Remote user verification in Internet-based applications is becoming increasingly important nowadays. A popular scenario for it consists of submitting a picture of the user's Identity Document (ID) to a service platform, authenticating its veracity, and then granting access to the requested digital service. An ID is well-suited to verify the identity of an individual, since it is government issued, unique, and nontransferable. However, with recent advances in Artificial Intelligence (AI), attackers can surpass security measures in IDs and create very realistic physical and synthetic fake IDs. Researchers are now trying to develop methods to detect an ever-growing number of these AI-based fakes that are almost indistinguishable from authentic (bona fide) IDs. In this counterattack effort, researchers are faced with an important challenge: the difficulty in using real data to train fake ID detectors. This real data scarcity for research and development is originated by the sensitive nature of these documents, which are usually kept private by the ID owners (the users) and the ID Holders (e.g., government, police, bank, etc.). The main contributions of our study are: 1) We propose and discuss a patch-based methodology to preserve privacy in fake ID detection research. 2) We provide a new public database, FakeIDet2-db, comprising over 900K real/fake ID patches extracted from 2,000 ID images, acquired using different smartphone sensors, illumination and height conditions, etc. In addition, three physical attacks are considered: print, screen, and composite. 3) We present a new privacy-aware fake ID detection method, FakeIDet2. 4) We release a standard reproducible benchmark that considers physical and synthetic attacks from popular databases in the literature.

Large language models (LLMs) struggle with cross-lingual knowledge transfer: they hallucinate when asked in one language about facts expressed in a different language during training. This work introduces a controlled setting to study the causes and dynamics of this phenomenon by training small Transformer models from scratch on synthetic multilingual datasets. We identify a learning phase wherein a model develops either separate or unified representations of the same facts across languages, and show that unification is essential for cross-lingual transfer. We also show that the degree of unification depends on mutual information between facts and training data language, and on how easy it is to extract that language. Based on these insights, we develop methods to modulate the level of cross-lingual transfer by manipulating data distribution and tokenization, and we introduce metrics and visualizations to formally characterize their effects on unification. Our work shows how controlled settings can shed light on pre-training dynamics and suggests new directions for improving cross-lingual transfer in LLMs. This behavior has been attributed to training and sampling noise, gaps in pretraining data (Xu et al., 2024), and misaligned incentives in post-training (Schul-man, 2023). However, these fail to explain cross-lingual factual errors: cases where models accurately answer questions when posed in the same language as the training data, yet hallucinate when prompted in a different (often lower-resource) language (Goldman et al., 2025). Failures of cross-lingual transfer exacerbate disadvantages faced by speakers of underrepresented languages, and increasing model scale does not solve the problem (Aggarwal et al., 2025; Qi et al., 2023). LLMs have been found to develop both a lingua franca for factual knowledge (typically based on English) and distinct language silos (Aggarwal et al., 2025; Lim et al., 2025b; Schut et al., 2025; Lim et al., 2025a, inter alia), and their hidden representations can be language-agnostic or language-specific depending on the layer (Wang et al., 2025). However the root cause of these phenomena is not understood, as most research on cross-lingual transfer analyzes models as static artifacts. Such analysis, while valuable, cannot explain how knowledge arises during training, and therefore cannot lead to effective pre-training interventions. While some have investigated the training dynamics of knowledge acquisition in multilingual LLMs (Zeng et al., 2025; Liu et al., 2025), their approach is non-interventional and does not establish a causal link between data properties and cross-lingual transfer.

Machine translation quality has steadily improved over the years, achieving near-perfect translations in recent benchmarks. These high-quality outputs make it difficult to distinguish between state-of-the-art models and to identify areas for future improvement. In this context, automatically identifying texts where machine translation systems struggle holds promise for developing more discriminative evaluations and guiding future research. In this work, we address this gap by formalizing the task of translation difficulty estimation, defining a text's difficulty based on the expected quality of its translations. We introduce a new metric to evaluate difficulty estimators and use it to assess both baselines and novel approaches. Finally, we demonstrate the practical utility of difficulty estimators by using them to construct more challenging benchmarks for machine translation. Our results show that dedicated models outperform both heuristic-based methods and LLM-as-a-judge approaches, with Sentinel-src achieving the best performance. Thus, we release two improved models for difficulty estimation, Sentinel-src-24 and Sentinel-src-25, which can be used to scan large collections of texts and select those most likely to challenge contemporary machine translation systems.

