Pavlos Sermpezis

Sofia Yfantidou, Pavlos Sermpezis, Athena Vakali et al.

Personal informatics (PI) systems, powered by smartphones and wearables, enable people to lead healthier lifestyles by providing meaningful and actionable insights that break down barriers between users and their health information. Today, such systems are used by billions of users for monitoring not only physical activity and sleep but also vital signs and women's and heart health, among others. Despite their widespread usage, the processing of sensitive PI data may suffer from biases, which may entail practical and ethical implications. In this work, we present the first comprehensive empirical and analytical study of bias in PI systems, including biases in raw data and in the entire machine learning life cycle. We use the most detailed framework to date for exploring the different sources of bias and find that biases exist both in the data generation and the model learning and implementation streams. According to our results, the most affected minority groups are users with health issues, such as diabetes, joint issues, and hypertension, and female users, whose data biases are propagated or even amplified by learning models, while intersectional biases can also be observed.

Thanasis Troboukis, Kelly Kiki, Antonis Galanopoulos et al.

This chapter introduces a research project titled "Analyzing the Political Discourse: A Collaboration Between Humans and Artificial Intelligence", which was initiated in preparation for Greece's 2023 general elections. The project focused on the analysis of political leaders' campaign speeches, employing Artificial Intelligence (AI), in conjunction with an interdisciplinary team comprising journalists, a political scientist, and data scientists. The chapter delves into various aspects of political discourse analysis, including sentiment analysis, polarization, populism, topic detection, and Named Entities Recognition (NER). This experimental study investigates the capabilities of large language model (LLMs), and in particular OpenAI's ChatGPT, for analyzing political speech, evaluates its strengths and weaknesses, and highlights the essential role of human oversight in using AI in journalism projects and potentially other societal sectors. The project stands as an innovative example of human-AI collaboration (known also as "hybrid intelligence") within the realm of digital humanities, offering valuable insights for future initiatives.

Vasilis Andritsoudis, Pavlos Sermpezis, Ilias Dimitriadis et al.

The Internet Yellow Pages (IYP) aggregates information from multiple sources about Internet routing into a unified, graph-based knowledge base. However, querying it requires knowledge of the Cypher language and the exact IYP schema, thus limiting usability for non-experts. In this paper, we propose ChatIYP, a domain-specific Retrieval-Augmented Generation (RAG) system that enables users to query IYP through natural language questions. Our evaluation demonstrates solid performance on simple queries, as well as directions for improvement, and provides insights for selecting evaluation metrics that are better fit for IYP querying AI agents.

Sofia Vei, Paolo Giudici, Pavlos Sermpezis et al.

The absolute dominance of Artificial Intelligence (AI) introduces unprecedented societal harms and risks. Existing AI risk assessment models focus on internal compliance, often neglecting diverse stakeholder perspectives and real-world consequences. We propose a paradigm shift to a human-centric, harm-severity adaptive approach grounded in empirical incident data. We present AI Harmonics, which includes a novel AI harm assessment metric (AIH) that leverages ordinal severity data to capture relative impact without requiring precise numerical estimates. AI Harmonics combines a robust, generalized methodology with a data-driven, stakeholder-aware framework for exploring and prioritizing AI harms. Experiments on annotated incident data confirm that political and physical harms exhibit the highest concentration and thus warrant urgent mitigation: political harms erode public trust, while physical harms pose serious, even life-threatening risks, underscoring the real-world relevance of our approach. Finally, we demonstrate that AI Harmonics consistently identifies uneven harm distributions, enabling policymakers and organizations to target their mitigation efforts effectively.

Pavlos Sermpezis, Stelios Karamanidis, Eva Paraschou et al.

Political discourse datasets are important for gaining political insights, analyzing communication strategies or social science phenomena. Although numerous political discourse corpora exist, comprehensive, high-quality, annotated datasets are scarce. This is largely due to the substantial manual effort, multidisciplinarity, and expertise required for the nuanced annotation of rhetorical strategies and ideological contexts. In this paper, we present AgoraSpeech, a meticulously curated, high-quality dataset of 171 political speeches from six parties during the Greek national elections in 2023. The dataset includes annotations (per paragraph) for six natural language processing (NLP) tasks: text classification, topic identification, sentiment analysis, named entity recognition, polarization and populism detection. A two-step annotation was employed, starting with ChatGPT-generated annotations and followed by exhaustive human-in-the-loop validation. The dataset was initially used in a case study to provide insights during the pre-election period. However, it has general applicability by serving as a rich source of information for political and social scientists, journalists, or data scientists, while it can be used for benchmarking and fine-tuning NLP and large language models (LLMs).

