In multi-instance partial label learning (MIPL), each sample is a bag of multiple instances linked to a candidate label set containing one true and multiple false labels, yielding inexact supervision in both instance features and label space. However, existing works adopt decoupled approaches that focus exclusively on either instance-level feature fusion or label-level disambiguation, failing to fully exploit the intrinsic dependencies between these two spaces. To overcome this limitation, graph-based methods are widely recognized as a powerful paradigm in weakly supervised learning, yet their success hinges on reliable features—precisely what MIPL lacks due to instance-level noise. To bridge this gap, we propose DualG, a novel framework that simultaneously addresses feature learning and label disambiguation through dual-level graph propagation. Specifically, we construct dual relevance graphs at both the bag and instance levels. At the bag level, we build a similarity graph based on fused feature representations; at the instance level, we employ attention scores to filter out irrelevant instances and construct a reliable instance-level relevance graph. These complementary graphs enable our joint label disambiguation framework to simultaneously address inexact supervision signals in both instance space and label space. Experimental results on five benchmark datasets demonstrate that DualG outperforms existing MIPL and partial label learning methods, validating its effectiveness and superiority.

Cross-lingual named entity recognition (NER) aims to build an NER model that generalizes to the low-resource target language with labeled data from the high-resource source language.Current state-of-the-art methods typically combine self-training mechanism with contrastive learning paradigm, in order to develop discriminative entity clusters for cross-lingual adaptation.Despite the promise, we identify that these methods neglect two key problems: distribution skewness and pseudo-label bias, leading to indistinguishable entity clusters with small margins.To this end, we propose a novel framework, MARAL, which optimizes an adaptively reweighted contrastive loss to handle the class skewness and theoretically guarantees the optimal feature arrangement with maximum margin.To further mitigate the adverse effects of unreliable pseudo-labels, MARAL integrates a progressive cross-lingual adaptation strategy, which first selects reliable samples as anchors and then refines the remaining unreliable ones.Extensive experiments demonstrate that MARAL significantly outperforms the current state-of-the-art methods on multiple benchmarks, e.g., +2.

In the era of big data, data annotation is integral to numerous applications. However, it is widely acknowledged as a laborious and time-consuming process, significantly impeding the scalability and efficiency of data-driven applications. To reduce the human cost, we demonstrate CORAL, a collaborative automatic labeling system driven by large language models (LLMs), which achieves high-quality annotation with the least human effort. Firstly, CORAL employs LLM to automatically annotate vast datasets, generating coarse-grained labels. Subsequently, a weakly-supervised learning module trains small language models (SLMs) using noisy label learning techniques to distill accurate labels from LLM's annotations. It also allows statistical analysis of model outcomes to identify potentially erroneous labels, reducing the human cost of error detection. Furthermore, CORAL supports iterative refinement by LLMs and SLMs using manually corrected labels, thereby ensuring continual enhancement in annotation quality and model performance. A visual interface enables annotation process monitoring and result analysis.

In order to enhance the utility of online educational digital resources, the authors propose a practical and efficient multi-strategy relation extraction (RE) model in online education scenarios. First, the effective relation discrimination model is used to make relation predictions for non-structured teaching resources and eliminate the noise data. Then, they extract relations from different path strategies using multiple low-computational resources and efficient relation extraction strategies and use their proposed multi-strategy weighting calculator to weigh the relation extraction strategies to derive the final target relations. To cope with the low-resource relation extraction scenario, the relation extraction results are complemented by using prompt learning with a big model paradigm. They also consider the model to serve the commercial scenario of online education, and they propose a global rate controller to adjust and adapt the rate and throughput requirements in different scenarios, so as to achieve the best balance of system stability, computation speed, and extraction performance.