Emotion detection through physiological signals has become an essential aspect of affective computing and provides an objective way to capture human emotions. However, physiological data characterized by cross-subject variability and noisy labels hinder the performance of emotion recognition models. Existing domain adaptation and continual learning methods struggle to address these issues, especially under realistic conditions where data is continuously streamed and unlabeled. To overcome these limitations, we introduce a novel bi-level self-supervised continual learning framework, SSOCL, based on a dynamic memory buffer. This bi-level architecture iteratively refines the dynamic buffer and pseudo-label assignments to effectively retain representative samples, enabling generalization from continuous, unlabeled physiological data streams for emotion recognition. Then assigned pseudo-labels are subsequently leveraged for accurate emotion prediction. Key components of the framework, including a fast adaptation module and clusters mapping module, enable robust learning and effective handling of evolving data streams. Experimental validation on two mainstream electroencephalogram (EEG) datasets demonstrates the framework’s ability to adapt to continuous data streams while maintaining strong generalization across subjects, outperforming existing approaches.

Abstract Icons are the first visual element users encounter when searching for applications in online store. Icons with eye-catching features can make an app stand out in user searches, playing a crucial role in attracting user attention and influencing selection. This increases the likelihood of downloads, which can expand the user base, improve revenue, and enhance engagement, contributing to the application’s overall success. However, the majority of research focused on evaluating appeal of apps through application icons is empirical in nature and may lack comprehensive data analytical approaches. While empirical research holds its significance, it may still be limited by the size of the dataset analyzed and could also be subjective. This proposed research presents a novel data-analytical methodology to analyze a large dataset of application icons from Google Play to determine their influence on downloads. It clusters the icons using three different techniques: $K$-means clustering with two distinct feature vectors and agglomerative clustering, extracting various visual features from the clusters that are strongly correlated with application installs. Subsequently, validation of results has revealed that factors of varied colors, the dominance of white or black colors, text, and exposure in the icons can be linked to downloads.

Motion-based association for Multi-Object Tracking (MOT) has recently re-achieved prominence with the rise of powerful object detectors. Despite this, little work has been done to incorporate appearance cues beyond simple heuristic models that lack robustness to feature degradation. In this paper, we propose a novel way to leverage objects’ appearances to adaptively integrate appearance matching into existing high-performance motion-based methods. Building upon the pure motion-based method OC-SORT, we achieve 1st place on MOT20 and 2nd place on MOT17 with 63.9 and 64.9 HOTA, respectively. We also achieve 61.3 HOTA on the challenging DanceTrack benchmark as a new state-of-the-art even compared to more heavily-designed methods. The code and models are available at https://github.com/GerardMaggiolino/Deep-OC-SORT.

The problem of statistical heterogeneity in Federated Learning has been a major challenge, with existing solutions making unrealistic assumptions about the availability of shared datasets and high bandwidth between clients and the server. Solving this problem is crucial for the success of Federated Learning in real-world scenarios. In this work, we propose a biologically inspired dual-network model FedDual, which mimics how the human brain learns and memorizes the information. The model consists of a neocortical and a hippocampal network similar to those in the human brain. The hippocampal network is comprised by an image classification model, while the neocortical network is a variational auto-encoder responsible for long-term and re-callable memory. In this manner, FedDual uses the neocortical network to generate pseudo-patterns (synthetic data) on the server (global model). This allows for the hippocampal network to be trained with these pseudo-patterns. The dual-network architecture allows devices to share information via the weight updates of the neocortical network to the server without sending the actual data. We compare FedDual against alternatives elsewhere in the literature when applied to widely available datasets. FedDual not only achieves a margin of accuracy improvement over the alternatives, but also converges faster, requiring less communication rounds.

The health of vegetation could be affected by oil leakage through micro-seepage phenomena. In the present study, we used spectroscopic analysis in the VIS-NIR region to monitor and characterize plant's health affected by the hydrocarbon micro-seepage. For this purpose, we used same plant species subjected to different soil conditions with the same soil type. The spectral acquisition has been performed using SVC-HR768i spectroradiometer on every 15-day interval. The plant's physical appearance and red-edge position along with vegetation indices like NDVI, CARI, PRI and WDRVI were derived to observe the temporal change in plant growth. The results indicated red-edge position shifted towards shorter wavelength region and vegetation indices were showing stress in the plant subjected to the hydrocarbon mixed soil condition, possibly due to hydrocarbon micro-seepage. This study could be used in the application of remote sensing data to identify the hydrocarbon micro-seepage region, especially in highly vegetated land cover.