Marinos Poiitis, Pavlos Sermpezis, Athena Vakali

Graph Representation Learning (GRL) has become essential for modern graph data mining and learning tasks. GRL aims to capture the graph's structural information and exploit it in combination with node and edge attributes to compute low-dimensional representations. While Graph Neural Networks (GNNs) have been used in state-of-the-art GRL architectures, they have been shown to suffer from over smoothing when many GNN layers need to be stacked. In a different GRL approach, spectral methods based on graph filtering have emerged addressing over smoothing; however, up to now, they employ traditional neural networks that cannot efficiently exploit the structure of graph data. Motivated by this, we propose PointSpectrum, a spectral method that incorporates a set equivariant network to account for a graph's structure. PointSpectrum enhances the efficiency and expressiveness of spectral methods, while it outperforms or competes with state-of-the-art GRL methods. Overall, PointSpectrum addresses over smoothing by employing a graph filter and captures a graph's structure through set equivariance, lying on the intersection of GNNs and spectral methods. Our findings are promising for the benefits and applicability of this architectural shift for spectral methods and GRL.

Sofia Yfantidou, Pavlos Sermpezis, Athena Vakali

In today's connected society, many people rely on mHealth and self-tracking (ST) technology to help them adopt healthier habits with a focus on breaking their sedentary lifestyle and staying fit. However, there is scarce evidence of such technological interventions' effectiveness, and there are no standardized methods to evaluate their impact on people's physical activity (PA) and health. This work aims to help ST practitioners and researchers by empowering them with systematic guidelines and a framework for designing and evaluating technological interventions to facilitate health behavior change (HBC) and user engagement (UE), focusing on increasing PA and decreasing sedentariness. To this end, we conduct a literature review of 129 papers between 2008 and 2022, which identifies the core ST HCI design methods and their efficacy, as well as the most comprehensive list to date of UE evaluation metrics for ST. Based on the review's findings, we propose PAST SELF, a framework to guide the design and evaluation of ST technology that has potential applications in industrial and scientific settings. Finally, to facilitate researchers and practitioners, we complement this paper with an open corpus and an online, adaptive exploration tool for the PAST SELF data.

Theodoros Giannakas, Pavlos Sermpezis, Anastasios Giovanidis et al.

As mobile traffic is dominated by content services (e.g., video), which typically use recommendation systems, the paradigm of network-friendly recommendations (NFR) has been proposed recently to boost the network performance by promoting content that can be efficiently delivered (e.g., cached at the edge). NFR increase the network performance, however, at the cost of being unfair towards certain contents when compared to the standard recommendations. This unfairness is a side effect of NFR that has not been studied in literature. Nevertheless, retaining fairness among contents is a key operational requirement for content providers. This paper is the first to study the fairness in NFR, and design fair-NFR. Specifically, we use a set of metrics that capture different notions of fairness, and study the unfairness created by existing NFR schemes. Our analysis reveals that NFR can be significantly unfair. We identify an inherent trade-off between the network gains achieved by NFR and the resulting unfairness, and derive bounds for this trade-off. We show that existing NFR schemes frequently operate far from the bounds, i.e., there is room for improvement. To this end, we formulate the design of Fair-NFR (i.e., NFR with fairness guarantees compared to the baseline recommendations) as a linear optimization problem. Our results show that the Fair-NFR can achieve high network gains (similar to non-fair-NFR) with little unfairness.

Savvas Kastanakis, Pavlos Sermpezis, Vasileios Kotronis et al.

