Event classification method based on binary hyperspherical prototype network
A classification method and prototype network technology, applied in reasoning methods, computer components, instruments, etc., can solve problems such as low-resource event extraction effect is not good
Active Publication Date: 2020-11-24
ZHEJIANG UNIV
View PDF6 Cites 4 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
The event extraction methods of these two technical solutions are not effective for low-resource event extraction.
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
experiment example
[0052] During the training process, the SGD optimizer was used for optimization, and 30,000 training iterations and 2,000 testing iterations were performed. The dimension of the word vector is set to 50, and the dimension of a single position vector is set to 10, so the entire position vector is 30 dimensions. The hyperparameters σ are set to 500, λ to 1, and α to 0.5. In order to avoid overfitting, a dropout ratio of 0.2 is set. The learning rate for model training is set to 0.001. The performance of the model is measured by the accuracy rate, recall rate and F1 value of event extraction.
[0053] The effect of event extraction on the causal event extraction dataset is shown in the following table:
[0054] Model Accuracy recall rate F1 value DMCNN 0.7033 0.7156 0.7156 JRNN 0.7156 0.6831 0.7088 JMEE 0.7491 0.7034 0.7418 Ours 0.7889 0.7438 0.7732 Ours (+25%) 0.7421 0.7132 0.7399 Ours (+50%) 0.7605 0.7204 0.7539...
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 event classification method based on a binary hyperspherical prototype network, and the method comprises the steps: obtaining an instance, and obtaining an instance representation through the coding of an instance representation model; constructing prototype representation of the event; constructing classification loss according to the distance difference between the hyperspherical representation of the instance and the prototype representation of the event to which the instance belongs, constructing inference loss representing a causal relationship according to thedistances from the prototype representations of the two events to the dielectric layer, and constructing total loss according to the classification loss and the inference loss; according to the totalloss optimization model parameters, obtaining an instance representation model with determined parameters and optimized prototype representation; and after obtaining the instance representation of thenew instance according to the instance representation model, calculating the similarity between the hyperspherical representation corresponding to the instance representation and all the optimized prototype representations, and selecting the event category of the prototype representation corresponding to the highest similarity as the event category of the new instance. Event classification is carried out by carrying out fusion reasoning through priori knowledge among the events, so that the event classification accuracy is improved.
Description
technical field [0001] The invention belongs to the technical field of information extraction and reasoning, and in particular relates to an event classification method based on a bipartite hypersphere prototype network. Background technique [0002] Traditional event extraction models are always based on supervised learning, thus requiring sufficient training samples, but it is often difficult to obtain a large number of labeled samples in the real world. Moreover, the effect of event extraction weakens as the number of samples decreases. However, most current models assume that sufficient training samples are an indispensable condition for learning event representations, which makes it difficult to achieve ideal results in event extraction tasks. Therefore, it is extremely important to allow the model to extract events with low resources. Traditional low-resource event extraction models are mainly based on supervised learning, migration learning / pre-training, or meta-lea...
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 IPC IPC(8): G06F30/27G06K9/62G06N5/04
CPCG06F30/27G06N5/04G06F18/241G06F18/25
Inventor 陈华钧邓淑敏张宁豫
Owner ZHEJIANG UNIV