Fake News Challenge Stage 1
FNC-1 was designed as a stance detection dataset and it contains 75,385 labeled headline and article pairs. The pairs are labelled as either agree, disagree, discuss, and unrelated. Each headline in the dataset is phrased as a statement
Source: Investigating Rumor News Using Agreement-Aware Search
Image Source: http://www.fakenewschallenge.org/
Variants: FNC-1
This dataset is used in 2 benchmarks:
| Task | Model | Paper | Date |
|---|---|---|---|
| Stance Detection | TESTED | Topic-Guided Sampling For Data-Efficient Multi-Domain … | 2023-06-01 |
| Fake News Detection | ZAINAB A. JAWAD, AHMED J. OBAID (CNN and DNN with SCM, 2022) | Combination Of Convolution Neural Networks … | 2022-10-15 |
| Fake News Detection | Bi-LSTM (max-pooling, attention) | Combining Similarity Features and Deep … | 2018-11-02 |
| Fake News Detection | Neural method from Mohtarami et al. + TF-IDF (Mohtarami et al., 2018) | Automatic Stance Detection Using End-to-End … | 2018-04-20 |
| Fake News Detection | Neural method from Mohtarami et al. (Mohtarami et al., 2018) | Automatic Stance Detection Using End-to-End … | 2018-04-20 |
| Fake News Detection | Bhatt et al. | On the Benefit of Combining … | 2017-12-11 |
| Fake News Detection | Neural baseline based on bi-directional LSTMs (Bhatt et al., 2017) | On the Benefit of Combining … | 2017-12-11 |
| Fake News Detection | Baseline based on skip-thought embeddings (Bhatt et al., 2017) | On the Benefit of Combining … | 2017-12-11 |
| Fake News Detection | Baseline based on word2vec + hand-crafted features (Bhatt et al., 2017) | On the Benefit of Combining … | 2017-12-11 |
| Fake News Detection | 3rd place at FNC-1 - Team UCL Machine Reading (Riedel et al., 2017) | A simple but tough-to-beat baseline … | 2017-07-11 |
Recent papers with results on this dataset: