This paper explores the use of large language models (LLMs) and prompt engineering to detect causal relationships in financial disclosures. The task was part of the FinCausal 2025 shared competition, which focuses on identifying cause-and-effect relationships in financial texts across languages. The study demonstrates the effectiveness of LLMs, specifically LLaMA 3.2, in tackling causality detection in English and Spanish financial reports. The paper introduces various prompt engineering techniques, including zero-shot, few-shot, and chain-of-thought (CoT) prompting, to improve performance. For English, the best results were achieved using the Few-Shot + CoT approach, while for Spanish, the Few-Shot method provided strong semantic alignment despite lower exact match accuracy. The evaluation used two metrics: Exact Match (EM) and Semantic Alignment Score (SAS). The results showed high SAS scores for both languages, indicating good semantic understanding, with English performing particularly well. The study emphasizes the importance of tailored prompt engineering techniques to handle language-specific nuances in financial contexts and suggests future research directions, including fine-tuning LLaMA 3.2 and testing additional LLM architectures to enhance multilingual causality detection in financial texts.

In this work, we propose Few Shot Domain Adapting Graph (FS-DAG), a scalable and efficient model architecture for visually rich document understanding (VRDU) in few-shot settings. FS-DAG leverages domain-specific and language/vision specific backbones within a modular framework to adapt to diverse document types with minimal data. The model is robust to practical challenges such as handling OCR errors, misspellings, and domain shifts, which are critical in real-world deployments. FS-DAG is highly performant with less than 90M parameters, making it well-suited for complex real-world applications for Information Extraction (IE) tasks where computational resources are limited. We demonstrate FS-DAG’s capability through extensive experiments for information extraction task, showing significant improvements in convergence speed and performance compared to state-of-the-art methods. Additionally, this work highlights the ongoing progress in developing smaller, more efficient models that do not compromise on performance.

This paper outlines the methodology for the automatic extraction of self-reported ages from social media posts as part of the Social Media Mining for Health (SMM4H) 2024 Workshop Shared Tasks. The focus was on Task 6: “Self-reported exact age classification with cross-platform evaluation in English.” The goal was to accurately identify age-related information from user-generated content, which is crucial for applications in public health monitoring, targeted advertising, and demographic research. A number of transformer-based models were employed, including RoBERTa-Base, BERT-Base, BiLSTM, and Flan T5 Base, leveraging their advanced capabilities in natural language understanding. The training strategies included fine-tuning foundational pre-trained language models and evaluating model performance using standard metrics: F1-score, Precision, and Recall. The experimental results demonstrated that the RoBERTa-Base model significantly outperformed the other models in this classification task. The best results achieved with the RoBERTa-Base model were an F1-score of 0.878, a Precision of 0.899, and a Recall of 0.858.