This process involves assigning a unique identification code which is key to maintaining operations tracking and risk valuation at an optimal level. In the current global market context, an accurate identification of foreign entities also helps regulatory authorities to better monitor credit institutions' economic and financial activities, reinforcing national and international standards compliance as well as financial information transparency and integrity. Additionally, these unique identifications play a critical role in preventing fraud and money laundering by providing a standardized method for the identification of counterparties involved in financial transactions. These identifications are currently assigned through a labor-intensive entity-matching process which consists of receiving a daily list of foreign entities whose details (name, address, legal form...) are compared against the available source of reference (hereinafter referred to as ASR). ASR includes a series of datasets sourced from a wide range of different national and international databases such as Einforma (mainly Spain and Portugal), Companies House (UK) or Bundesanzeiger (Germany). If the information of the incoming record matches all attributes in the ASR, the identification will be approved and given a unique code (a new or an existing one). On the contrary, if there is no match (or a very poor matching) between the incoming data and the ASR, the incoming record will be rejected. Therefore, there is a permanent entity-matching challenge since small differences between incoming data and the ASR could easily lead to wrong conclusions, for example, considering two datasets as different entities when they are actually referring to the same one and vice versa.

Large language models (LLMs) have made significant strides in code translation tasks. However, ensuring both the correctness and readability of translated code remains a challenge, limiting their effective adoption in real-world software development. In this work, we propose F2STrans, a function-to-style guiding paradigm designed to progressively improve the performance of LLMs in code translation. Our approach comprises two key stages: (1) Functional learning, which optimizes translation correctness using high-quality source-target code pairs mined from online programming platforms, and (2) Style learning, which improves translation readability by incorporating both positive and negative style examples. Additionally, we introduce a novel code translation benchmark that includes up-to-date source code, extensive test cases, and manually annotated ground-truth translations, enabling comprehensive functional and stylistic evaluations. Experiments on both our new benchmark and existing datasets demonstrate that our approach significantly improves code translation performance. Notably, our approach enables Qwen-1.5B to outperform prompt-enhanced Qwen-32B and GPT-4 on average across 20 diverse code translation scenarios.

Log anomaly detection using traditional rule based or deep learning based methods is often challenging due to the large volume and highly complex nature of log sequence. So effective way of detection of anomalous sequence of logs is crucial for system maintenance and development. This paper proposes parameter efficient finetuning specifically low rank adaptation (LoRA) and adapter based approaches for finding contextual anomalies in sequence of logs in large log data set. It compares different tiny large language models (LLMs) on the Thunderbird dataset. The results show that LoRA based finetuning provides substantial performance improvements of 18 to 19 percentage over LogBert based full finetuning approach, achieving accuracy scores between 97.76% and 98.83% compared to 79.37%.

Timely identification and accurate risk stratification of cardiovascular disease (CVD) remain essential for reducing global mortality. While existing prediction models primarily leverage structured data, unstructured clinical notes contain valuable early indicators. This study introduces a novel LLM-augmented clinical NLP pipeline that employs domain-adapted large language models for symptom extraction, contextual reasoning, and correlation from free-text reports. Our approach integrates cardiovascular-specific fine-tuning, prompt-based inference, and entity-aware reasoning. Evaluations on MIMIC-III and CARDIO-NLP datasets demonstrate improved performance in precision, recall, F1-score, and AUROC, with high clinical relevance (kappa = 0.82) assessed by cardiologists. Challenges such as contextual hallucination, which occurs when plausible information contracts with provided source, and temporal ambiguity, which is related with models struggling with chronological ordering of events are addressed using prompt engineering and hybrid rule-based verification. This work underscores the potential of LLMs in clinical decision support systems (CDSS), advancing early warning systems and enhancing the translation of patient narratives into actionable risk assessments.

This paper presents HerO 2, Team HUMANE's system for the AVeriTeC shared task at the FEVER-25 workshop. HerO 2 is an enhanced version of HerO, the best-performing open-source model from the previous year's challenge. It improves evidence quality through document summarization and answer reformulation, optimizes veracity prediction via post-training quantization under computational constraints, and enhances overall system performance by integrating updated language model (LM) backbones. HerO 2 ranked second on the leaderboard while achieving the shortest runtime among the top three systems, demonstrating both high efficiency and strong potential for real-world fact verification. The code is available at https://github.com/ssu-humane/HerO2.

