As the popularity of wearables increases, so does their utility for studying emotions. Using new technologies points to several ethical challenges to be considered to improve research designs. There are several ethical recommendations for utilizing wearables to study human emotions, but they focus on emotion recognition systems applications rather than research design and implementation. To address this gap, we have developed a perspective on wearables, especially in daily life, adapting the ReCODE Health - Digital Health Framework and companion checklist. Therefore, our framework consists of four domains: (1) participation experience, (2) privacy, (3) data management, and (4) access and usability. We identified 33 primary risks of using wearables to study emotions, including research-related negative emotions, collecting, processing, storing, sharing personal and biological information, commercial technology validity and reliability, and exclusivity issues. We also proposed possible strategies for minimizing risks. We consulted the new ethical guidelines with members of ethics committees and relevant researchers. The judges (N= 26) positively rated our solutions and provided useful feedback that helped us refine our guidance. Finally, we summarized our proposals with a checklist for researchers' convenience. Our guidelines contribute to future research by providing improved protection of participants' and scientists' interests.

Ubiquitous sensing from wearable devices in the wild holds promise for enhancing human well-being, from diagnosing clinical conditions and measuring stress to building adaptive health promoting scaffolds. But the large volumes of data therein across heterogeneous contexts pose challenges for conventional supervised learning approaches. Representation Learning from biological signals is an emerging realm catalyzed by the recent advances in computational modeling and the abundance of publicly shared databases. The electrocardiogram (ECG) is the primary researched modality in this context, with applications in health monitoring, stress and affect estimation. Yet, most studies are limited by small-scale controlled data collection and over-parameterized architecture choices. We introduce WildECG, a pre-trained state-space model for representation learning from ECG signals. We train this model in a self-supervised manner with 275 000 10 s ECG recordings collected in the wild and evaluate it on a range of downstream tasks. The proposed model is a robust backbone for ECG analysis, providing competitive performance on most of the tasks considered, while demonstrating efficacy in low-resource regimes.

Kamil Kanclerz, Konrad Karanowski, Julita Bielaniewicz, Marcin Gruza, Piotr Miłkowski, Jan Kocon, Przemyslaw Kazienko. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. 2023.

Smart wearables, equipped with sensors monitoring physiological parameters, are becoming an integral part of our life. In this work, we investigate the possibility of utilizing such wearables to recognize emotions in the wild. In most reviewed papers, the authors apply a similar procedure consisting of participant recruitment, stimuli preparation and annotation, signal collection and processing, self-assessment, and machine learning model learning and validation. Besides, we identified seven emotion recognition scenarios and analyzed the transition from psychological models to machine learning tasks.

Recognizing emotions in everyday life requires a user-friendly and reliable system based on a smartphone and wearables. For over a year, we have been developing the Emognition system, which enables collecting emotionally annotated physiological signals in real-life scenarios. In this work, we describe the system architecture, the components and libraries used, as well as the development, testing, and implementation strategies. We explain in detail the integration with wearables - smartwatch Samsung Galaxy Watch 3 and chest strap Polar H10. The encountered problems and developed solutions are thoroughly discussed. We also provide the advantages and limitations of several frameworks for embedding machine learning models into a resource-restricted mobile application.

Kamil Kanclerz, Alicja Figas, Marcin Gruza, Tomasz Kajdanowicz, Jan Kocon, Daria Puchalska, Przemyslaw Kazienko. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). 2021.

Many tasks in natural language processing like offensive, toxic, or emotional text classification are subjective by nature. Humans tend to perceive textual content in their own individual way. Existing methods commonly rely on the agreed output values, the same for all consumers. Here, we propose personalized solutions to subjective tasks. Our four new deep learning models take into account not only the content but also the specificity of a given human. The models represent different approaches to learning the representation and processing data about text readers. The experiments were carried out on four datasets: Wikipedia discussion texts labelled with attack, aggression, and toxicity, as well as opinions annotated with ten numerical emotional categories. Emotional data was considered as multivariate regression (multitask), whereas Wikipedia data as independent classifications. All our models based on human biases and their representations significantly improve the prediction quality in subjective tasks evaluated from the individual’s perspective.

Analysis of subjective texts like offensive content or hate speech is a great challenge, especially regarding annotation process. Most of current annotation procedures are aimed at achieving a high level of agreement in order to generate a high quality reference source. However, the annotation guidelines for subjective content may restrict the annotators’ freedom of decision making . Motivated by a moderate annotation agreement in offensive content datasets, we hypothesize that personalized approaches to offensive content identification should be in place. Thus, we propose two novel perspectives of perception: group-based and individual. Using demographics of annotators as well as embeddings of their previous decisions (annotated texts), we are able to train multimodal models (including transformer-based) adjusted to personal or community profiles. Based on the agreement of individuals and groups, we experimentally showed that annotator group agreeability strongly correlates with offensive content recognition quality. The proposed personalized approaches enabled us to create models adaptable to personal user beliefs rather than to agreed offensiveness understanding. Overall, our individualized approaches to offensive content classification outperform classic data-centric methods that generalize offensiveness perception and it refers to all six tested models. Additionally, we developed requirements for annotation procedures, personalization and content processing to make the solutions human-centered.

