We study the problem of fair cohort selection from an unknown population, with a focus on university admissions. We start with the one-shot setting, where the admission policy must be fixed in advance and remain transparent, before observing the actual applicant pool. In contrast, the sequential setting allows the policy to be updated across stages as new applicant data becomes available. This is achieved by optimizing admission policies using a population model, trained on data from previous admission cycles. We also study the fairness properties of the resulting policies in the one-shot setting, including meritocracy and group parity.

Machine Learning software systems are frequently used in our day-to-day lives. Some of these systems are used in various sensitive environments to make life-changing decisions. Therefore, it is crucial to ensure that these AI/ML systems do not make any discriminatory decisions for any specific groups or populations. In that vein, different bias detection and mitigation open-source software libraries (aka API libraries) are being developed and used. In this paper, we conduct a qualitative study to understand in what scenarios these open-source fairness APIs are used in the wild, how they are used, and what challenges the developers of these APIs face while developing and adopting these libraries. We have analyzed 204 GitHub repositories (from a list of 1885 candidate repositories) which used 13 APIs that are developed to address bias in ML software. We found that these APIs are used for two primary purposes (i.e., learning and solving real-world problems), targeting 17 unique use-cases. Our study suggests that developers are not well-versed in bias detection and mitigation; they face lots of troubleshooting issues, and frequently ask for opinions and resources. Our findings can be instrumental for future bias-related software engineering research, and for guiding educators in developing more state-of-the-art curricula.

The five idioms (i.e., varieties) of the Romansh language are largely standardized and are taught in the schools of the respective communities in Switzerland. In this paper, we present the first parallel corpus of Romansh idioms. The corpus is based on 291 schoolbook volumes, which are comparable in content for the five idioms. We use automatic alignment methods to extract 207k multi-parallel segments from the books, with more than 2M tokens in total. A small-scale human evaluation confirms that the segments are highly parallel, making the dataset suitable for NLP applications such as machine translation between Romansh idioms. We release the parallel and unaligned versions of the dataset under a CC-BY-NC-SA license and demonstrate its utility for machine translation by training and evaluating an LLM on a sample of the dataset.

Water Distribution Networks (WDNs), critical to public well-being and economic stability, face challenges such as pipe blockages and background leakages, exacerbated by operational constraints such as data non-stationarity and limited labeled data. This paper proposes an unsupervised, online learning framework that aims to detect two types of faults in WDNs: pipe blockages, modeled as collective anomalies, and background leakages, modeled as concept drift. Our approach combines a Long Short-Term Memory Variational Autoencoder (LSTM-VAE) with a dual drift detection mechanism, enabling robust detection and adaptation under non-stationary conditions. Its lightweight, memory-efficient design enables real-time, edge-level monitoring. Experiments on two realistic WDNs show that the proposed approach consistently outperforms strong baselines in detecting anomalies and adapting to recurrent drift, demonstrating its effectiveness in unsupervised event detection for dynamic WDN environments.

Recently, Vision-Language-Action (VLA) models have demonstrated strong performance on a range of robotic tasks. These models rely on multimodal inputs, with language instructions playing a crucial role -- not only in predicting actions, but also in robustly interpreting user intent, even when the requests are impossible to fulfill. In this work, we investigate how VLAs can recognize, interpret, and respond to false-premise instructions: natural language commands that reference objects or conditions absent from the environment. We propose Instruct-Verify-and-Act (IVA), a unified framework that (i) detects when an instruction cannot be executed due to a false premise, (ii) engages in language-based clarification or correction, and (iii) grounds plausible alternatives in perception and action. Towards this end, we construct a large-scale instruction tuning setup with structured language prompts and train a VLA model capable of handling both accurate and erroneous requests. Our approach leverages a contextually augmented, semi-synthetic dataset containing paired positive and false-premise instructions, enabling robust detection and natural language correction. Our experiments show that IVA improves false premise detection accuracy by 97.56% over baselines, while increasing successful responses in false-premise scenarios by 50.78%.

