Interpersonal conflict between couples is a significant source of stress with long-lasting effects on partners' physical and psychological health. Motivated by findings in psychological science, we study how couples with distinct relationship functioning characteristics experience conflict in real life. We propose sub-population specific machine learning models using hierarchical and adaptive learning frameworks to automatically detect interpersonal conflict through the ambulatory monitoring of couples' physiological signals, audio samples, and linguistic indices. Our results indicate that the proposed models reach an unweighted F1-score of 0.61 and outperform a general model learned for the entire population, providing a foundation toward personalized health applications.
The proliferation of connected embedded devices, or the Internet of Things (IoT), together with recent advances in machine intelligence, will change the profile of future cloud services and introduce a variety of new research problems centered around empowering resource-limited edge devices to exhibit intelligent behavior, both in sensing and control. Cloud services will enable learning from data, performing inference, and executing control, all with assurances on outcomes. The paper discusses such emerging services and outlines five resulting new research directions towards enabling and optimizing intelligent, cloud-assisted sensing and control in the age of the Internet of Things.
Word embeddings based on context are lack of capturing the sentiment information.This paper takes advantage of an emotional psychology model to learn the emotional embeddings in Chinese.We present two different purifying models (LPM and GPM) based on Plutchik?s wheel of emotions to add the emotional information into word vectors.
Detecting misogyny and xenophobia in Spanish tweets using language technologies
To address the real-time and reliability requirements of industrial IoT applications, IEEE 802.15.4-based wireless sensor-actuator networks made unique design choices such as employing the Time-Synchronized Channel Hopping (TSCH). The function-based channel hopping used in TSCH simplifies the network operations at the cost of security. Our study shows that an attacker can reverse engineer the channel hopping sequences and routes by silently observing the transmission activities and put the network in danger of selective jamming attacks. To our knowledge, this paper represents the first systematic study that investigates the security vulnerability of TSCH channel hopping and graph routing under realistic settings.
Fog nodes are dimensioned for an average traffic load and cannot handle sudden bursts. This paper addresses the problem of controlling Fog/Cloud admission to guarantee application response time. An Admission Control strategy based on online learning of the request popularity distribution via a Least-Recently-Used filter to maximize the number of requests handled by the Fog is introduced and analytically assessed. An implementation on FPGA hardware is proposed using Ageing Bloom Filters for memory efficiency. The implementation achieves a throughput of 16.7Mpps and a latency lower than 3µs, while multiplying the Fog acceptance-rate by 10 in the evaluated scenario.
The IoT domain has been significantly expanded with the proliferation of drones and unmanned robotic devices. In this new landscape the communication between the end device and the fixed infrastructure is similarly expanded to include new messages of varying importance. To control the exchange of such messages subject to the stochastic nature of the underlying wireless network we design a solution that exploits optimal stopping and change detection theories. Our findings are quite promising and solidly supportive to a vast spectrum of applications with quality of service requirements.
To solve the cold start and data sparsity problem, we propose a novel sparse trust recommendation model, SSL-SVD. Specifically, we decompose the aspects influencing trust building into finer-grained factors, and combine these factors to mine the implicit sparse trust relationships among users by employing the Transductive Support Vector Machine algorithm. Then we extend an existing state-of-the-art recommendation model with social trust and sparse trust information for rating prediction in the recommendation system. Experiments show that our SSL-SVD increases the trust density degree of each data set by more than 65% and improves the recommendation accuracy by up to 4.3%.
In recent years, the increasing propagation of hate speech in online social networks and the need for effective counter-measures have drawn significant investment from social network companies, and researchers. This had as a result, the development of many web platforms and mobile applications for reporting and monitoring online hate speech incidents. In this paper, we present the MANDOLA, a big data processing system that monitors, detects, visualizes and reports the spread and penetration of online hate-related speech using machine learning and big data approaches.
For a better quality of service, a source node often opts for a premium network path to send packets to a destination. However, the current Internet provides no assurance that packets follow the designated path. Network path validation enables on-path nodes to validate the path traversed by packets. We introduce path privacy and index privacy in the context of path validation. A path validation scheme is vulnerable to practical attacks if it lacks these properties. We design PrivNPV, a privacy-preserving network path validation protocol, that satisfies these properties. We discuss practicality of PrivNPV and how PrivNPV defends against various attacks.
Emergency communication networks are crucial for monitoring and providing assistance to affected people during long-persisting disasters. Given ubiquity of smartphones, we envision future emergency networks to seamlessly integrate the smartphones in the disaster area. The paper proposes such an architecture called Energy Aware Disaster Response Network using WiFi Tethering (E-DARWIN). We analyze the performance of our network using a combination of mathematical modeling, large scale simulations, and a prototype implementation on an Android platform.