Multilingual understanding is crucial for the cross-cultural applicability of Large Language Models (LLMs). However, evaluation benchmarks designed for Hong Kong's unique linguistic landscape, which combines Traditional Chinese script with Cantonese as the spoken form and its cultural context, remain underdeveloped. To address this gap, we introduce HKMMLU, a multi-task language understanding benchmark that evaluates Hong Kong's linguistic competence and socio-cultural knowledge. The HKMMLU includes 26,698 multi-choice questions across 66 subjects, organized into four categories: Science, Technology, Engineering, and Mathematics (STEM), Social Sciences, Humanities, and Other. To evaluate the multilingual understanding ability of LLMs, 90,550 Mandarin-Cantonese translation tasks were additionally included. We conduct comprehensive experiments on GPT-4o, Claude 3.7 Sonnet, and 18 open-source LLMs of varying sizes on HKMMLU. The results show that the best-performing model, DeepSeek-V3, struggles to achieve an accuracy of 75\%, significantly lower than that of MMLU and CMMLU. This performance gap highlights the need to improve LLMs' capabilities in Hong Kong-specific language and knowledge domains. Furthermore, we investigate how question language, model size, prompting strategies, and question and reasoning token lengths affect model performance. We anticipate that HKMMLU will significantly advance the development of LLMs in multilingual and cross-cultural contexts, thereby enabling broader and more impactful applications.

Automatically grading the diverse range of question types in high school physics problem is a challenge that requires automated grading techniques from different fields. We report the findings of a Systematic Literature Review of potential physics grading techniques. We propose a multi-modal AI grading framework to address these challenges and examine our framework in light of Australia's AI Ethical Principles.

We address a fundamental challenge in Reinforcement Learning from Interaction Demonstration (RLID): demonstration noise and coverage limitations. While existing data collection approaches provide valuable interaction demonstrations, they often yield sparse, disconnected, and noisy trajectories that fail to capture the full spectrum of possible skill variations and transitions. Our key insight is that despite noisy and sparse demonstrations, there exist infinite physically feasible trajectories that naturally bridge between demonstrated skills or emerge from their neighboring states, forming a continuous space of possible skill variations and transitions. Building upon this insight, we present two data augmentation techniques: a Stitched Trajectory Graph (STG) that discovers potential transitions between demonstration skills, and a State Transition Field (STF) that establishes unique connections for arbitrary states within the demonstration neighborhood. To enable effective RLID with augmented data, we develop an Adaptive Trajectory Sampling (ATS) strategy for dynamic curriculum generation and a historical encoding mechanism for memory-dependent skill learning. Our approach enables robust skill acquisition that significantly generalizes beyond the reference demonstrations. Extensive experiments across diverse interaction tasks demonstrate substantial improvements over state-of-the-art methods in terms of convergence stability, generalization capability, and recovery robustness.

Evaluating the quality of slide-based multimedia instruction is challenging. Existing methods like manual assessment, reference-based metrics, and large language model evaluators face limitations in scalability, context capture, or bias. In this paper, we introduce LecEval, an automated metric grounded in Mayer's Cognitive Theory of Multimedia Learning, to evaluate multimodal knowledge acquisition in slide-based learning. LecEval assesses effectiveness using four rubrics: Content Relevance (CR), Expressive Clarity (EC), Logical Structure (LS), and Audience Engagement (AE). We curate a large-scale dataset of over 2,000 slides from more than 50 online course videos, annotated with fine-grained human ratings across these rubrics. A model trained on this dataset demonstrates superior accuracy and adaptability compared to existing metrics, bridging the gap between automated and human assessments. We release our dataset and toolkits at https://github.com/JoylimJY/LecEval.

Computer-assisted language learning -- CALL -- is an established research field. We review how artificial intelligence can be applied to support language learning and teaching. The need for intelligent agents that assist language learners and teachers is increasing: the human teacher's time is a scarce and costly resource, which does not scale with growing demand. Further factors contribute to the need for CALL: pandemics and increasing demand for distance learning, migration of large populations, the need for sustainable and affordable support for learning, etc. CALL systems are made up of many components that perform various functions, and AI is applied to many different aspects in CALL, corresponding to their own expansive research areas. Most of what we find in the research literature and in practical use are prototypes or partial implementations -- systems that perform some aspects of the overall desired functionality. Complete solutions -- most of them commercial -- are few, because they require massive resources. Recent advances in AI should result in improvements in CALL, yet there is a lack of surveys that focus on AI in the context of this research field. This paper aims to present a perspective on the AI methods that can be employed for language learning from a position of a developer of a CALL system. We also aim to connect work from different disciplines, to build bridges for interdisciplinary work.

