Shuo Yang, Caren Han, Xueqi Ma, Yan Li, Mohammad Reza Ghasemi Madani, Eduard Hovy. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Mental health challenges are increasingly prevalent among university students, yet often go undetected due to reliance on traditional assessments that are subjective, infrequent, and lack behavioral context. Digital phenotyping through passively collected smartphone data offers a scalable alternative, but existing approaches often fail to integrate predictive accuracy with narrative-based insights. To overcome these limitations, we present NarrativeSense, a novel framework that combines machine learning models with narrative-based descriptions of daily life events inferred from smartphone sensing data to predict weekly affective states. The system incorporates language model components to transform behavioral patterns into contextualized, human-readable narratives that ground affective predictions in everyday experiences. This narrative layer complements structured prediction by offering intuitive, user-centered insights. Applied to longitudinal data from 58 university students over 119 days, NarrativeSense outperforms baseline machine learning models, standalone LLMs, and ensemble methods, while providing richer insights. Our findings demonstrate the potential of narrative-enhanced digital phenotyping for scalable and explainable mental health monitoring in educational and clinical settings.

Transformer-based Large Language Models (LLMs) struggle with inputs exceeding their training context window due to positional outof-distribution (O.O.D.) issues that disrupt attention.Existing solutions, including finetuning and training-free methods, face challenges like inefficiency, redundant interpolation, logit outliers, or loss of local positional information.We propose Greedy Attention Logit Interpolation (GALI), a training-free method that improves length extrapolation by greedily reusing pretrained positional intervals and interpolating attention logit to eliminate outliers.GALI achieves stable and superior performance across a wide range of long-context tasks without requiring input-length-specific tuning.Our analysis further reveals that LLMs interpret positional intervals unevenly and that restricting interpolation to narrower ranges improves performance, even on short-context tasks.GALI represents a step toward more robust and generalizable long-text processing in LLMs.Our implementation of GALI, along with the experiments from our paper, is open-sourced at https://github.com/adlnlp/Gali.

Visually Rich Document Understanding (VRDU) has emerged as a critical field in document intelligence, enabling automated extraction of key information from complex documents across domains such as medical, financial, and educational applications. However, form-like documents pose unique challenges due to their complex layouts, multi-stakeholder involvement, and high structural variability. Addressing these issues, the VRD-IU Competition was introduced, focusing on extracting and localizing key information from multi-format forms within the Form-NLU dataset, which includes digital, printed, and handwritten documents. This paper presents insights from the competition, which featured two tracks: Track A, emphasizing entity-based key information retrieval, and Track B, targeting end-to-end key information localization from raw document images. With over 20 participating teams, the competition showcased various state-of-the-art methodologies, including hierarchical decomposition, transformer-based retrieval, multimodal feature fusion, and advanced object detection techniques. The top-performing models set new benchmarks in VRDU, providing valuable insights into document intelligence.