This innovative practice article reports on the piloting of vibe coding (using natural language to create software applications with AI) for English as a Foreign Language (EFL) education. We developed a human-AI meta-languaging framework with three dimensions: talking to AI (prompt engineering), talking through AI (negotiating authorship), and talking about AI (mental models of AI). Using backward design principles, we created a four-hour workshop where two students designed applications addressing authentic EFL writing challenges. We adopted a case study methodology, collecting data from worksheets and video recordings, think-aloud protocols, screen recordings, and AI-generated images. Contrasting cases showed one student successfully vibe coding a functional application cohering to her intended design, while another encountered technical difficulties with major gaps between intended design and actual functionality. Analysis reveals differences in students' prompt engineering approaches, suggesting different AI mental models and tensions in attributing authorship. We argue that AI functions as a beneficial languaging machine, and that differences in how students talk to, through, and about AI explain vibe coding outcome variations. Findings indicate that effective vibe coding instruction requires explicit meta-languaging scaffolding, teaching structured prompt engineering, facilitating critical authorship discussions, and developing vocabulary for articulating AI mental models.

The integration of Large Language Models (LLMs) into recommender systems has enabled zero-shot, personality-based personalization through prompt-based interactions, offering a new paradigm for user-centric recommendations. However, incorporating user personality traits via the OCEAN model highlights a critical tension between achieving psychological alignment and ensuring demographic fairness. To address this, we propose PerFairX, a unified evaluation framework designed to quantify the trade-offs between personalization and demographic equity in LLM-generated recommendations. Using neutral and personality-sensitive prompts across diverse user profiles, we benchmark two state-of-the-art LLMs, ChatGPT and DeepSeek, on movie (MovieLens 10M) and music (Last.fm 360K) datasets. Our results reveal that personality-aware prompting significantly improves alignment with individual traits but can exacerbate fairness disparities across demographic groups. Specifically, DeepSeek achieves stronger psychological fit but exhibits higher sensitivity to prompt variations, while ChatGPT delivers stable yet less personalized outputs. PerFairX provides a principled benchmark to guide the development of LLM-based recommender systems that are both equitable and psychologically informed, contributing to the creation of inclusive, user-centric AI applications in continual learning contexts.

Large Video Models (LVMs) build on the semantic capabilities of Large Language Models (LLMs) and vision modules by integrating temporal information to better understand dynamic video content. Despite their progress, LVMs are prone to hallucinations-producing inaccurate or irrelevant descriptions. Current benchmarks for video hallucination depend heavily on manual categorization of video content, neglecting the perception-based processes through which humans naturally interpret videos. We introduce MESH, a benchmark designed to evaluate hallucinations in LVMs systematically. MESH uses a Question-Answering framework with binary and multi-choice formats incorporating target and trap instances. It follows a bottom-up approach, evaluating basic objects, coarse-to-fine subject features, and subject-action pairs, aligning with human video understanding. We demonstrate that MESH offers an effective and comprehensive approach for identifying hallucinations in videos. Our evaluations show that while LVMs excel at recognizing basic objects and features, their susceptibility to hallucinations increases markedly when handling fine details or aligning multiple actions involving various subjects in longer videos.

The rapid advancement of large language models (LLMs) and domain-specific AI agents has greatly expanded the ecosystem of AI-powered services. User queries, however, are highly diverse and often span multiple domains and task types, resulting in a complex and heterogeneous landscape. This diversity presents a fundamental routing challenge: how to accurately direct each query to an appropriate execution unit while optimizing both performance and efficiency. To address this, we propose MoMA (Mixture of Models and Agents), a generalized routing framework that integrates both LLM and agent-based routing. Built upon a deep understanding of model and agent capabilities, MoMA effectively handles diverse queries through precise intent recognition and adaptive routing strategies, achieving an optimal balance between efficiency and cost. Specifically, we construct a detailed training dataset to profile the capabilities of various LLMs under different routing model structures, identifying the most suitable tasks for each LLM. During inference, queries are dynamically routed to the LLM with the best cost-performance efficiency. We also introduce an efficient agent selection strategy based on a context-aware state machine and dynamic masking. Experimental results demonstrate that the MoMA router offers superior cost-efficiency and scalability compared to existing approaches.

The performance of large language models (LLMs) is deeply influenced by the quality and composition of their training data. While much of the existing work has centered on English, there remains a gap in understanding how to construct effective training corpora for other languages. We explore scalable methods for building web-based corpora for LLMs. We apply them to build a new 120B token corpus in Portuguese that achieves competitive results to an industrial-grade corpus. Using a continual pretraining setup, we study how different data selection and preprocessing strategies affect LLM performance when transitioning a model originally trained in English to another language. Our findings demonstrate the value of language-specific filtering pipelines, including classifiers for education, science, technology, engineering, and mathematics (STEM), as well as toxic content. We show that adapting a model to the target language leads to performance improvements, reinforcing the importance of high-quality, language-specific data. While our case study focuses on Portuguese, our methods are applicable to other languages, offering insights for multilingual LLM development.

