Current research on event extraction primarily utilizes intra-document information and overlooks inter-document relationships. Meanwhile, existing instance learning methods in event extraction overly rely on similarity-based retrieval, failing to capture event complexities and document semantics. To address these issues, we propose an event extraction model based on multi-instance learning, exploring the connections between documents through event types and event arguments. We design multiple instance selection strategies and construction methods to facilitate a comprehensive grasp of event semantics. Furthermore, we implemented a two-stage training approach to optimize the model’s ability to learn from instances obtained through various selection strategies. Experiments conducted on the RAMS and WIKIEVENTS datasets demonstrate that our method surpasses the current state-of-the-art models in certain metrics, validating its effectiveness and superiority.1

In this paper, we propose a document-level event argument extraction model called BSDM based on bidirectional span detection, mainly to address the problems of insufficient use of known information and overlooked correlations between span boundaries in document-level event argument extraction tasks. Firstly, in order to fully utilize known information, we design event types and event roles as event templates in the data preprocessing stage. Then, we integrate trigger and event type information into event roles in the information fusion layer, and identify argument boundaries under the guidance of role information. Secondly, in order to establish correlations between span boundaries, we design forward/backward decoders for the model, decoding argument spans from starting/ending boundaries respectively. Instead of using independent span boundary recognition methods, our model uses de-coded argument boundaries to recognize pending argument boundaries. Finally, considering the core role of trigger in events, we regard event detection as an auxiliary task for document-level event argument extraction. Through multitask joint training, the model will have better performance. We conducted sufficient experiments on the RAMS dataset. The results show that when using BERT-base as the encoder, BSDM achieved an F1 score 0.7% higher than the current SOTA result, and when using BERT-large as the encoder, it showed performance only second to the current SOTA model.

Early detection of lung cancer is crucial for five-year survival of patients. Compared with the pathological analysis and CT scans, the circulating tumor DNA (ctDNA) methylation based approach is noninvasive and cost-effective, and thus is one of the most promising methods for early detection of lung cancer. Existing studies on ctDNA methylation data measure the methylation level of each region with a predefined metric, ignoring the positions of methylated CpG sites and methylation patterns, thus are not able to capture the early cancer signals. In this paper, we propose a blood-based lung cancer detection method, and present the first ever study to represent methylation regions by continuous vectors. Specifically, we propose DeepMeth to regard each region as a one-channel image and develop an auto-encoder model to learn its representation. For each ctDNA methylation sample, DeepMeth achieves its representation via concatenating the region vectors. We evaluate DeepMeth on a multicenter clinical dataset collected from 14 hospitals. The experiments show that DeepMeth achieves about 5%-8% improvements compared with the baselines in terms of Area Under the Curve (AUC). Moreover, the experiments also demonstrate that DeepMeth can be combined with traditional scalar metrics to enhance the diagnostic power of ctDNA methylation classifiers. DeepMeth has been clinically deployed and applied to 450 patients from 94 hospitals nationally since April 2020.

In this paper, the problem of stabilization with optimal L1-gain for positive T-S fuzzy systems is investigated with the use of linear Lyapunov function. A T-S fuzzy model for positive nonlinear system is established to study the stabilization control for the positive system. Sufficient condition for stabilization is presented in term of linear programming. The static output-feedback fuzzy controller is constructed to guarantee that the closed-loop system is controlled positive, asymptotically stable and the L1-gains from the exogenous inputs to the regulated output is minimized, respectively. Moreover, the stabilization problem with optimal L∞-gain for positive T-S fuzzy systems is solved. Finally, three examples are presented to show the effectiveness of the theoretical results.