Intestinal cancer diagnosis electronic medical record attribute value extraction method based on multi-task learning
A multi-task learning, electronic medical record technology, applied in the field of attribute value extraction for colorectal cancer electronic medical records, colorectal cancer diagnosis electronic medical record attribute value extraction, can solve the problem of not considering the global word co-occurrence problem, to prevent overfitting Combine, improve the experimental effect, improve the effect of the experimental effect
Active Publication Date: 2020-09-15
DONGHUA UNIV
View PDF4 Cites 4 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
However, these models prioritize the information brought by the text order, and do not consider the word co-occurrence problem in the global context, and they will carry a large amount of long-distance information
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0046] Below in conjunction with specific embodiment, further illustrate the present invention. It should be understood that these examples are only used to illustrate the present invention and are not intended to limit the scope of the present invention. In addition, it should be understood that after reading the teachings of the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms also fall within the scope defined by the appended claims of the present application.
[0047] The framework as figure 2shown. The present invention uses an end-to-end neural network model to extract attribute values from multiple instances of text. First, use pre-trained word embeddings for each instance to better initialize the parameters in the neural network model. Second, it is fine-tuned using a domain corpus (training data) to capture domain-specific semantics / knowledge. Then, a BiLSTM layer is used ...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses an intestinal cancer diagnosis electronic medical record attribute value extraction method based on multi-task learning, and particularly relates to that an end-to-end neural network model extracts attribute values from a plurality of instances of a text. First, for each instance, pre-trained word embedding is used to better initialize parameters in a neural network model.Secondly, a domain corpus (training data) is used to finely tune it to capture semantics / knowledge of a particular domain. Then, a BiLSTM layer is used to consider the context information of the plurality of sentences to obtain a better sentence representation. Next, considering that not all sentences are useful for each attribute extractor, the intestinal cancer diagnosis electronic medical record attribute value extraction method uses an attention mechanism to select the most important instance for different attribute extractors, and correspondingly reduces noise brought by other instances.And finally, a multi-task learning mechanism is used in an output layer, and related tasks are learned together to solve multiple multi-class problem tasks at the same time, so that a better result isobtained, and the risk of over-fitting is reduced. And meanwhile, the loss contribution of each task is differentiated according to different importance.
Description
technical field [0001] The invention relates to a multi-task learning-based method for extracting attribute values from electronic medical records for intestinal cancer diagnosis, in particular to extracting attribute values from electronic medical records for intestinal cancer and completing structured tasks, belonging to the field of information technology. Background technique [0002] With the rapid development of information technology, it has brought technical support to the information construction of hospitals, and many hospitals have started the construction of hospital information system (hospital information system, HIS). Our country has a large population, and medical events of different sizes occur at each time point, resulting in massive medical data. Among them, electronic medical record (EMR) data contains a large number of patients' disease information and medical knowledge, which has attracted extensive attention of researchers. [0003] Electronic med...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Applications(China)
IPC IPC(8): G06F40/289G06F40/211G16H10/60G06N20/20G06N3/04
CPCG06F40/289G06F40/211G16H10/60G06N20/20G06N3/044G06N3/045Y02A90/10
Inventor 杜明周军锋徐波刘国华左彦飞庞敏敏张弘王文坤王璿
Owner DONGHUA UNIV