AI coding agents are increasingly being adopted to autonomously author pull requests (PRs). While these agents can perform a wide range of tasks, little is known about how much human intervention is required to collaborate with them and integrate AI Agents contributions in practice. In this paper, we conduct an empirical study on human intervention in agent-authored PRs (APRs) as a measure of human effort and oversight. Using the AIDev dataset, we first compare how often humans intervene in APRs vs human-authored PRs (HPRs), as well as the resulting outcomes of this intervention. We then conduct a qualitative thematic analysis of human interventions in APRs and derive a taxonomy including 4 high-level categories and 42 intervention actions. Our results show that, human interventions occur less frequently in APRs than in HPRs (52.17% vs. 83.59%), but when it occurs in APRs, it requires higher review effort, including larger code churn and longer durations. Our taxonomy results show that most human effort is spent on guidance-level interventions with 58.02%, focusing on restricting the agent’s actions and enforcing project-conventions, followed by decision-level interventions at 21.16%, direct code changes at 17.05% and operational-level intervention at 3.69%. Which indicates that, collaboration with coding agents, is shifting developer work from implementation to supervision, guidance and quality control.

This research promotes innovative, dynamic learning scenarios by supporting gamified flipped learning through a model-driven authoring tool. It also incorporates Agile principles, namely Scrum, to promote flexibility and adaptability based on learners' diversity and feedback. Moreover, the approach integrates APIs for IoT devices (e.g., Raspberry Pi) and AI-powered APIs for real-time data analysis. Furthermore, to facilitate the upcoming enactment, it employs Atlas Transformation Language (ATL)-based model-to-model transformation to generate Business Process Management Notation (BPMN) 2.0 models from learning scenarios. It reduces the gap between pedagogical design and BPMN platforms. Preliminary results show positive educators' user feedback and an accurate, coherent BPMN model transformation

Meta-heuristic algorithms were effectively utilized in solving a wide array of problems in business, science, data mining, and finance. They are population-based approaches that develop novel optimal solutions based on the natural behaviors of creatures. A meta-heuristic method implied in reducing the number of present features and improving classification accuracy is the wrapper feature selection approach which was specifically proposed in this paper. This approach was based on the Capuchin Search Algorithm (CapSA) [1], which is mainly inspired by the functional behavior of Capuchin monkeys in food-searching actions through leaping, swinging, and climbing, all implemented in solving constrained and global optimization problems. This paper aims to enhance the CapSA algorithm to search in a binary search space rather than a continuous one. The proposed improvements on the algorithm can efficiently outperform other present algorithms with fewer selected features, less computational time, and high solution accuracy. The enhanced algorithm, named (EbCapSA) was tested afterwards on several biological benchmark datasets and then followed by a broad comparative study with four other algorithms, including the Gray Wolf Optimizer (GWO) [2], Particle Swarm Optimization (PSO) [3], the Firefly Algorithm (FA) [4], the Whale Optimization Algorithm (WOA) [5] and the original CapSA Algorithm. Overall results rendered a superiority of the proposed approach and clear outperformance based on the quality of the final solution in searching for the most favorable feature subsets.

In Technology-Enhanced-Learning (TEL) field, Learning Design (LD) research aims to support educators in assisting their learners in achieving the Intended Learning Objectives (ILOs) faster and more effectively. We focus on non-technical savvy educators that are aiming for efficient flipped learning scenarios, including Agile methodology, and providing a gamified learning environment to influence learners' engagement and motivation. However, to the best of our knowledge, no pedagogical standards support the description of this type of scenario, and no authoring tools assist in its creation. Therefore, in this paper, we introduce the implementation of (i) an authoring tool that is based on (ii) IMS LD extension approach. After the examination of a group of well-known authoring tools, we specify the requirements to which it should respond. Moreover, we present its evaluation according to LD experts’ criteria.

Human activity recognition has increasingly received attention in recent years to track regular activities of people. Existing activity recognition approaches considerably contributed to the analysis of human behavior. However, they still confront numerous issues related to the variability of activities performed by people within dynamic environments. Generally, this variability renders the used training or ontology models with predefined activities unsuitable. Therefore, creating an activity recognition approach that is able to leverage dynamically new and unknown activities at runtime becomes important. In this paper, we propose a novel knowledge-driven activity recognition framework using smartphone. This framework envisions taking a knowledge-driven approach to reinforce the recognition accuracy and people's quality of life in the context of dynamic environments at runtime. More specifically, we propose an ontology-based context evolution along with a dynamic decision-making, so that new and unknown performed activities can be accurately recognized. Furthermore, we use a public activity recognition dataset to demonstrate the effectiveness of the proposed framework and show its advantage over a data-driven baselines in terms of accuracy. Experimental results reveal that our framework not only reinforces the accuracy, but also enables an effective activity learning when facing unknown activities at runtime.

