Neural Temporality Adaptation for Document Classification: Diachronic Word Embeddings and Domain Adaptation Models (2019-07-01T00:00:00.000000Z)

TL;DR

Two complementary ways to adapt classifiers to shifts across time are described, and it is shown that diachronic word embeddings, which were originally developed to study language change, can also improve document classification.

Abstract

Language usage can change across periods of time, but document classifiers models are usually trained and tested on corpora spanning multiple years without considering temporal variations. This paper describes two complementary ways to adapt classifiers to shifts across time. First, we show that diachronic word embeddings, which were originally developed to study language change, can also improve document classification, and we show a simple method for constructing this type of embedding. Second, we propose a time-driven neural classification model inspired by methods for domain adaptation. Experiments on six corpora show how these methods can make classifiers more robust over time.

Authors

Michael J. Paul

2 papers

Xiaolei Huang

1 papers

TL;DR

Abstract

Authors

References39 items

Examining Temporality in Document Classification

Time-evolving Text Classification with Deep Neural Networks

Diachronic word embeddings and semantic shifts: a survey

Deep Neural Models of Semantic Shift

Dynamic Embeddings for Language Evolution

Outta Control: Laws of Semantic Change and Inherent Biases in Word Representation Models

Temporal Word Analogies: Identifying Lexical Replacement with Diachronic Word Embeddings

Wasserstein Distance Guided Representation Learning for Domain Adaptation

Learning Word Relatedness over Time

Dynamic Word Embeddings for Evolving Semantic Discovery

Dynamic Word Embeddings

Analysing the Semantic Change Based on Word Embedding

Enriching Word Vectors with Subword Information

Hierarchical Attention Networks for Document Classification

Diachronic Word Embeddings Reveal Statistical Laws of Semantic Change

Ups and Downs: Modeling the Visual Evolution of Fashion Trends with One-Class Collaborative Filtering

Domain-Adversarial Training of Neural Networks

Recurrent Convolutional Neural Networks for Text Classification

Adam: A Method for Stochastic Optimization

Statistically Significant Detection of Linguistic Change

Convolutional Neural Networks for Sentence Classification

Explicit factor models for explainable recommendation based on phrase-level sentiment analysis

Temporal Analysis of Language through Neural Language Models

Distributed Representations of Words and Phrases and their Compositionality

An Empirical Investigation of Statistical Significance in NLP

Word Epoch Disambiguation: Finding How Words Change Over Time

Scikit-learn: Machine Learning in Python

Data for Everyone

Frustratingly Easy Domain Adaptation

NLTK: The Natural Language Toolkit

Digital selection and analogue amplification coexist in a cortex-inspired silicon circuit

Long Short-Term Memory

Calculation of the Wasserstein Distance Between Probability Distributions on the Line

Meituan-dianping official website

Chollet et al

Keras

Dropout: a simple way to prevent neural networks from overfitting

Lecture 6.5-rmsprop: Divide the gradient by a running average of its recent magnitude

NLTK: The Natural Language Toolkit

Field of Study

Venue Information

Name

Type

URL

Alternate Names