2019 · 14 citations · 16 references
ComputingComputer ScienceComputational LinguisticsMachine LearningNatural Language ProcessingLanguage ModelingMisinformation DetectionOpinion MiningSemantic NetworksWord EmbeddingsAttentionBenchmark Dataset DesignBenchmark Dataset EvaluationFake News AnalysisLearned Feature RepresentationsLong Short-term Memory NetworksNeural Text RepresentationSiamese Networks
An important component of fake news detection is to evaluate the stance, different news sources take towards the assertion. Automatic stance detection, would facilitate the process of fact checking. In this paper, we present our stance detection system which comprises of siamese adaptation of Long Short Term Memory (LSTM) networks augmented with an attention mechanism, as siamese adaptation forces the LSTM to entirely capture the semantic differences during training, rather than supplementing the network with a more complex learner that can help resolve shortcomings in the learned representations. Our experiments on a public benchmark dataset, FakeNewsChallenge (FNC), demonstrate the effectiveness of our approach. It focuses on classifying the stance of a news article body relative to a headline as agree, disagree, discuss, or unrelated.
16
Hierarchical Attention Networks for Document Classification
Zichao Yang, Diyi Yang, Chris Dyer et al. · 2016 · 4.8K citations · Full text
SemEval-2016 Task 6: Detecting Stance in Tweets
Saif M. Mohammad, Svetlana Kiritchenko, Parinaz Sobhani et al. · 2016 · 857 citations · Full text