Information on the web, such as scientific publications and Wikipedia, often surpasses users' reading level. To help address this, we used a self-refinement approach to develop a LLM capability for minimally lossy text simplification. To validate our approach, we conducted a randomized study involving 4563 participants and 31 texts spanning 6 broad subject areas: PubMed (biomedical scientific articles), biology, law, finance, literature/philosophy, and aerospace/computer science. Participants were randomized to viewing original or simplified texts in a subject area, and answered multiple-choice questions (MCQs) that tested their comprehension of the text. The participants were also asked to provide qualitative feedback such as task difficulty. Our results indicate that participants who read the simplified text answered more MCQs correctly than their counterparts who read the original text (3.9% absolute increase, p<0.05). This gain was most striking with PubMed (14.6%), while more moderate gains were observed for finance (5.5%), aerospace/computer science (3.8%) domains, and legal (3.5%). Notably, the results were robust to whether participants could refer back to the text while answering MCQs. The absolute accuracy decreased by up to ~9% for both original and simplified setups where participants could not refer back to the text, but the ~4% overall improvement persisted. Finally, participants' self-reported perceived ease based on a simplified NASA Task Load Index was greater for those who read the simplified text (absolute change on a 5-point scale 0.33, p<0.05). This randomized study, involving an order of magnitude more participants than prior works, demonstrates the potential of LLMs to make complex information easier to understand. Our work aims to enable a broader audience to better learn and make use of expert knowledge available on the web, improving information accessibility.

The increasing sophistication of AI-generated texts highlights the urgent need for accurate and transparent detection tools, especially in educational settings, where verifying authorship is essential. Existing literature has demonstrated that the application of stylometric features with machine learning classifiers can yield excellent results. Building on this foundation, this study proposes a comprehensive framework that integrates stylometric analysis with psycholinguistic theories, offering a clear and interpretable approach to distinguishing between AI-generated and human-written texts. This research specifically maps 31 distinct stylometric features to cognitive processes such as lexical retrieval, discourse planning, cognitive load management, and metacognitive self-monitoring. In doing so, it highlights the unique psycholinguistic patterns found in human writing. Through the intersection of computational linguistics and cognitive science, this framework contributes to the development of reliable tools aimed at preserving academic integrity in the era of generative AI.

In the rapidly evolving educational landscape, the unbiased assessment of soft skills is a significant challenge, particularly in higher education. This paper presents a fuzzy logic approach that employs a Granular Linguistic Model of Phenomena integrated with multimodal analysis to evaluate soft skills in undergraduate students. By leveraging computational perceptions, this approach enables a structured breakdown of complex soft skill expressions, capturing nuanced behaviours with high granularity and addressing their inherent uncertainties, thereby enhancing interpretability and reliability. Experiments were conducted with undergraduate students using a developed tool that assesses soft skills such as decision-making, communication, and creativity. This tool identifies and quantifies subtle aspects of human interaction, such as facial expressions and gesture recognition. The findings reveal that the framework effectively consolidates multiple data inputs to produce meaningful and consistent assessments of soft skills, showing that integrating multiple modalities into the evaluation process significantly improves the quality of soft skills scores, making the assessment work transparent and understandable to educational stakeholders.

Web-based educational videos offer flexible learning opportunities and are becoming increasingly popular. However, improving user engagement and knowledge retention remains a challenge. Automatically generated questions can activate learners and support their knowledge acquisition. Further, they can help teachers and learners assess their understanding. While large language and vision-language models have been employed in various tasks, their application to question generation for educational videos remains underexplored. In this paper, we investigate the capabilities of current vision-language models for generating learning-oriented questions for educational video content. We assess (1) out-of-the-box models' performance; (2) fine-tuning effects on content-specific question generation; (3) the impact of different video modalities on question quality; and (4) in a qualitative study, question relevance, answerability, and difficulty levels of generated questions. Our findings delineate the capabilities of current vision-language models, highlighting the need for fine-tuning and addressing challenges in question diversity and relevance. We identify requirements for future multimodal datasets and outline promising research directions.

-- The widespread adoption of Artificial Intelligence (AI) has been driven by significant advances in intelligent system research. However, this progress has raised concerns about data privacy, leading to a growing awareness of the need for privacy - preserving AI. In response, there has been a seismic shift in interest towards the leading paradigm for training Machine Learning (ML) models on decentralized data silos while maintaining data privacy, Federated Learning (FL). This research paper presents a comprehensive performance analysis of a cutting - edge approach to personalize ML model while preserving privacy achieved through Privacy Preserving Machine Learning with the innovative framework of Federated Personalized Learning (PPMLFPL). Regarding the increasing concerns about data privacy, this study evaluates the effectiveness of PPMLFPL addressing the critical balance between person - alized model refinement and maintaining the confidentiality of individual user data. According to our analysis, Adaptive Person - alized Cross - Silo Federated Learning with Differential Privacy (APPLE+DP) offering efficient execution whereas overall, the use of the Adaptive Personalized Cross - Silo Federated Learning with Homomorphic Encryption (APPLE+HE) algorithm for privacy - preserving machine learning tasks in federated personalized learning settings is strongly suggested. The results offer valuable insights creating it a promising scope for future advancements in the field of privacy - conscious data - driven technologies. Traditional ML models are often centralized, where all data is collected and stored in a single location for training. Privacy concerns in ML have been further exacerbated with the origin of Deep Learning (DL) models, which require even more data to achieve state - of - the - art performance.