DeepStance at SemEval-2016 Task 6: Detecting Stance in Tweets Using Character and Word-Level CNNs

Vijayaraghavan, Prashanth; Sysoev, Ivan; Vosoughi, Soroush; Roy, Deb

doi:10.18653/v1/s16-1067

Cited by 40 publications

(29 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Cosine distance between embeddings of reference source tweets and those of unlabeled candidate tweets is used as a measurement of semantic similarity. Cosine similarity between vector representation of two sentences is a common metric for measuring semantic similarity [20]. Two semantically equivalent embeddings have a cosine similarity of 1, and two vectors with no relation have that of 0.…”

Section: A Overview Of the Proposed Methodsmentioning

confidence: 99%

“…In particular, character-level CNNs trained on augmented data achieves the best performance. Recent research [19,20] applies this method to tweets, and shows that data augmentation can bring performance gains in deep learning tasks on noisy and short social media texts. Vosoughi et al [19] augment domainindependent English tweets for training an encoder-decoder embedding model built with character-level CNN and long short-term memory (LSTM).…”

Section: Related Workmentioning

confidence: 99%

“…The number of tweets before data augmentation is not presented, but the author report that 3 million tweets in total are available after data augmentation. Another work [20] on tweet stance classification employs the same technique but uses Word2Vec instead of the WordNet thesaurus to replace words in text. Synonyms of a given word are ranked based on cosine similarity between the Word2Vec vector of given word and that of each synonym.…”

Section: Related Workmentioning

confidence: 99%

See 2 more Smart Citations

Neural language model based training data augmentation for weakly supervised early rumor detection

Han

Gao

Ciravegna

2019

Proceedings of the 2019 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining

View full text Add to dashboard Cite

The scarcity and class imbalance of training data are known issues in current rumor detection tasks. We propose a straight-forward and general-purpose data augmentation technique which is beneficial to early rumor detection relying on event propagation patterns. The key idea is to exploit massive unlabeled event data sets on social media to augment limited labeled rumor source tweets. This work is based on rumor spreading patterns revealed by recent rumor studies and semantic relatedness between labeled and unlabeled data. A state-ofthe-art neural language model (NLM) and large credibilityfocused Twitter corpora are employed to learn context-sensitive representations of rumor tweets. Six different real-world events based on three publicly available rumor datasets are employed in our experiments to provide a comparative evaluation of the effectiveness of the method. The results show that our method can expand the size of an existing rumor data set nearly by 200% and corresponding social context (i.e., conversational threads) by 100% with reasonable quality. Preliminary experiments with a state-of-the-art deep learning-based rumor detection model show that augmented data can alleviate over-fitting and class imbalance caused by limited train data and can help to train complex neural networks (NNs). With augmented data, the performance of rumor detection can be improved by 12.1% in terms of F-score. Our experiments also indicate that augmented training data can help to generalize rumor detection models on unseen rumors.

show abstract

Section: A Overview Of the Proposed Methodsmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Neural language model based training data augmentation for weakly supervised early rumor detection

Han

Gao

Ciravegna

2019

Proceedings of the 2019 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining

View full text Add to dashboard Cite

show abstract

“…MITRE [14] provided the best deep learning solution in the contest, initializing weights from a 256-dimensional word embeddings learned using the word2vec skip-gram algorithm [6], followed by a second layer with 128 LSTM units. Among others, pkudblab [12] and DeepStance [11] use deep CNN models. Augenstein et al [1] employ a bidirectional attention model.…”

Section: Related Workmentioning

confidence: 99%

Topical Stance Detection for Twitter: A Two-Phase LSTM Model Using Attention

Dey

Shrivastava

Kaushik

2018

Lecture Notes in Computer Science

View full text Add to dashboard Cite

The topical stance detection problem addresses detecting the stance of the text content with respect to a given topic: whether the sentiment of the given text content is in FAVOR of (positive), is AGAINST (negative), or is NONE (neutral) towards the given topic. Using the concept of attention, we develop a two-phase solution. In the first phase, we classify subjectivity -whether a given tweet is neutral or subjective with respect to the given topic. In the second phase, we classify sentiment of the subjective tweets (ignoring the neutral tweets) -whether a given subjective tweet has a FAVOR or AGAINST stance towards the topic. We propose a Long Short-Term memory (LSTM) based deep neural network for each phase, and embed attention at each of the phases. On the SemEval 2016 stance detection Twitter task dataset [7], we obtain a best-case macro F-score of 68.84% and a best-case accuracy of 60.2%, outperforming the existing deep learning based solutions. Our framework, T-PAN, is the first in the topical stance detection literature, that uses deep learning within a two-phase architecture.

show abstract

“…Task A External resources: Bag-of-Words and word vectors. DeepStance [19] Overall approach: A set of naive bayes classifiers using deep learning.…”

Section: Our Approachmentioning

confidence: 99%

Friends and Enemies of Clinton and Trump: Using Context for Detecting Stance in Political Tweets

Lai

Farías

Patti

et al. 2017

Advances in Computational Intelligence

View full text Add to dashboard Cite

Abstract. Stance detection, the task of identifying the speaker's opinion towards a particular target, has attracted the attention of researchers. This paper describes a novel approach for detecting stance in Twitter. We define a set of features in order to consider the context surrounding a target of interest with the final aim of training a model for predicting the stance towards the mentioned targets. In particular, we are interested in investigating political debates in social media. For this reason we evaluated our approach focusing on two targets of the SemEval-2016 Task 6 on Detecting stance in tweets, which are related to the political campaign for the 2016 U.S. presidential elections: Hillary Clinton vs. Donald Trump. For the sake of comparison with the state of the art, we evaluated our model against the dataset released in the SemEval-2016 Task 6 shared task competition. Our results outperform the best ones obtained by participating teams, and show that information about enemies and friends of politicians help in detecting stance towards them.

show abstract

DeepStance at SemEval-2016 Task 6: Detecting Stance in Tweets Using Character and Word-Level CNNs

Cited by 40 publications

References 10 publications

Neural language model based training data augmentation for weakly supervised early rumor detection

Neural language model based training data augmentation for weakly supervised early rumor detection

Topical Stance Detection for Twitter: A Two-Phase LSTM Model Using Attention

Friends and Enemies of Clinton and Trump: Using Context for Detecting Stance in Political Tweets

Contact Info

Product

Resources

About