在生物医学文档中进行顺序句子分类的边界意识双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双
概括
一个新的边界意识的双双形模型通过明确建模语义结构和边界信息来改进生物医学文档中的句子角色标签,增强医疗决策.
科学领域:
- 生物医学自然语言处理
- 计算语言学 计算语言学
- 医疗信息学 医疗信息学
背景情况:
- 在生物医学文本中,准确的修辞角色分配 (例如背景,干预,结果) 有助于临床决策.
- 现有的基于序列和跨度的方法在捕捉文档语义结构和处理句子边界方面存在局限性.
研究的目的:
- 开发一种先进的模型,用于生物医学文档中精确的句子标签.
- 解决修辞角色分配中的语义连贯性和边界信息挑战.
主要方法:
- 提出了一个具有边界意识的双双关联模型,包含文档语义结构和边界信息.
- 用于显式语义建模的双双线层.
- 实现了一个动态编程算法,以实现最佳解码,并最大限度地减少预测错误.
主要成果:
- 具有边界意识的双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双双实现了PubMed 20k RCT和PubMed-PICO数据集上的最新性能.
- 与强大的基线方法相比,表现出优越的性能.
- 在NICTA-PIBOSO数据集上取得了竞争性结果.
结论:
- 拟议的模型有效地捕捉了文档语义结构和边界信息,以改善句子角色标签.
- 这种方法提高了医生检索证据的效率.
- 该模型代表了临床应用生物医学文本分析的重大进步.
相关概念视频
02:28Leaky Scanning
5.1K
During most eukaryotic translation processes, the small 40S ribosome subunit scans an mRNA from its 5' end until it encounters the first start AUG codon. The large 60S ribosomal subunit then joins the smaller one to initiate protein synthesis. The location of the translation initiation is largely determined by the nucleotides near the start codon as there may be multiple translation initiation sites present on the mRNA. Marilyn Kozak discovered that the sequence RCCAUGG (where R...
5.1K
02:02Cis-regulatory Sequences
9.9K
Cis-regulatory sequences are short fragments of non-coding DNA that are present on the same chromosomes as the genes that they regulate. These fragments serve as binding sites for transcriptional regulators, proteins that are responsible for controlling gene transcription and differential gene expression across cell types in eukaryotes. Cis-regulatory sequences can be close to the gene of interest or thousands of bases away in the DNA sequence; however, those sequences that are further away are...
9.9K