MuSeM: Detecting Incongruent News Headlines using Mutual Attentive Semantic Matching

Abstract

Measuring congruence between two texts has several useful applications, such as detecting the prevalent deceptive and misleading news headlines on the web. Many works have proposed machine learning based solutions such as text similarity between the headline and body text to detect the incongruence. Text similarity based methods fail to perform well due to different inherent challenges such as relative length mismatch between the news headline and its body content and non-overlapping vocabulary. On the other hand, more recent works that use headline guided attention to learn a headline derived contextual representation of the news body also result in convoluting overall representation due to the news body's lengthiness. This paper proposes a method that uses inter-mutual attention-based semantic matching between the original and synthetically generated headlines, which utilizes the difference between all pairs of word embeddings of words involved. The paper also investigates two more variations of our method, which use concatenation and dot-products of word embeddings of the words of original and synthetic headlines. We observe that the proposed method outperforms prior-arts significantly for two publicly available datasets.

Detecting Documents With Inconsistent Context

“…It seems similar to ours, but its document is made by permuting the sentence order within a document rather than another document. • congruence / incongruence: In [12], [18], [19], [37], [38], the goal is comparing relationships with different kinds of texts such as headline and body text in news articles. However, its document generating method is the same as our task.…”

“…For these reasons, we consider the incongruent news detection tasks with slight modifications to evaluate our model and compare to the proposed models. A. INCONGRUENT NEWS DETECTION TASKS [18], [19], [37], [38] performed incongruent news headline detection tasks and proposed their attention based models with auxiliary techniques.…”

“…Sentences from the headline and corresponding body can be directly compared as in [8], [33]. Attention based models have been proposed recently [12], [18], [19], [37], [38] showing promising performance for detecting incongruent headline news. However, detecting incongruent headline news is only one field of interest.…”

Learning to Detect Incongruence in News Headline and Body Text via a Graph Neural Network

“…An unsupervised learning technique was introduced to detect the stance of users in social media [11]. Most recently, a study proposed a method that detects headline incongruity via a semantic matching framework between the original and synthetically generated headlines [12].…”

Adversarial Training for Fake News Classification

“…Therefore, detecting incongruent news is vital to fight social media misinformation. Researchers have currently exploited different methods for detecting fake news, ranging from simple n-gram features based methods [4], hierarchical encoding based models [5], summarization based models [6] to artificially intelligent systems [7]- [9]. Normally, a system based on artificial intelligence encounters a bottleneck when optimization and tuning of different parameters [10] are essential.…”