Abstract--Label concepts in multi-label data streams often experience drift in non-stationary environments, either independently or in relation to other labels. Transferring knowledge between related labels can accelerate adaptation, yet research on multi-label transfer learning for data streams remains limited. T o address this, we propose two novel transfer learning methods: BR-MARLENE leverages knowledge from different labels in both source and target streams for multi-label classification; BRPW-MARLENE builds on this by explicitly modelling and transferring pairwise label dependencies to enhance learning performance. Comprehensive experiments show that both methods outperform state-of-the-art multi-label stream approaches in non-stationary environments, demonstrating the effectiveness of inter-label knowledge transfer for improved predictive performance. Index T erms--Concept drift, non-stationary environment, multi-source, multi-label, class imbalance, transfer learning. Most research on data stream learning concentrates on streams with single labels [1]. However, many practical data streaming applications naturally adopt a multi-label paradigm, where each incoming data example has more than one label [2]. For example, a social media post could be tagged with several descriptors, or a movie might be classified under various predefined genres (e.g., Action, Crime, Historical), with each tag or genre representing a unique label.

This paper presents a personalized lecture concept using educational blocks and its demonstrative application in a new university lecture. Higher education faces daily challenges: deep and specialized knowledge is available from everywhere and accessible to almost everyone. University lecturers of specialized master courses confront the problem that their lectures are either too boring or too complex for the attending students. Additionally, curricula are changing more rapidly than they have in the past 10-30 years. The German education system comprises different educational forms, with universities providing less practical content. Consequently, many university students do not obtain the practical skills they should ideally gain through university lectures. Therefore, in this work, a new lecture concept is proposed based on the extension of the just-in-time teaching paradigm: Personalized and Demand-Based Education. This concept includes: 1) an initial assessment of students' backgrounds, 2) selecting the appropriate educational blocks, and 3) collecting ongoing feedback during the semester. The feedback was gathered via Pingo, ensuring anonymity for the students. Our concept was exemplarily tested in the new lecture "Practical Tools for Control Engineers" at the Karlsruhe Institute of Technology. The initial results indicate that our proposed concept could be beneficial in addressing the current challenges in higher education.

Large Language Models (LLMs), when paired with prompt-based tasks, have significantly reduced data annotation costs and reliance on human annotators. However, evaluating the quality of their annotations remains challenging in dynamic, unsupervised environments where oracle feedback is scarce and conventional methods fail. To address this challenge, we propose a novel agentic annotation paradigm, where a student model collaborates with a noisy teacher (the LLM) to assess and refine annotation quality without relying on oracle feedback. The student model, acting as an unsupervised feedback mechanism, employs a user preference-based majority voting strategy to evaluate the consistency of the LLM outputs. To systematically measure the reliability of LLM-generated annotations, we introduce the Consistent and Inconsistent (CAI) Ratio, a novel unsupervised evaluation metric. The CAI Ratio not only quantifies the annotation quality of the noisy teacher under limited user preferences but also plays a critical role in model selection, enabling the identification of robust LLMs in dynamic, unsupervised environments. Applied to ten open-domain NLP datasets across four LLMs, the CAI Ratio demonstrates a strong positive correlation with LLM accuracy, establishing it as an essential tool for unsupervised evaluation and model selection in real-world settings.

As humans delegate more tasks and decisions to artificial intelligence (AI), we risk losing control of our individual and collective futures. Relatively simple algorithmic systems already steer human decision-making, such as social media feed algorithms that lead people to unintentionally and absent-mindedly scroll through engagement-optimized content. In this paper, we develop the idea of human agency by integrating philosophical and scientific theories of agency with AI-assisted evaluation methods: using large language models (LLMs) to simulate and validate user queries and to evaluate AI responses. We develop HumanAgencyBench (HAB), a scalable and adaptive benchmark with six dimensions of human agency based on typical AI use cases. HAB measures the tendency of an AI assistant or agent to Ask Clarifying Questions, Avoid Value Manipulation, Correct Misinformation, Defer Important Decisions, Encourage Learning, and Maintain Social Boundaries. We find low-to-moderate agency support in contemporary LLM-based assistants and substantial variation across system developers and dimensions. For example, while Anthropic LLMs most support human agency overall, they are the least supportive LLMs in terms of Avoid Value Manipulation. Agency support does not appear to consistently result from increasing LLM capabilities or instruction-following behavior (e.g., RLHF), and we encourage a shift towards more robust safety and alignment targets.

Federated Learning (FL) enables decentralized model training without sharing raw data, offering strong privacy guarantees. However, existing FL protocols struggle to defend against Byzantine participants, maintain model utility under non-independent and identically distributed (non-IID) data, and remain lightweight for edge devices. Prior work either assumes trusted hardware, uses expensive cryptographic tools, or fails to address privacy and robustness simultaneously. We propose DSFL, a Dual-Server Byzantine-Resilient Federated Learning framework that addresses these limitations using a group-based secure aggregation approach. Unlike LSFL, which assumes non-colluding semi-honest servers, DSFL removes this dependency by revealing a key vulnerability: privacy leakage through client-server collusion. DSFL introduces three key innovations: (1) a dual-server secure aggregation protocol that protects updates without encryption or key exchange, (2) a group-wise credit-based filtering mechanism to isolate Byzantine clients based on deviation scores, and (3) a dynamic reward-penalty system for enforcing fair participation. DSFL is evaluated on MNIST, CIFAR-10, and CIFAR-100 under up to 30 percent Byzantine participants in both IID and non-IID settings. It consistently outperforms existing baselines, including LSFL, homomorphic encryption methods, and differential privacy approaches. For example, DSFL achieves 97.15 percent accuracy on CIFAR-10 and 68.60 percent on CIFAR-100, while FedAvg drops to 9.39 percent under similar threats. DSFL remains lightweight, requiring only 55.9 ms runtime and 1088 KB communication per round.