Wearables equipped with pervasive sensors enable us to monitor physiological and behavioral signals in our everyday life. We propose the WellAff system able to recognize affective states for wellbeing support. It also includes health care scenarios, in particular patients with chronic kidney disease suffering from bipolar disorders. For the need of a large-scale field study, we revised over 50 off-the-shelf devices in terms of usefulness for emotion, stress, meditation, sleep, and physical activity recognition and analysis. Their usability directly comes from the types of sensors they possess as well as the quality and availability of raw signals. We found there is no versatile device suitable for all purposes. Using Empatica E4 and Samsung Galaxy Watch, we have recorded physiological signals from 11 participants over many weeks. The gathered data enabled us to train a classifier that accurately recognizes strong affective states.

Currently available artificial network generation models are characterized by consistency and low variance due to the rigidity of models' underlying assumptions. Networks generated from these models are usually too regular and do not contain noise and imbalance inherent in networks induced by human behavior. An important consequence is that much research on social network analysis presented in recent years used idealistic artificial networks that did not conform to reality. In order to alleviate this problem we introduce a new network generation model capable of modeling a broad spectrum of networks. In our model, the network formation process is not hard-coded into the model. Rather, we propose a simple mechanism for network creation based on priority ranking, and we encode the guiding principle of network formation as a distance function. By only changing the distance function definition and using the same priority ranking mechanism we are able to model very diverse networks. Our preliminary results show that we can mimic the behavior of popular artificial network generation models, such as the Erdös-Rényi random network model, the Watts-Strogatz small world model, or the Albert-Barabási preferential attachment model, but we can generate new types of networks as well. Following the principles of Open Science we publish the source code used to perform experiments and publish results in a public repository.

It is a common problem that cost of extracting data for network analysis could be very high. Also sometimes in the Internet is it hard to find graph with desired features such as node degree or clustering level. Because of that graph generators can than be very helpful. In the past bunch of models of such generators was developed: random graphs, small worlds and scale free networks. All of these generators were developed to quickly and efficiently create networks with desired parameters. However all of this models produce single layer graphs. Domain of multiplexes or multilayer graphs has not already been so deeply analysed, also because it is hard to collect multilayer data among real datasets or there is hard to define what kind of information layers exactly should represent. Proposed MuNeG --- Multilayer Network Generator can produce, based on set of input parameters, multiplex networks - networks where each node has its counterpart in each layer. The carried out experiments proved that MuNeG graphs have different network and social parameters depends on input values. This feature gives user a very handful tool to generate multiplex networks on purpose of social network or complex network analysis. Generator features, input parameters and their influence on so called graph theory measures such as: node degree, average shortest path, diameter or clustering are described in the following article.

It has been shown in this paper that simplistic Bag of Words (BoW) lexicon methods for sentiment polarity assignment with ensemble classifiers are much faster than a supervised approach to sentiment classification while yielding similar accuracy. BoW methods also proved to be efficient and fast across all examined datasets. Moreover, a new approach to lexicon extraction that can be successfully used for sentiment polarity assignment is presented in the paper. It has been shown that accuracy obtained from such lexicons outperforms other lexicon based approaches.

Methods of social network analysis are evolving into the practice of analyzing multi-layer networks, which represent different dimensions of activity and relationships. This also applies to methods oriented to detect communities in order to analyze the relationship between groups on different layers. In this context, the purpose of this article is to answer the question of how much the communities in cyberspace overlap when studying a variety of forms of interactions. The authors analyzed the structure of user groups formed on different layers of a social network created among users in a virtual world. By extracting groups and analyzing their structure and relationship with one another, some interesting properties are shown, especially in the area of the groups' overlapping. By performing this research, the authors were able to discover how similar the worlds we create can be when we are using different forms of communication in the same virtual environment.

