Imbalanced data are commonly present in real-world applications. While data synthesis can effectively mitigate data scarcity for rare classes, and LLMs have revolutionized text generation, the application of LLMs to the synthesis of relational/structured tabular data remains underexplored. Moreover, existing approaches lack an effective feedback mechanism to guide LLMs in continuously optimizing the quality of the generated data throughout the synthesis process. In this work, we propose RDDG, Relational Data generator with Dynamic Guidance, which is a unified in-context learning framework that employs progressive chain-of-thought (CoT) steps to generate tabular data for enhancing downstream imbalanced classification performance. RDDG first uses core set selection to identify representative samples from the original data, then utilizes in-context learning to discover the inherent patterns and correlations among attributes within the core set, and subsequently generates tabular data while preserving the aforementioned constraints. More importantly, it incorporates a self-reinforcing feedback mechanism that provides automatic assessments of the quality of the generated data, enabling continuous quality optimization throughout the generation process. Experimental results on multiple real and synthetic datasets demonstrate that RDDG outperforms existing approaches in both data fidelity and downstream imbalanced classification performance.

Ancient manuscripts are the primary source of ancient linguistic corpora. However, some of them exhibit duplications due to unintentional repeated publication or deliberate forgery. Notable examples include counterfeit fragments within the Dead Sea Scrolls and the repeated publication of Oracle Bones (OB). Identifying such duplicates is critical for both archaeological curation and historical research. We present OBD-Finder, an explainable coarse-to-fine framework for large-scale duplicate discovery in ancient manuscript collections, with a primary focus on Oracle Bones. OBD-Finder integrates unsupervised low-level keypoint matching with high-level character-centric content similarity to progressively filter and rank duplicate candidates, balancing retrieval accuracy with computational efficiency. Designed as a practical, transparent, and domain-expert-friendly tool, the system produces candidate duplicates and supports expert inspection and validation. In real-world deployment, OBD-Finder has enabled domain experts to uncover 63 previously unknown duplicate Oracle Bone pairs that had remained overlooked for decades. Extensive experiments also confirm that OBD-Finder achieves competitive retrieval accuracy while offering substantially superior efficiency and interpretability compared to existing approaches. Code and deployed system are publicly available: https://github.com/cszhangLMU/OBD-Finder/.

Open-ended text generation faces a critical challenge: balancing coherence with diversity in LLM outputs. While contrastive search-based decoding strategies have emerged to address this trade-off, their practical utility is often limited by hyperparameter dependence and high computational costs. We introduce GUARD, a self-adaptive decoding method that effectively balances these competing objectives through a novel “Glocal” uncertainty-driven framework. GUARD combines global entropy estimates with local entropy deviations to integrate both long-term and short-term uncertainty signals. We demonstrate that our proposed global entropy formulation effectively mitigates abrupt variations in uncertainty, such as sudden overconfidence or high entropy spikes, and provides theoretical guarantees of unbiasedness and consistency. To reduce computational overhead, we incorporate a simple yet effective token-count-based penalty into GUARD. Experimental results demonstrate that GUARD achieves a good balance between text diversity and coherence, while exhibiting substantial improvements in generation speed. In a more nuanced comparison study across different dimensions of text quality, both human and LLM evaluators validated its remarkable performance. Our code is available at https://github.com/YecanLee/GUARD.