Joint caching and recommendation has been recently proposed as a new paradigm for increasing the efficiency of mobile edge caching. Early findings demonstrate significant gains for the network performance. However, previous works evaluated the proposed schemes exclusively on simulation environments. Hence, it still remains uncertain whether the claimed benefits would change in real settings. In this paper, we propose a methodology that enables to evaluate joint network and recommendation schemes in real content services by only using publicly available information. We apply our methodology to the YouTube service, and conduct extensive measurements to investigate the potential performance gains. Our results show that significant gains can be achieved in practice; e.g., 8 to 10 times increase in the cache hit ratio from cache-aware recommendations. Finally, we build an experimental testbed and conduct experiments with real users; we make available our code and datasets to facilitate further research. To our best knowledge, this is the first realistic evaluation (over a real service, with real measurements and user experiments) of the joint caching and recommendations paradigm. Our findings provide experimental evidence for the feasibility and benefits of this paradigm, validate assumptions of previous works, and provide insights that can drive future research.

Dimitra Karanatsiou, Pavlos Sermpezis, Jon Gruda et al.

Users in Online Social Networks (OSN) leaves traces that reflect their personality characteristics. The study of these traces is important for a number of fields, such as a social science, psychology, OSN, marketing, and others. Despite a marked increase on research in personality prediction on based on online behavior the focus has been heavily on individual personality traits largely neglecting relational facets of personality. This study aims to address this gap by providing a prediction model for a holistic personality profiling in OSNs that included socio-relational traits (attachment orientations) in combination with standard personality traits. Specifically, we first designed a feature engineering methodology that extracts a wide range of features (accounting for behavior, language, and emotions) from OSN accounts of users. Then, we designed a machine learning model that predicts scores for the psychological traits of the users based on the extracted features. The proposed model architecture is inspired by characteristics embedded in psychological theory, i.e, utilizing interrelations among personality facets, and leads to increased accuracy in comparison with the state of the art approaches. To demonstrate the usefulness of this approach, we applied our model to two datasets, one of random OSN users and one of organizational leaders, and compared their psychological profiles. Our findings demonstrate that the two groups can be clearly separated by only using their psychological profiles, which opens a promising direction for future research on OSN user characterization and classification.

Pavlos Sermpezis, Savvas Kastanakis, João Ismael Pinheiro et al.

In this paper we propose that recommendation systems (RSs) for multimedia services should be "QoS-aware", i.e., take into account the expected QoS with which a content can be delivered, to increase the user satisfaction. Network-aware recommendations have been very recently proposed as a promising solution to improve network performance. However, the idea of QoS-aware RSs has been studied from the network perspective. Its feasibility and performance performance advantages for the content-provider or user perspective have only been speculated. Hence, in this paper we aim to provide initial answers for the feasibility of the concept of QoS-aware RS, by investigating its impact on real user experience. To this end, we conduct experiments with real users on a testbed, and present initial experimental results. Our analysis demonstrates the potential of the idea: QoS-aware RSs could be beneficial for both the users (better experience) and content providers (higher user engagement). Moreover, based on the collected dataset, we build statistical models to (i) predict the user experience as a function of QoS, relevance of recommendations (QoR) and user interest, and (ii) provide useful insights for the design of QoS-aware RSs. We believe that our study is an important first step towards QoS-aware recommendations, by providing experimental evidence for their feasibility and benefits, and can help open a future research direction.

Pavlos Sermpezis, Vasileios Kotronis, Alberto Dainotti et al.

BGP prefix hijacking is a threat to Internet operators and users. Several mechanisms or modifications to BGP that protect the Internet against it have been proposed. However, the reality is that most operators have not deployed them and are reluctant to do so in the near future. Instead, they rely on basic - and often inefficient - proactive defenses to reduce the impact of hijacking events, or on detection based on third party services and reactive approaches that might take up to several hours. In this work, we present the results of a survey we conducted among 75 network operators to study: (a) the operators' awareness of BGP prefix hijacking attacks, (b) presently used defenses (if any) against BGP prefix hijacking, (c) the willingness to adopt new defense mechanisms, and (d) reasons that may hinder the deployment of BGP prefix hijacking defenses. We expect the findings of this survey to increase the understanding of existing BGP hijacking defenses and the needs of network operators, as well as contribute towards designing new defense mechanisms that satisfy the requirements of the operators.