DOC: Deep Open Classification of Text Documents

By Lei Shu, Hu Xu, B. Liu. 2017

345 citations recorded by the source corpus.

Traditional supervised learning makes the closed-world assumption that the classes appeared in the test data must have appeared in training. This also applies to text learning or text classification. As learning is used increasingly in dynamic open environments where some new/test documents may not belong to any of the training classes, identifying these novel documents during classification presents an important problem. This problem is called open-world classification or open classification. This paper proposes a novel deep learning based approach. It outperforms existing state-of-the-art techniques dramatically.

Subfields: Artificial Intelligence, Natural Language Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Bing Liu
2. Han-Jia Ye
3. De-chuan Zhan
4. Caiming Xiong
5. Philip S. Yu
6. Richard Socher
7. Lei Shu
8. Hu Xu

In Natural Language Processing

1. Bing Liu
2. Andrea Madotto
3. Pascale Fung
4. Richard Socher
5. Caiming Xiong
6. Philip S. Yu
7. Hu Xu
8. Lei Shu

Citations in the corpus

• Simple and Principled Uncertainty Estimation with Deterministic Deep Learning via Distance Awareness (2020, methodology): “These are achieved by either replacing the loss function [33, 49, 50, 66, 67], modifying the output layer [6, 70, 11, 48], or computing a closed-form posterior for the last dense layer only [60, 68, 39].”
• Learning Placeholders for Open-Set Recognition (2021, methodology): “A similar work [28] replaced the softmax layer with one-vs-rest units.”
• Open-world Learning and Application to Product Classification (2018, methodology): “It has been shown in [32] that DOC signi€cantly outperforms the methods CL-cbsSVM and cbsSVM in [9] and OpenMax in [3].”
• Learning on the Job: Online Lifelong and Continual Learning (2020, background): “Steps 1 and 3 have been studied in open-world learning (Shu et al., 2017a, 2018; Chen and Liu, 2018) and continual learning (Chen and Liu, 2018; Parisi et al., 2018) respectively.”
• Transferable Multi-Domain State Generator for Task-Oriented Dialogue Systems (2019, background): “The applications within NLP has been comparatively limited, e.g., Shu et al. (2016, 2017b) for opinion mining, Shu et al. (2017a) for document classification, and Lee (2017) for hybrid code networks (Williams et al., 2017).”
• Open Category Detection with PAC Guarantees (2018, background): “There is also recent interest in open category detection for deep neural networks applied to vision and text classification (Bendale & Boult, 2016; Shu et al., 2017).”
• Classification-Reconstruction Learning for Open-Set Recognition (2018, methodology): “While DOC and Openmax performed almost on a par with each other, the improvement of CROSR over Openmax was also significant in this dataset.”
• A Unifying Review of Deep and Shallow Anomaly Detection (2020, background): “Currently, the standard evaluation practices in deep AD [130], [135], [137], [141], [144], [148], [152]–[155], [233], [531], OOD detection [241], [277], [490], [494]–[496], [532], [533], and open set recognition [240], [497]–[500] still extensively repurpose classification data sets by deeming some ”
• Hybrid Models for Open Set Recognition (2020, methodology): “Several methods [31,30] have also been proposed to apply open set models to text classification.”
• Conditional Gaussian Distribution Learning for Open Set Recognition (2020, methodology): “[33] proposed the Deep Open Classifier (DOC) model, which replaces the Softmax layer with a 1-vs-rest layer containing sigmoid functions.”
• CLEAR: Contrastive-Prototype Learning with Drift Estimation for Resource Constrained Stream Mining (2021, background): “Whenever the new classes come, the classifier architecture should be modified to adjust the amount of current total classes (for example, adding new neurons/layers for the new categories)[36, 57].”
• Zero-Shot Out-of-Distribution Detection Based on the Pre-trained Model CLIP (2021, methodology): “We adapted DOC’s code for images to generate its results.”
• Deep Learning for Anomaly Detection: A Survey (2019, neutral): “[2017],Yang et al. [2015],Chen et al. [2015],Gutoski et al.,D’Avino et al. [2017],Dotti et al. [2017],Yang et al. [2015],Chen et al. [2015],Sabokrou et al. [2016b],Tran and Hogg [2017],Chen et al.”
• Deep Unknown Intent Detection with Margin Loss (2019, methodology): “Shu et al. (2017) achieve the state-of-the-art performance by replacing the softmax layer of convolution neural network (CNN) with a 1-vs-rest layer consist of sigmoid and tightening the decision threshold of probability output for detection.”
• Unseen Class Discovery in Open-world Classification (2018, methodology): “It will also employs a latest algorithm for (1) (Shu et al., 2017) as (2) is after (1) and we want to test our proposed technique for (2) based on the rejected examples from (1).”
• Unknown Intent Detection Using Gaussian Mixture Model with an Application to Zero-shot Intent Classification (2020, methodology): “• DOC (Shu et al., 2017) builds m 1-vs-rest sigmoid classifiers for m seen classes respectively.”