Semi-supervised deep learning based named entity recognition model to parse education section of resumes

Abstract

A job seeker’s resume contains several sections, including educational qualifications. Educational qualifications capture the knowledge and skills relevant to the job. Machine processing of the education sections of resumes has been a difficult task. In this paper, we attempt to identify educational institutions’ names and degrees from a resume’s education section. Usually, a significant amount of annotated data is required for neural network-based named entity recognition techniques. A semi-supervised approach is used to overcome the lack of large annotated data. We trained a deep neural network model on an initial (seed) set of resume education sections. This model is used to predict entities of unlabeled education sections and is rectified using a correction module. The education sections containing the rectified entities are augmented to the seed set. The updated seed set is used for retraining, leading to better accuracy than the previously trained model. This way, it can provide a high overall accuracy without the need of large annotated data. Our model has achieved an accuracy of 92.06% on the named entity recognition task.

Digital Document Analysis Using Weighted Score Convolutional Neural Network and Arc Factored Graph Based Dependency Parsing

“…Above figure shows the graphical representations of false positive rate using the proposed method, namely, WSCN-AGDP and existing methods namely Semi-supervised deep learning [1],…”

“…With vector, meta-analysis between rating progression for each sentence is analyzed that in turn retuned relevant feature extraction. With the relevant features extracted, parsing process was performed which in turn reduces the overhead incurred using WSCN-AGDP method by 16% compared to [1], 30% compared to [2] and 40% compared to [3].…”

“…Step 4: Formulate baseline embedding vector representation as in equation (1) Step 5: Estimate overall feature embedding representation for each word as in equation (2) Step 6: Estimate feature representation for each word as in equation (4)…”

“…Text mining and summarization methods utilize different types of mechanisms to classify and extract data from unstructured digital documents. A semi-supervised deep learning method has been proposed in [1] for parsing the educational section of the institutions' names and degrees of the corresponding resume. To be more specific, a proportion of data which is annotated is necessitated for named entity recognition methods based on neural network.…”

NeuronFair: Interpretable White-Box Fairness Testing through Biased Neuron Identification

“…Taking gender discrimination in salary prediction as an example, for two identical instances except for the gender attribute, male's annual income predicted by the DNN is often higher than female's [35]. Thus, it is of great importance for stakeholders to uncover fairness violations and then to reduce DNNs' discrimination so as to responsibly deploy fair and trustworthy deep learning systems in many sensitive scenarios [12,25,30,42].…”