From professional research to everyday planning, many tasks are bottlenecked by wide-scale information seeking, which is more repetitive than cognitively complex. With the rapid development of Large Language Models (LLMs), automated search agents powered by LLMs offer a promising solution to liberate humans from this tedious work. However, the capability of these agents to perform such "wide-context" collection reliably and completely remains largely unevaluated due to a lack of suitable benchmarks. To bridge this gap, we introduce WideSearch, a new benchmark engineered to evaluate agent reliability on these large-scale collection tasks. The benchmark features 200 manually curated questions (100 in English, 100 in Chinese) from over 15 diverse domains, grounded in real user queries. Each task requires agents to collect large-scale atomic information, which could be verified one by one objectively, and arrange it into a well-organized output. A rigorous five-stage quality control pipeline ensures the difficulty, completeness, and verifiability of the dataset. We benchmark over 10 state-of-the-art agentic search systems, including single-agent, multi-agent frameworks, and end-to-end commercial systems. Most systems achieve overall success rates near 0\%, with the best performer reaching just 5\%. However, given sufficient time, cross-validation by multiple human testers can achieve a near 100\% success rate. These results demonstrate that present search agents have critical deficiencies in large-scale information seeking, underscoring urgent areas for future research and development in agentic search. Our dataset, evaluation pipeline, and benchmark results have been publicly released at https://widesearch-seed.github.io/

Large Language Models (LLMs) are trained on a vast amount of procedural texts, but they do not directly observe real-world phenomena. In the context of cooking recipes, this poses a challenge, as intermediate states of ingredients are often omitted, making it difficult for models to track ingredient states and understand recipes accurately. In this paper, we apply state probing, a method for evaluating a language model's understanding of the world, to the domain of cooking. We propose a new task and dataset for evaluating how well LLMs can recognize intermediate ingredient states during cooking procedures. We first construct a new Japanese recipe dataset with clear and accurate annotations of ingredient state changes, collected from well-structured and controlled recipe texts. Using this dataset, we design three novel tasks to evaluate whether LLMs can track ingredient state transitions and identify ingredients present at intermediate steps. Our experiments with widely used LLMs, such as Llama3.1-70B and Qwen2.5-72B, show that learning ingredient state knowledge improves their understanding of cooking processes, achieving performance comparable to commercial LLMs. The dataset are publicly available at: https://huggingface.co/datasets/mashi6n/nhkrecipe-100-anno-1

Parkinson's Disease (PD) affects over 10 million people globally, with speech impairments often preceding motor symptoms by years, making speech a valuable modality for early, non-invasive detection. While recent deep-learning models achieve high accuracy, they typically lack the explainability required for clinical use. To address this, we propose RECA-PD, a novel, robust, and explainable cross-attention architecture that combines interpretable speech features with self-supervised representations. RECA-PD matches state-of-the-art performance in Speech-based PD detection while providing explanations that are more consistent and more clinically meaningful. Additionally, we demonstrate that performance degradation in certain speech tasks (e.g., monologue) can be mitigated by segmenting long recordings. Our findings indicate that performance and explainability are not necessarily mutually exclusive. Future work will enhance the usability of explanations for non-experts and explore severity estimation to increase the real-world clinical relevance.

