The structure of a document contains rich information such as logical relations in context, hierarchy, affiliation, dependence, and applicability. It will greatly affect the accuracy of document information processing, particularly of legal documents and business contracts. Therefore, intelligent document structural analysis is important to information extraction and data mining. However, unlike the well-studied field of text semantic analysis, current work in document structural analysis is still scarce. In this paper, we propose an intelligent document structural analysis framework through data pre-processing, feature engineering, and structural classification with a dynamic sample weighting algorithm. As a typical application, we collect more than 11,000 insurance document content samples and carry out the machine learning experiments to check the efficiency of our framework. Meanwhile, to address the sample imbalance problem in the hierarchy classification task, a dynamic sample weighting algorithm is incorporated into our Dynamic Weighting Structural Analysis (DWSA) framework, in which the weights of different category tags according to the structural levels are iterated dynamically in training. Our results show that the DWSA has significantly improved the comprehensive accuracy and the classification F1-score of each category. The comprehensive accuracy is as high as 94.68% (3.36% absolute improvement) and the Macro F1-score is 88.29% (5.1% absolute improvement).

Document classification is an important area in Natural Language Processing (NLP). Because a huge amount of scientific papers have been published at an accelerating rate, it is beneficial to carry out intelligent paper classifications, especially fine-grained classification for researchers. However, a public scientific paper dataset for fine-grained classification is still lacking, so the existing document classification methods have not been put to the test. To fill this vacancy, we designed and collected the PaperNet-Dataset that consists of multi-modal data (texts and figures). PaperNet 1.0 version contains hierarchical categories of papers in the fields of computer vision (CV) and NLP, 2 coarse-grained and 20 fine-grained (7 in CV and 13 in NLP). We ran current mainstream models on the PaperNet-Dataset, along with a multi-modal method that we propose. Interestingly, none of these methods reaches an accuracy of 80% in fine-grained classification, showing plenty of room for improvement. We hope that PaperNet-Dataset will inspire more work in this challenging area.

Multimodal sarcasm detection (MSD) aims to identify sarcastic expressions by integrating multimodal and contextual cues to capture cross-modal semantic inconsistencies. However, existing studies face challenges: fine-grained sarcastic cues are implicit and dispersed, large generative models suffer from gradient vanishing in classification tasks, and cross-domain generalization remains limited. To address these limitations, we propose InterARM, an interpretable affective reasoning model that introduces a structured sarcasm reasoning paradigm. Specifically, we introduce a three-stage training strategy based on curriculum learning, consisting of 1) sarcasm classification learning, 2) structured sarcasm reasoning learning, and 3) step-selective hierarchical reward reinforcement learning. The model generates reasoning chains across modalities to analyze sentiment polarity and semantic conflicts. We also construct the MSD-CoT dataset with 3,200 image-text pairs with 19,200 human-annotated reasoning steps. Experiments show that InterARM achieves superior performance on the in-domain and out-of-domain datasets, outperforming larger MLLMs such as Qwen2.5-VL-7B and GPT-5-mini while maintaining high interpretability and strong cross-domain generalization.