Clickbait Detection in Tweets Using Self-attentive Network

TL;DR

This paper presents a self-attentive neural network model for clickbait detection in tweets, reformulating the problem as multi-class classification and achieving top performance in the 2017 Clickbait Challenge.

Contribution

It introduces a token-level self-attentive mechanism on biGRU hidden states for clickbait detection, eliminating manual feature engineering.

Findings

Ranked first in Clickbait Challenge 2017

Effective end-to-end training without manual features

Improved detection accuracy over previous methods

Abstract

Clickbait detection in tweets remains an elusive challenge. In this paper, we describe the solution for the Zingel Clickbait Detector at the Clickbait Challenge 2017, which is capable of evaluating each tweet's level of click baiting. We first reformat the regression problem as a multi-classification problem, based on the annotation scheme. To perform multi-classification, we apply a token-level, self-attentive mechanism on the hidden states of bi-directional Gated Recurrent Units (biGRU), which enables the model to generate tweets' task-specific vector representations by attending to important tokens. The self-attentive neural network can be trained end-to-end, without involving any manual feature engineering. Our detector ranked first in the final evaluation of Clickbait Challenge 2017.