Gamification is the use of game elements in domains other than games. Gamification use is often suggested for difficult activities because it enhances users’ engagement and motivation level. Due to such benefits, the use of gamification is also proposed in education environments to improve students’ performance, engagement, and satisfaction. Computer science in higher education is a tough area of study and thus needs to utilize various already explored benefits of gamification. This research develops an empirical study to evaluate the effectiveness of gamification in teaching computer science in higher education. Along with the learning outcomes, the effect of group size on students’ satisfaction level is also measured. Furthermore, the impact of gamification over time is analyzed throughout a semester to observe its effectiveness as a long-term learning technique. The analysis, covering both learning outcome and students’ satisfaction, suggests that gamification is an effective tool to teach tough courses at higher education level; however, group size should be taken into account for optimal classroom size and better learning experience.

Smart grid is one of the most advanced technologies which plays a key role in maintaining balance between demand and supply by implementing demand response (DR). Residential users basically effect the overall performance of traditional grid due to maximum requirement of their energy demand. Home energy management (HEM) benefit the end user by monitoring, managing and controlling their energy consumption. Appliance scheduling is integral part of HEM as it manages energy demand according to supply by automatically controlling the appliances or by shifting the load from peak to off peak hours. Recently different techniques based on artificial intelligence (AI) are used to meet these objectives. In this research work, we evaluate the performance of HEM which is designed on the basis of heuristic algorithms, wind driven optimization (WDO), ganetic algorithm (GA) and binary particle swarm optimisation (BPSO). Finally, simulations are conducted in MATLAB to validate the performance of scheduling techniques in terms of cost, reduced peak to average ratio (PAR) and equally distributed energy consumption pattern. The simulation results prove that WDO algorithm based HEM proves to perform efficiently than BPSO and GA.

Online Social Networks (OSN) are platforms that let users build relationships by interacting with each other and adding objects. They differ from simple technical systems in having to satisfy social as well as technical requirements, so OSN access control is both more complex and more subtle than traditional. Currently, it is managed by local management of individual domains and local roles like friend. But making friend gives them rights, raising the issue of meta-rights, the right to issue a right. As user move from friend dyads to groups to communities, a systematic scheme to handle meta-rights (e.g. transferring, delegating, multiplying and dividing rights) is required. This paper outlines a general model to manage meta-rights for OSN in particular and socio-technical systems in general. The model's validity derives from socio-technical design, where social requirements like ownership and fairness give technical axioms.

Access control is the process by which access to information is granted to users for certain actions based on their identity. Traditional access control models that map every system resource directly to every system user work for organizations with thousands of users but struggle for social network sites like Facebook with millions of users. The problems faced are firstly the technical complexity of mapping millions of users to billions of resources and secondly the social need of users to own the items they post and to control their access, so access policies beyond just public/private are needed. And finally, that if ordinary users are to manage their own access control, they need software support. This paper argues that only distributed access control can meet these challenges and proposes a model based on the socio-technical design paradigm: first define the social requirements then design a technical solution to fulfill them.

Social networks are online platforms where users form relationships with others by sharing resources. Access control for these social networks is different from other systems as it fulfills the social requirements of community as well as the technical requirements of the system. This paper presents a classification of access control models for social networks based on lattice taxonomy where axes represent the properties of the models. The proposed taxonomy has eight axes representing: requestor identity, mapping authority, resource control, relationship management, credential distribution, access control decisions, rights delegation and transparency. Analysis of existing models using this taxonomy highlights the tradeoffs between user control, state distribution and social needs. The taxonomy reveals that various interesting features of social networks have not been implemented yet and there is a gap between the social requirements and access control features of social networks.

A novel video coding scheme is presented with superior performance against packet losses and hence is suitable for wireless communication compared to standard approaches like scalable video coding. The scheme also offers a better multi- rate video solution by offering multiple options for receiving videos at each sub-rate. The approach is suitable for use on multiple channels as well as on a single channel. The proposed hybrid scheme is innovative in the manner it combines the two standard styles of video coding, namely Scalable video coding and Multiple Description Coding, to provide a scalable bit- stream. The goal is to provide a better overall comprise between quality and resilience over different loss rates than is offered by either of the two separately as the congestion and hence the loss rate may vary with time. The scheme may offer a good potential to cater needs of emerging multimedia applications based on wireless communication and/or ubiquitous computing. Results of the proposed and existing schemes on different packet loss rates are reported in order to evaluate and compare their performances.