DataStories at SemEval-2017 Task 6: Siamese LSTM with Attention for
            Humorous Text Comparison

Baziotis, Christos; Pelekis, Nikos; Doulkeridis, Christos

doi:10.18653/v1/s17-2065

Cited by 18 publications

(22 citation statements)

References 13 publications

(21 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Considering neural network-based systems, LSTMs were used the most, which is expected given the sequential nature of text data. Plain LSTM models alone, using pretrained word embeddings, achieved competitive results, and DataStories (Baziotis et al, 2017) ranked third using a siamese bidirectional LSTM model with attention.…”

Section: System Analysismentioning

confidence: 99%

SemEval-2017 Task 6: #HashtagWars: Learning a Sense of Humor

Potash¹,

Romanov²,

Rumshisky³

2017

Proceedings of the 11th International Workshop on Semantic Evaluation (SemEval-2017)

View full text Add to dashboard Cite

This paper describes a new shared task for humor understanding that attempts to eschew the ubiquitous binary approach to humor detection and focus on comparative humor ranking instead. The task is based on a new dataset of funny tweets posted in response to shared hashtags, collected from the 'Hashtag Wars' segment of the TV show @midnight. The results are evaluated in two subtasks that require the participants to generate either the correct pairwise comparisons of tweets (subtask A), or the correct ranking of the tweets (subtask B) in terms of how funny they are. 7 teams participated in subtask A, and 5 teams participated in subtask B. The best accuracy in subtask A was 0.675. The best (lowest) rank edit distance for subtask B was 0.872.

show abstract

Section: System Analysismentioning

confidence: 99%

SemEval-2017 Task 6: #HashtagWars: Learning a Sense of Humor

Potash¹,

Romanov²,

Rumshisky³

2017

Proceedings of the 11th International Workshop on Semantic Evaluation (SemEval-2017)

View full text Add to dashboard Cite

show abstract

“…We also compare the results with sentiment specific word embeddings (Tang et al, 2014), where we use Fully connected layers along with attention as the downstream model. For Sem-3 dataset we compare our results with RCNN (Yin et al, 2017) and Siamese network (Baziotis et al, 2017), which were top performing teams in the task. In addition, we separately train each task with same model parameters Figure 3: Plot of how intensity of the tweet changes with the words, signifying the importance of the sequence without multi-task frame work, to observe the improvement due to multi-task and also to access the ability of the model architecture proposed.…”

Section: Resultsmentioning

confidence: 99%

Learning representations for sentiment classification using Multi-task framework

Meisheri¹,

Khadilkar²

2018

Proceedings of the 9th Workshop on Computational Approaches to Subjectivity, Sentiment and Social Media Analysis

View full text Add to dashboard Cite

Most of the existing state of the art sentiment classification techniques involve the use of pre-trained embeddings. This paper postulates a generalized representation that collates training on multiple datasets using a Multitask learning framework. We incorporate publicly available, pre-trained embeddings with Bidirectional LSTM's to develop the multitask model. We validate the representations on an independent test Irony dataset that can contain several sentiments within each sample, with an arbitrary distribution. Our experiments show a significant improvement in results as compared to the available baselines for individual datasets on which independent models are trained. Results also suggest superior performance of the representations generated over Irony dataset.

show abstract

“…The weights we used for the embedding matrix of these models were the same as Baziotis et al (2017b), pre-trained on 330 millions of english tweets messages posted from 12/2012 to 07/2016 with GloVe.…”

Section: Cat-lstm and Bin-lstmmentioning

confidence: 99%

“…For the specific task of emotion classification in textual conversation, Gupta et al (2017) achieved a F1µ score of 0.7134 on the same dataset, using an architecture based on LSTM. For sentiment analysis, other successful approaches also used Bi-LSTM (Baziotis et al, 2017b) as well as transfer learning (Daval-Frerot et al, 2018).…”

Section: Introductionmentioning

confidence: 99%

SCIA at SemEval-2019 Task 3: Sentiment Analysis in Textual Conversations Using Deep Learning

Rebiai¹,

Andersen²,

Debrenne³

et al. 2019

Proceedings of the 13th International Workshop on Semantic Evaluation

View full text Add to dashboard Cite

In this paper we present our submission for SemEval-2019 Task 3: EmoContext. The task consisted of classifying a textual dialogue into one of four emotion classes: happy, sad, angry or others. Our approach tried to improve on multiple aspects, preprocessing with an emphasis on spell-checking and ensembling with four different models: Bi-directional contextual LSTM (BC-LSTM), categorical Bi-LSTM (CAT-LSTM), binary convolutional Bi-LSTM (BIN-LSTM) and Gated Recurrent Unit (GRU). On the leader-board, we submitted two systems that obtained a micro F1 score (F1µ) of 0.711 and 0.712. After the competition, we merged our two systems with ensembling, which achieved a F1µ of 0.7324 on the test dataset.

show abstract

DataStories at SemEval-2017 Task 6: Siamese LSTM with Attention for Humorous Text Comparison

Cited by 18 publications

References 13 publications

SemEval-2017 Task 6: #HashtagWars: Learning a Sense of Humor

SemEval-2017 Task 6: #HashtagWars: Learning a Sense of Humor

Learning representations for sentiment classification using Multi-task framework

SCIA at SemEval-2019 Task 3: Sentiment Analysis in Textual Conversations Using Deep Learning

Contact Info

Product

Resources

About