More and more large data collections are gathered worldwide in various IT systems. Many of them possess a networked nature and need to be processed and analysed as graph structures. Due to their size they very often require the usage of a parallel paradigm for efficient computation. Three parallel techniques have been compared in the paper: MapReduce, its map-side join extension and Bulk Synchronous Parallel (BSP). They are implemented for two different graph problems: calculation of single source shortest paths (SSSP) and collective classification of graph nodes by means of relational influence propagation (RIP). The methods and algorithms are applied to several network datasets differing in size and structural profile, originating from three domains: telecommunication, multimedia and microblog. The results revealed that iterative graph processing with the BSP implementation always and significantly, even up to 10 times outperforms MapReduce, especially for algorithms with many iterations and sparse communication. The extension of MapReduce based on map-side join is usually characterized by better efficiency compared to its origin, although not as much as BSP. Nevertheless, MapReduce still remains a good alternative for enormous networks, whose data structures do not fit in local memories.

The relational label propagation problem for large data sets using MapReduce programming model was considered in the paper. The method we propose estimates class probability in relational domain in the networks. The method was examined on large real telecommunication data set. The results indicated that it could be used successfully to classify networks’ nodes and, thanks to that, new offerings or tariffs might be proposed to customers who belong to other providers. Moreover, basic properties of relational label propagation were examined and reported.

Predicting the future direction of community evolution is a problem with high theoretical and practical significance. It allows to determine which characteristics describing communities have importance from the point of view of their future behaviour. Knowledge about the probable future career of the community aids in the decision concerning investing in contact with members of a given community and carrying out actions to achieve a key position in it. It also allows to determine effective ways of forming opinions or to protect group participants against such activities. In the paper, a new approach to group identification and prediction of future events is presented together with the comparison to existing method. Performed experiments prove a high quality of prediction results. Comparison to previous studies shows that using many measures to describe the group profile, and in consequence as a classifier input, can improve predictions.

Diffusion of information in social networks takes more and more attention from marketers. New methods and algorithms are constantly developed towards maximizing reach of the campaigns and increasing their effectiveness. One of the important research directions in this area is related to selecting initial nodes of the campaign to result with maximizing its effects represented as total number of infections. To achieve this goal, several strategies were developed and they are based on different network measures and other characteristics of users. The problem is that most of these strategies base on static network properties while typical online networks change over time and are sensitive to varying activity of users. In this work a novel strategy is proposed which is based on multiple measures with additional parameters related to nodes availability in time periods prior to the campaign. Presented results show that it is possible to compensate users with high network measures by others having high frequency of system usage, which, instead, may be easier or cheaper to acquire.

In relational learning tasks such as within network classification the main problem arises from the inference of nodes' labels based on the the ground true labels of remaining nodes. The problem becomes even harder if the nodes from initial network do not have any labels assigned and they have to be acquired. However, labels of which nodes should be obtained in order to provide fair classification results? Active learning and inference is a practical framework to study this problem. The method for active learning and inference in within network classification based on node selection is proposed in the paper. Based on the structure of the network it is calculated the utility score for each node, the ranking is formulated and for selected nodes the labels are acquired. The paper examines several distinct proposals for utility scores and selection methods reporting their impact on collective classification results performed on various real-world networks.

The problem of automated video categorization in large datasets is considered in the paper. A new Iterative Multi-label Propagation (IMP) algorithm for relational learning in multi-label data is proposed. Based on the information of the already categorized videos and their relations to other videos, the system assigns suitable categories—multiple labels to the unknown videos. The MapReduce approach to the IMP algorithm described in the paper enables processing of large datasets in parallel computing. The experiments carried out on 5-million videos dataset revealed the good efficiency of the multi-label classification for videos categorization. They have additionally shown that classification of all unknown videos required only several parallel iterations.

The rapid development and expansion of the Internet and the social–based services comprised by the common Web 2.0 idea provokes the creation of the new area of research interests, i.e. social networks on the Internet called also virtual or online communities. Social networks can be either maintained and presented by social networking sites like MySpace , LinkedIn or indirectly extracted from the data about user interaction, activities or achievements such as emails, chats, blogs, homepages connected by hyperlinks, commented photos in multimedia sharing system, etc. A social network is the set of human beings or rather their digital representations that refer to the registered users who are linked by relationships extracted from the data about their activities, common communication or direct links gathered in the internet–based systems. Both digital representations named in the paper internet identities as well as their relationships can be characterized in many different ways. Such diversity yields for building a comprehensive and coherent view onto the concept of internet–based social networks. This survey provides in–depth analysis and classification of social networks existing on the Internet together with studies on selected examples of different virtual communities.

The paper addresses a problem of change identification in social group evolution. A new SGCI method for discovering of stable groups was proposed and compared with existing GED method. The experimental studies on a Polish blogosphere service revealed that both methods are able to identify similar evolution events even though both use different concepts. Some differences were demonstrated as well.