--Service robots are increasingly deployed in diverse and dynamic environments, where both physical layouts and social contexts change over time and across locations. In these unstructured settings, conventional navigation systems that rely on fixed parameters often fail to generalize across scenarios, resulting in degraded performance and reduced social acceptance. Although recent approaches have leveraged reinforcement learning to enhance traditional planners, these methods often fail in real-world deployments due to poor generalization and limited simulation diversity, which hampers effective sim-to-real transfer . T o tackle these issues, we present LE-Nav, an interpretable and scene-aware navigation framework that leverages multi-modal large language model reasoning and conditional variational autoencoders to adaptively tune planner hyperpa-rameters. T o achieve zero-shot scene understanding, we utilize one-shot exemplars and chain-of-thought prompting strategies. Experiments show that LE-Nav can generate hyperparameters achieving human-level tuning across diverse planners and scenarios. Real-world navigation trials and a user study on a smart wheelchair platform demonstrate that it outperforms state-of-the-art methods on quantitative metrics such as success rate, efficiency, safety, and comfort, while receiving higher subjective scores for perceived safety and social acceptance. Note to Practitioners--Service robots often experience degraded performance of traditional local planners due to changing and dynamic environmental conditions during navigation. This work investigates automatic hyperparameter tuning for planners such as DW A and TEB, and our framework LE-Nav can be used to adjust hyperparameters of any optimization-based planner . Existing navigation frameworks are typically either end-to-end, lacking safety guarantees, or rely on reinforcement learning-based tuning with limited generalization. By designing two prompting strategies, we enable the MLLM to generate stable and accurate scene descriptions. We use a conditional variational autoencoder to learn human expert tuning strategies, enhanced with data augmentation and attention masking to address inevitable MLLM packet loss in real applications. The decoupling of the MLLM and action modules improves decision transparency, allowing clear insight into how scene analysis informs navigation behavior . Experiments demonstrate that our method adaptively generates hyperparameters comparable to human experts, while being robust to packet loss and compatible with various MLLMs. Future work includes enhancing real-time scene understanding with advanced MLLMs, expanding support to more planners with personalized tuning, and extending the framework to collaborative multi-robot systems.

This paper presents our approach to EXIST 2025 Task 1, addressing text-based sexism detection in English and Spanish tweets through hierarchical Low-Rank Adaptation (LoRA) of Llama 3.1 8B. Our method introduces conditional adapter routing that explicitly models label dependencies across three hierarchically structured subtasks: binary sexism identification, source intention detection, and multilabel sexism categorization. Unlike conventional LoRA applications that target only attention layers, we apply adaptation to all linear transformations, enhancing the model's capacity to capture task-specific patterns. In contrast to complex data processing and ensemble approaches, we show that straightforward parameter-efficient fine-tuning achieves strong performance. We train separate LoRA adapters (rank=16, QLoRA 4-bit) for each subtask using unified multilingual training that leverages Llama 3.1's native bilingual capabilities. The method requires minimal preprocessing and uses standard supervised learning. Our multilingual training strategy eliminates the need for separate language-specific models, achieving 1.7-2.4\% F1 improvements through cross-lingual transfer. With only 1.67\% trainable parameters compared to full fine-tuning, our approach reduces training time by 75\% and model storage by 98\%, while achieving competitive performance across all subtasks (ICM-Hard: 0.6774 for binary classification, 0.4991 for intention detection, 0.6519 for multilabel categorization).

ABSTRACT Large language models, with their strong reasoning ability and rich knowledge, have brought revolution to many tasks of AI, but their impact on sign language generation remains limited due to its complexity and unique rules. In this paper, we propose TEAch Me Sign (TEAM-Sign), treating sign language as another natural language. By fine-tuning an LLM, we enable it to learn the correspondence between text and sign language, and facilitate generation. Considering the differences between sign and spoken language, we employ a stepwise prompting strategy to extract the inherent sign language knowledge within the LLM, thereby supporting the learning and generation process. Experimental results on How2Sign and Phoenix14T datasets demonstrate that our approach effectively leverages both the sign language knowledge and reasoning capabilities of LLM to align the different distribution and grammatical rules between sign and spoken language.

Prior human-robot interaction (HRI) research has primarily focused on single-user interactions, where robots do not need to consider the timing or recipient of their responses. However, in multi-party interactions, such as at malls and hospitals, social robots must understand the context and decide both when and to whom they should respond. In this paper, we propose a Transformer-based multi-task learning framework to improve the decision-making process of social robots, particularly in multi-user environments. Considering the characteristics of HRI, we propose two novel loss functions: one that enforces constraints on active speakers to improve scene modeling, and another that guides response selection towards utterances specifically directed at the robot. Additionally, we construct a novel multi-party HRI dataset that captures real-world complexities, such as gaze misalignment. Experimental results demonstrate that our model achieves state-of-the-art performance in respond decisions, outperforming existing heuristic-based and single-task approaches. Our findings contribute to the development of socially intelligent social robots capable of engaging in natural and context-aware multi-party interactions.

In this paper, we explore how large language models (LLMs) approach financial decision - making by systematically comparing their responses to those of human participants across the globe. We posed a set of commonly used financial decision - making questions t o seven leading LLMs, including five models from the GPT series (GPT - 4o, GPT - 4.5, o1, o3 - mini), Gemini 2.0 Flash, and DeepSeek R1 . We then compared their outputs to human responses drawn from a dataset covering 53 nations. Our analysis reveals three main r esults. First, LLMs generally exhibit a risk - neutral decision - making pattern, favoring choices aligned with expected value calculations when faced with lottery - type questions . Second, when evaluating trade - offs between present and future, LLMs occasionally produce responses that appear inconsistent with normative reasoning . Third, when we examine cross - national similarities, we f ind that the LLMs' aggregate responses most closely resemble those of participants from Tanzania. These findings contribute to the understanding of how LLMs emulate human - like decision behaviors and highlight potential cultural and training influences embedded within their outputs.