection (CFS) technique has been evaluated. By reducing the complexity in binary classification, better results can be obtained, and by selecting a reduced set of the most relevant attributes, less time is required (above 30% of decrease in the time required to test the model) at the cost of a small performance loss. The tree-based techniques with CFS attribute selection (six attributes selected) reached F1 values above 0.990 in the complete dataset. Finally, a conventional tool like Zeek has been used to process the raw traffic captures to identify the traffic flows and to obtain a reduced set of attributes from these flows. The classification results obtained using tree-based techniques (with 14 Zeek-based attributes) were also very high, with F1 above 0.997 (average obtained in the complete dataset) and low execution times (allowing several hundred thousand flows/s to be processed). These classification results obtained on the CICIDS2017 dataset allow us to affirm that the tree-based machine learning techniques may be appropriate in the flow-based intrusion detection problem and that algorithms, such as PART or J48, may offer a faster alternative solution to the RF technique. [ABSTRACT FROM AUTHOR]

This study evaluates the impacts of slot tagging and training data length on joint natural language understanding (NLU) models for medication management scenarios using chatbots in Spanish. In this study, we define the intents (purposes of the sentences) for medication management scenarios and two types of slot tags. For training the model, we generated four datasets, combining long/short sentences with long/short slots, while for testing, we collect the data from real interactions of users with a chatbot. For the comparative analysis, we chose six joint NLU models (SlotRefine, stack-propagation framework, SF-ID network, capsule-NLU, slot-gated modeling, and a joint SLU-LM model) from the literature. The results show that the best performance (with a sentence-level semantic accuracy of 68.6%, an F 1 -score of 76.4% for slot filling, and an accuracy of 79.3% for intent detection) is achieved using short sentences and short slots. Our results suggest that joint NLU models trained with short slots yield better results than those trained with long slots for the slot filling task. The results also indicate that short slots could be a better choice for the dialog system because of their simplicity. Importantly, the work demonstrates that the performance of the joint NLU models can be improved by selecting the correct slot configuration according to the usage scenario. [ABSTRACT FROM AUTHOR]

The widespread adoption of smartphones and the new-generation wireless networks have changed the way that people interact among themselves and with their environment. The use of messaging platforms, such as WhatsApp, has become deeply ingrained in peoples' lives, and many digital services have started to be delivered using these communication channels. In this work, we propose a new OAuth grant type to be used when the interaction between the resource owner and the client takes place through a messaging platform. This new grant type firstly allows the authorization server to be sure that no Man-in-the-Middle risk exists between the resource owner and the client before issuing an access token. Secondly, it allows the authorization server to interact with the resource owner through the same user-agent already being used to interact with the client, i.e., the messaging platform, which is expected to improve the overall user experience of the authorization process. To verify this assumption, we conducted a usability study in which subjects were required to perform the full authorization process using both the standard authorization code grant type (through a web-browser) and the new grant type defined in this work. They have also been required to fill in a small questionnaire including some demographic information and their impressions about both authorization flows. The results suggest that the proposed grant type eases the authorization process in most cases. [ABSTRACT FROM AUTHOR]

Heart rate variability (HRV) analysis is a noninvasive tool widely used to assess autonomic nervous system state. The market for wearable devices that measure the heart rate has grown exponentially, as well as their potential use for healthcare and wellbeing applications. Still, there is a lack of validation of these devices. In particular, this work aims to validate the Apple Watch in terms of HRV derived from the RR interval series provided by the device, both in temporal (HRM (mean heart rate), SDNN, RMSSD and pNN50) and frequency (low and high frequency powers, LF and HF) domain. For this purpose, a database of 20 healthy volunteers subjected to relax and a mild cognitive stress was used. First, RR interval series provided by Apple Watch were validated using as reference the RR interval series provided by a Polar H7 using Bland-Altman plots and reliability and agreement coefficients. Then, HRV parameters derived from both RR interval series were compared and their ability to identify autonomic nervous system (ANS) response to mild cognitive stress was studied. Apple Watch measurements presented very good reliability and agreement (>0.9). RR interval series provided by Apple Watch contain gaps due to missing RR interval values (on average, 5 gaps per recording, lasting 6.5 s per gap). Temporal HRV indices were not significantly affected by the gaps. However, they produced a significant decrease in the LF and HF power. Despite these differences, HRV indices derived from the Apple Watch RR interval series were able to reflect changes induced by a mild mental stress, showing a significant decrease of HF power as well as RMSSD in stress with respect to relax, suggesting the potential use of HRV measurements derived from Apple Watch for stress monitoring. [ABSTRACT FROM AUTHOR]

The photoplethysmogram (PPG) is a biomedical signal that can be used to estimate volumetric blood flow changes in the peripheral circulation. During the past few years, several works have been published in order to assess the potential for PPGs to be used in biometric authentication systems, but results are inconclusive. In this paper we perform an analysis of the feasibility of using the PPG as a realistic biometric alternative in the long term. Several feature extractors (based on the time domain and the Karhunen–Loève transform) and matching metrics (Manhattan and Euclidean distances) have been tested using four different PPG databases (PRRB, MIMIC-II, Berry, and Nonin). We show that the false match rate (FMR) and false non-match rate (FNMR) values remain constant in different time instances for a selected threshold, which is essential for using the PPG for biometric authentication purposes. On the other hand, obtained equal error rate (EER) values for signals recorded during the same session range from 1.0% for high-quality signals recorded in controlled conditions to 8% for those recorded in conditions closer to real-world scenarios. Moreover, in certain scenarios, EER values rise up to 23.2% for signals recorded over different days, signaling that performance degradation could take place with time. [ABSTRACT FROM AUTHOR]