Search

The technology of military equipment entity extraction, a crucial component in constructing military knowledge bases, holds significant research value and theoretical importance for guiding the development and improvement of equipment support forces. In the military domain, equipment entities exhibit a phenomenon of nesting, where one entity is contained within another, and abbreviations or codes are frequently used to represent these entities. To address this complexity, this paper proposes a method named CoTNER for extracting entities. Initially, a large-scale language model is used to perform data augmentation with chain-of-thought on the original dataset, providing additional semantic and contextual information. Subsequently, the augmented dataset is fine-tuned on a small-scale language model to adapt it to the task of military equipment entity extraction and to enhance its ability to learn complex rules specific to the domain of military equipment. Additionally, a high-quality data filtering strategy based on instruction-following difficulty scoring is proposed to address the catastrophic forgetting issue that may occur during the fine-tuning of large language models. The experimental results demonstrate that the proposed military equipment entity extraction method outperforms mainstream traditional deep learning methods, validating the effectiveness of CoTNER. [ABSTRACT FROM AUTHOR]

In experiments with univalent target stimuli, task-switching costs can be eliminated if participants are unaware of the task rules and apply cue-target-response associations. However, in experiments with bivalent target stimuli, participants show task-switching costs. Participants may exhibit switch costs even when no task rules are provided in the instructions because they can infer the task rules. We tested this prediction by controlling the meaningfulness of cues and targets and therefore the ability to apply the task rules in 2 groups of participants. We compared the performance of Chinese and non-Chinese participants, who responded to Chinese numerals in an odd/even and high/low number task. In Experiment 1, Chinese participants, who knew Chinese characters and understood the task rules, showed task-switching costs. Non-Chinese participants on the other hand, who did not know Chinese characters, exhibited no switch costs. They applied a "target-first" strategy which means that they processed the target stimulus before the cue. In Experiment 2, we confirmed the absence of task-switching costs in Chinese participants using traditional Chinese numerals as target stimuli. Further, to determine how the target-first strategy affects switch costs, we manipulated the sequence of cue and target presentations. We conclude that task-switching costs can be eliminated more easily than previously thought, even for bivalent stimuli. The occurrence of task-switching costs depends on the approach used by participants and this may answer the puzzling question why humans typically do show task-switching costs whereas pigeons and monkeys do not.

Objective To study the effect of down-regulation of long non-coding RNA (lncRNA) HOXA cluster antisense RNA 2 (HOXA-AS2)/miR-20a pathway on apoptosis and invasion of skin squamous cell carcinoma cells in vitro. Methods qRT-PCR was used to detect the expression of HOXA-AS2 and miR-20a in 50 clinical specimens of skin squamous cell carcinoma and paired adjacent tissues. Skin squamous carcinoma cell line A431 was transfected with HOXA-AS2 siRNA, and the changes in cell proliferation, clonogenic ability, apoptosis, migration and invasion were evaluated by using MTT assay, clone formation experiment, flow cytometry, and Transwell assay respectively. Western blotting was performed to detect the expression of MMP-2, MMP-9, and cleaved caspase-3 proteins in the cells. Bioinformatics software predicted that HOXA-AS2 and miR-20a had complementary binding sites, and a luciferase reporter system was employed to verify their targeting relationship. A431 cells were co-transfected with HOXA-AS2 siRNA and miR-20a inhibitor, and the changes in cell proliferation, clone formation, apoptosis, migration and invasion were assessed. Results The expression level of HOXA-AS2 was increased in skin squamous cell carcinoma tissues. Transfection of A431 cells with HOXA-AS2 siRNA significantly suppressed the cell proliferation, clone formation, invasion and migration, promoted cell apoptosis, decreased the expression of MMP-2 and MMP-9 proteins, and increased the expression of cleaved caspase-3 protein. Down-regulation of HOXA-AS2 significantly up-regulated miR-20a expression. The expression level of miR-20a in skin squamous cell carcinoma tissues was negatively correlated with that of HOXA-AS2. In A431 cells, miR-20a inhibitor obviously reversed the effects of HOXA-AS2 siRNA on proliferation, clone formation, invasion, migration and apoptosis of the cells. Conclusion Down-regulation of HOXA-AS2/miR-20a pathway can promote the expression of miR-20a, suppress the proliferation, clone formation, invasion and migration, and induce apoptosis of skin squamous cell carcinoma cells in vitro.

Common ontology languages, construction principles and construction methods were summarized. It was considered that constructing coal mine safety ontology should adopt OWL-DL language and the construction method of cyclic acquisition method or seven-step method. Current research status of the coal mine safety ontology was analyzed from aspects of coal mine digging and coal mine disaster event ontology, coal mine safety database query ontology and coal mine safety knowledge concept ontology. Advantages, disadvantages, application scene and construction process of the three coal mine safety ontologies were summarized, and an ontology integration strategy of the above three coal mine safety ontologies was proposed.

Polymeric composites with the hybrid filler network for high-performance electromagnetic interference (EMI) shielding brought increasing interest. Herein, UHMWPE/ground tire rubber (GTR)/conductive carbon black (CCB) (UGC) composites with a hybrid segregated structural CCB-GTR were constructed by solid-phase shear milling (S3M), mechanical blending, and compacted molding processes. The percolation behavior of electrical conductivity and EMI shielding effectiveness (SE) of composites was described using Sigmoidal/Growth models, and its EMI shielding mechanism was discussed. Results showed that the EMI SE of U30G70C20, U50G50C20, and U70G30C20 composites were found to be 37.5, 41.7, and 43.8 dB, respectively, due to constructing a hybrid network. The attenuation rate of microwave radiation is above 99.99%, and the shielding mechanism is mainly electrical loss. Moreover, the relationship between electrical conductivity and EMI SE is consistent with Sigmoidal/Growth models, indicating that Sigmoidal/Growth models can successfully predict the percolation threshold of the EMI SE of UGC composites. [ABSTRACT FROM AUTHOR]