Sound source localization (SSL) adds a spatial dimension to auditory perception, allowing a system to pinpoint the origin of speech, machinery noise, warning tones, or other acoustic events, capabilities that facilitate robot navigation, human-machine dialogue, and condition monitoring. While existing surveys provide valuable historical context, they typically address general audio applications and do not fully account for robotic constraints or the latest advancements in deep learning. This review addresses these gaps by offering a robotics-focused synthesis, emphasizing recent progress in deep learning methodologies. We start by reviewing classical methods such as Time Difference of Arrival (TDOA), beamforming, Steered-Response Power (SRP), and subspace analysis. Subsequently, we delve into modern machine learning (ML) and deep learning (DL) approaches, discussing traditional ML and neural networks (NNs), convolutional neural networks (CNNs), convolutional recurrent neural networks (CRNNs), and emerging attention-based architectures. The data and training strategy that are the two cornerstones of DL-based SSL are explored. Studies are further categorized by robot types and application domains to facilitate researchers in identifying relevant work for their specific contexts. Finally, we highlight the current challenges in SSL works in general, regarding environmental robustness, sound source multiplicity, and specific implementation constraints in robotics, as well as data and learning strategies in DL-based SSL. Also, we sketch promising directions to offer an actionable roadmap toward robust, adaptable, efficient, and explainable DL-based SSL for next-generation robots.

Large Language Models (LLMs) can achieve strong performance on everyday coding tasks, but they can fail on complex tasks that require non-trivial reasoning about program semantics. Finding training examples to teach LLMs to solve these tasks can be challenging. In this work, we explore a method to synthetically generate code reasoning training data based on a semantic inequivalence game SInQ: a generator agent creates program variants that are semantically distinct, derived from a dataset of real-world programming tasks, while an evaluator agent has to identify input examples that cause the original programs and the generated variants to diverge in their behaviour, with the agents training each other semi-adversarially. We prove that this setup enables theoretically unlimited improvement through self-play in the limit of infinite computational resources. We evaluated our approach on multiple code generation and understanding benchmarks, including cross-language vulnerability detection (Lu et al., 2021), where our method improves vulnerability detection in C/C++ code despite being trained exclusively on Python code, and the challenging Python builtin identifier swap benchmark (Miceli-Barone et al., 2023), showing that whereas modern LLMs still struggle with this benchmark, our approach yields substantial improvements. We release the code needed to replicate the experiments, as well as the generated synthetic data, which can be used to fine-tune LLMs.

DNA microarray technology enables the simultaneous measurement of expression levels of thousands of genes, thereby facilitating the understanding of the molecular mechanisms underlying complex diseases such as brain tumors and the identification of diagnostic genetic signatures. To derive meaningful biological insights from the high-dimensional and complex gene features obtained through this technology and to analyze gene properties in detail, classical AI-based approaches such as machine learning and deep learning are widely employed. However, these methods face various limitations in managing high-dimensional vector spaces and modeling the intricate relationships among genes. In particular, challenges such as hyperparameter tuning, computational costs, and high processing power requirements can hinder their efficiency. To overcome these limitations, quantum computing and quantum AI approaches are gaining increasing attention. Leveraging quantum properties such as superposition and entanglement, quantum methods enable more efficient parallel processing of high-dimensional data and offer faster and more effective solutions to problems that are computationally demanding for classical methods. In this study, a novel model called "Deep VQC" is proposed, based on the Variational Quantum Classifier approach. Developed using microarray data containing 54,676 gene features, the model successfully classified four different types of brain tumors-ependymoma, glioblastoma, medulloblastoma, and pilocytic astrocytoma-alongside healthy samples with high accuracy. Furthermore, compared to classical ML algorithms, our model demonstrated either superior or comparable classification performance. These results highlight the potential of quantum AI methods as an effective and promising approach for the analysis and classification of complex structures such as brain tumors based on gene expression features.

Large language models (LLMs) have demonstrated strong performance in sentence-level machine translation, but scaling to document-level translation remains challenging, particularly in modeling long-range dependencies and discourse phenomena across sentences and paragraphs. In this work, we propose a method to improve LLM-based long-document translation through targeted fine-tuning on high-quality document-level data, which we curate and introduce as DocBlocks. Our approach supports multiple translation paradigms, including direct document-to-document and chunk-level translation, by integrating instructions both with and without surrounding context. This enables models to better capture cross-sentence dependencies while maintaining strong sentence-level translation performance. Experimental results show that incorporating multiple translation paradigms improves document-level translation quality and inference speed compared to prompting and agent-based methods.