Personalized recommendation is a key feature of intelligent tutoring systems, typically relying on accurate models of student knowledge. Knowledge Tracing (KT) models enable this by estimating a student's mastery based on their historical interactions. Many KT models rely on human-annotated knowledge concepts (KCs), which tag each exercise with one or more skills or concepts believed to be necessary for solving it. However, these KCs can be incomplete, error-prone, or overly general. In this paper, we propose a deep learning model that learns sparse binary representations of exercises, where each bit indicates the presence or absence of a latent concept. We refer to these representations as auxiliary KCs. These representations capture conceptual structure beyond human-defined annotations and are compatible with both classical models (e.g., BKT) and modern deep learning KT architectures. We demonstrate that incorporating auxiliary KCs improves both student modeling and adaptive exercise recommendation. For student modeling, we show that augmenting classical models like BKT with auxiliary KCs leads to improved predictive performance. For recommendation, we show that using auxiliary KCs enhances both reinforcement learning-based policies and a simple planning-based method (expectimax), resulting in measurable gains in student learning outcomes within a simulated student environment.

Usability describes a set of essential quality attributes of user interfaces (UI) that influence human-computer interaction. Common evaluation methods, such as usability testing and inspection, are effective but resource-intensive and require expert involvement. This makes them less accessible for smaller organizations. Recent advances in multimodal LLMs offer promising opportunities to automate usability evaluation processes partly by analyzing textual, visual, and structural aspects of software interfaces. To investigate this possibility, we formulate usability evaluation as a recommendation task, where multimodal LLMs rank usability issues by severity. We conducted an initial proof-of-concept study to compare LLM-generated usability improvement recommendations with usability expert assessments. Our findings indicate the potential of LLMs to enable faster and more cost-effective usability evaluation, which makes it a practical alternative in contexts with limited expert resources.

Multimodal Large Language Models (MLLMs) have demonstrated remarkable multimodal understanding capabilities in Visual Question Answering (VQA) tasks by integrating visual and textual features. However, under the challenging ten-choice question evaluation paradigm, existing methods still exhibit significant limitations when processing PDF documents with complex layouts and lengthy content. Notably, current mainstream models suffer from a strong bias toward English training data, resulting in suboptimal performance for Japanese and other language scenarios. To address these challenges, this paper proposes a novel Japanese PDF document understanding framework that combines multimodal hierarchical reasoning mechanisms with Colqwen-optimized retrieval methods, while innovatively introducing a semantic verification strategy through sub-question decomposition. Experimental results demonstrate that our framework not only significantly enhances the model's deep semantic parsing capability for complex documents, but also exhibits superior robustness in practical application scenarios.

Instruction tuning is vital for aligning large language models (LLMs) with human intent, but current methods typically rely on costly human-annotated seed data or powerful external teacher models. While instruction back-translation techniques reduce this dependency, they remain fundamentally tethered to an initial seed set, which limits full automation, introduces biases, and can lead to inefficient use of unlabeled corpora. In this paper, we propose Cycle-Instruct, a novel framework that achieves fully seed-free instruction tuning. Inspired by cycle consistency, Cycle-Instruct employs a dual self-training loop where two models-an answer generator and a question generator-are bootstrapped solely from raw, unlabeled text. These models mutually supervise each other by reconstructing original text segments from their counterpart's generated pseudo-labels, effectively learning from the intrinsic structure of the data without any human-provided seeds. We demonstrate Cycle-Instruct's efficacy across four diverse data tracks, including general instruction-following, domain-specific tasks, dialogue logs, and plain text. Our extensive experiments show that Cycle-Instruct not only outperforms seed-driven back-translation baselines but also achieves performance comparable to strongly supervised methods.

Ensuring liveability and comfort is one of the fundamental objectives of urban planning. Numerous studies have employed computational methods to assess and quantify factors related to urban comfort such as greenery coverage, thermal comfort, and walkability. However, a clear definition of urban comfort and its comprehensive evaluation framework remain elusive. Our research explores the theoretical interpretations and methodologies for assessing urban comfort within digital planning, emphasising three key dimensions: multidimensional analysis, data support, and AI assistance.