Image recognition and classification apps are considered to be one of the most popular apps in recent times due to its extremely important role in daily life. To improve the energy efficiency and performance of compute-demanding CNN, FPGA-based acceleration appears to be the best solution. In this article, we design and implement a hardware / software accelerator to efficiently accelerate the entificient and reusable FPGA-based accelerator that maximizes the FPGA compute capacity by exploiting the reorganization and parallelism of weights is proposed. The accelerator also supports computation of$3\times 3$convolutional layers and MLPs layers without interaction with the CPU. This accelerator is integrated into the tensorflow deep learning framework to provide software programmers with an easy-to-use interface so that they can declare a network definition while taking advantage of an FPGA engine. This system implemented on a Xilinx Zynq SoC using the PYNQ-Z1 platform achieves a frame rate equivalent to 5.91 fps using 16-bit fixed point and an energy efficiency of 186.25x the Intel® Xeon® processre CNN on FPGAs. First, we use tiling techniques to partition the input data. Second, we integrate the proposed accelerator into the tensorflow deep learning framework. We are evaluating the proposed hardware/software system and its integration with tensorflow by implementing the Deep Network In Network. The proposed accelerator achieves peak performance of 57.6 GOPS on a Xilinx PYNQ-Z1 FPGA board. End-to-end evaluation shows performance and power savings of up to 105.91x compared to an Intel® Xeon® CPU E5-2620 V4 under the working frequency of 200 MHz and a frame rate equivalent to 3.42 fps using 16-bit fixed point. Using the system with a high-end FPGA shows even higher gains and performance.

Nowadays, the rapidly changing Information and Communication Technologies bring new opportunities to the e-learning field. One of these opportunities is the use of LMS (Learning Management System) platforms. LMSs have to be able to interact with external applications and thus reconfigure their functionalities. In this context, the present paper aims at proposing an approach for integrating external tools into LMSs. In our approach, we suggest a solution based on Semantic Web Services (SWS) and Enterprise Service Bus (ESB)s. The main advantage of this solution is that it can be applied to any integration based e-Learning scenario.

Many research works had emphasized that successful learning processes can't be achieved without established methods and standards, consistent learner support or also collaborative activities. Applying intelligent features to simulate collaborative behavior could bring new challenging issues in e-learning processes. In this paper, we propose a Semantic web based approach which extends the Instructional Management Systems Learning Design information model to take into account intelligent virtual roles required by Collaborative Learning processes.

Many companies aim to use Web services to integrate heterogeneous and /or remote application in SOA (Service Oriented Architecture) contexts. The SaaS (Software as a Service) economical model allows to link service consumption and pricing. We aim to consider e learning as a set of services, hosted according to Cloud Computing techniques. We based our work on a concrete industrial product.

A few e-Learning platforms propose a solution for ubiquity and context aware adaptability. We decided to propose a solution with meta modeling approach based on previous related works. We used a concrete software engineering industrial product that was promoted by French Government.

Choose the best way to move from one place to another can involve different information: offers of different transport modes, their combination in the same journey and other information about services (such as restaurants, libraries, etc) that can be available in the route and useful for the passenger. Different approaches have been proposed to support the passenger's planning considering some part of this information. In this paper we present a public transportation domain ontology that considers different concepts related to the best and more relevant planning for the passenger. This ontology is formalized with OWL in Protégè tool. Using real instances and inferences, we show the ontology application, its relevance and consistency.

Research in e-learning has shifted in recent years its emphasis from almost purely technical aspects to pedagogy oriented interest. This shift is an indication of a positive trend which testifies to the preoccupations of teachers and educators. Nevertheless, and despite declarations of good intentions, more research effort should be made to attract more teachers towards the use of learning technologies. In our work, we are in favor of developing pedagogical approaches. However, since the field of learning technologies (LTs) has made significant progress in recent years (LOM, DC, EML, LD), we have to support this trend. This paper describes our approach and some of our findings out of several experiments that we have conducted. We also give the architectural design of an e-learning environment that is being deployed, whereby advances in LTs information models and tools such as LD and MOT+LD are utilized.