The link prediction problem in social networks defined as a task to predict whether a link between two particular nodes will appear in the future is still a broadly researched topic in the field of social network analysis. However, another relevant problem is solved in the paper instead of individual link forecasting: prediction of key network measures values, what is a more time saving approach. Two machine learning techniques were examined: time series forecasting and classification. Both of them were tested on two real-life social network datasets.

New technologies allow to store vast amount of data about users interaction. From those data the social network can be created. Additionally, because usually also time and dates of this activities are stored, the dynamic of such network can be analyzed by splitting it into many timeframes representing the state of the network during specific period of time. One of the most interesting issue is group evolution over time. To track group evolution the GED method can be used. However, choice of the timeframe type and length might have great influence on the method results. Therefore, in this paper, the influence of timeframe type as well as timeframe length on the GED method results is extensively analyzed.

Knowledge exchange between employees is an essential feature of recent commercial organisations on the competitive market. Based on the data gathered by various information technology (IT) systems, social links can be extracted and exploited in knowledge exchange systems of a new kind. Users of such a system ask their queries and the system recommends known and unknown experts selected out of user's friends. The friends either provide the solution or forward the query to their friends. By means of the established social paths to experts, the system facilitates informal learning and exchange of latent knowledge between organisation members in their workplace. The overall concept, limitations and detailed features of this novel knowledge exchange system are discussed in the article.

In recent world, we have been surrounded by various social networks (SNs). In every company, every institution and every place worldwide, people call each other, exchange emails, text messages, post in forums, co-author documents, meet at diverse events, etc. In other words, they communicate and collaborate with each other creating and maintaining mutual relationships in a complex SN. The traditional expert finding systems try to locate a relevant expert to whom the query should be sent. However, most of the experts are not willing to solve the problems for people they do not know. In our proposed novel system, a social paradigm is used to encourage experts to send their solutions. By means of recommendations, the system propagates the query not directly to the expert, but to friends and colleagues of the expert through the acquaintance chain existing in the SN. The experts are more likely to answer if the requests come from their acquaintances. The general idea, model and simulations on the recommender system for query propagation in the SN are presented in the paper.

Group extraction and their evolution are among the topics which arouse the greatest interest in the domain of social network analysis. However, while the grouping methods in social networks are developed very dynamically, the methods of group evolution discovery and analysis are still 'uncharted territory' on the social network analysis map. Therefore the new method for the group evolution discovery called GED is proposed in this paper. Additionally, the results of the first experiments on the email based social network together with comparison with two other methods of group evolution discovery are presented.

Multi-layered social networks consist of the fixed set of nodes linked by multiple connections. These connections may be derived from different types of user activities logged in the IT system. To calculate any structural measures for multi-layered networks this multitude of relations should be coped with in the parameterized way. Two separate algorithms for evaluation of shortest paths in the multi-layered social network are proposed in the paper. The first one is based on pre-processing -- aggregation of multiple links into single multi-layered edges, whereas in the second approach, many edges are processed 'on the fly' in the middle of path discovery. Experimental studies carried out on the DBLP database are presented as well.

Two approaches for identifying key persons in the blogosphere-based social network are analysed in the paper: discovery of the most important individuals either in persistent or in global social communities existing on web blogs. A new method for the separation of stable groups fulfilling given conditions is presented. Additionally, a new concept for extraction of user roles and key persons in such groups is proposed. It has been compared to the general clustering method and structural node position measure applied to rank users in the time-aggregated data. Experimental, comparative studies have been conducted on real blogosphere data gathered over one year.

All online sharing systems gather data that reflects users' collective behavior and their shared activities. This data can be used to extract different kinds of relationships which can be grouped into layers and which are basic components of the multidimensional social network (MSN) proposed in the paper. The layers are created on the basis of two types of relations between humans, i.e., direct and object-based ones which, respectively, correspond to either social or semantic links between individuals. For better understanding of the complexity of the social network structure, layers and their profiles were identified and studied on two, spanned in time, snapshots of the `Flickr' population. Additionally, for each layer, a separate strength measure was proposed. The experiments on the `Flickr' photo sharing system revealed that the relationships between users result either from semantic links between objects they operate on or from social connections of these users. Moreover, the density of the social network increases in time. The second part of this paper is devoted to building a social recommender system that supports the creation of new relations between users in a multimedia sharing system. Its main goal is to generate personalized suggestions that are continuously adapted to users' needs depending on the personal weights assigned to each layer in the MSN. The conducted experiments confirmed the usefulness of the proposed model.

Association rules are often utilized in Web recommendation systems for creation of suggested items lists. However, lists obtained in this way may be too short. Indirect association rules are introduced to extend classical, direct association rules and supplement the knowledge they contribute. The relation between ranking lists created on the basis of association rules and hyperlinks existing on Web pages has also been examined in the paper.