CARER: Contextualized Affect Representations for Emotion Recognition
By Elvis Saravia, Hsien-Chi Toby Liu, Yen-Hao Huang, Junlin Wu, Yi-Shin Chen. 2018
503 citations recorded by the source corpus.
Emotions are expressed in nuanced ways, which varies by collective or individual experiences, knowledge, and beliefs. Therefore, to understand emotion, as conveyed through text, a robust mechanism capable of capturing and modeling different linguistic nuances and phenomena is needed. We propose a semi-supervised, graph-based algorithm to produce rich structural descriptors which serve as the building blocks for constructing contextualized affect representations from text. The pattern-based representations are further enriched with word embeddings and evaluated through several emotion recognition tasks. Our experimental results demonstrate that the proposed method outperforms state-of-the-art techniques on emotion recognition tasks.
Subfields: Artificial Intelligence, Experimental and Cognitive Psychology, Sentiment & Opinion Analysis
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Iryna Gurevych
- Luke S. Zettlemoyer
- Mohit Bansal
- Percy Liang
- Noah A. Smith
- Jonas Pfeiffer
- Diyi Yang
- Graham Neubig
In Experimental and Cognitive Psychology
In Sentiment & Opinion Analysis
Citations in the corpus
- Continual Pre-Training Mitigates Forgetting in Language and Vision (2022, methodology): “We leverage 3 FC datasets: sentiment analysis from tweets (Saravia et al., 2018), Question Answering Natural Language Inference ( QNLI ) (Wang et al., 2018) and the SentEval benchmark (Conneau & Kiela, 2018), which contains 20 different datasets.”
- An Empirical Investigation of the Role of Pre-training in Lifelong Learning (2021, methodology): “(13) SciTail (Khot et al., 2018); (14) Implicit Discourse Relation Classification (Prasad et al., 2019; Kim et al., 2020); and (15) Emotion Detection (Saravia et al., 2018).”
- Demystifying Prompts in Language Models via Perplexity Estimation (2022, background): “2015); (f) Emotion (classification to emotions; Saravia et al. 2018); (g) Tweet Offensive (classification to offensive vs.”
- MTEB: Massive Text Embedding Benchmark (2022, background): “For each user query the label is an intent among 77 intents like ’activate_my_card’, ’apple_pay’, ’bank_transfer’, etc. Emotion (Saravia et al., 2018) Dataset of English Twitter messages with six basic emotions: anger, fear, joy, love, sadness, and surprise.”
- Efficient Few-Shot Learning Without Prompts (2022, unknown): “(Saravia et al., 2018).”
- Can Large Language Models Transform Computational Social Science? (2023, methodology): “We evaluate emotion detection with weakly labeled Twitter data from Saravia et al. (2018), which uses Plutchik’s 8 emotional categories: anger, anticipation, disgust, fear, joy, sadness, surprise, and trust.”
- What to Pre-Train on? Efficient Intermediate Task Selection (2021, background): “84.72 85.26 CoLA (Warstadt et al., 2019) 8.5K linguistic acceptability books, journals Matthews 59.18 62.18 Emotion (Saravia et al., 2018) 16K emotion classification Twitter acc.”
- NV-Embed: Improved Techniques for Training LLMs as Generalist Embedding Models (2024, unknown): “…AmazonCounterfactual (O’Neill et al., 2021), Banking77 (Casanueva et al., 2020), Emotion (Saravia et al., 2018), IMDB (Maas et al., 2011), MTOPDomain/MTOPIntent (Li et al., 2021), ToxicConversations (Adams et al., 2019), TweetSentimentExtraction (Maggie, 2020), AmazonPolarity (McAuley &…”
- Layer-wise Representation Dynamics: An Empirical Investigation Across Embedders and Base LLMs (2026, unknown): “…Accuracy Banking77Classification [6] Classification Accuracy DBpediaClassification [23] Classification Accuracy EmotionClassification [58] Classification Accuracy ImdbClassification [41] Classification Accuracy MassiveIntentClassification [16] Classification Accuracy…”
- Beyond Matryoshka: Revisiting Sparse Coding for Adaptive Representation (2025, unknown): “• Emotion (Saravia et al., 2018): The dataset consists of English Twitter messages categorized into basic emotions, including anger, fear, joy, love, sadness, and surprise.”
- Length-Induced Embedding Collapse in PLM-based Models (2024, unknown): “Emotion (Saravia et al., 2018) comprises Twitter messages categorized by six fundamental emotions: anger, fear, joy, love, sadness, and surprise.”
- Bagging-Based Model Merging for Robust General Text Embeddings (2026, unknown): “• Classification: AmazonReviews-Classification [40], Banking77-Classification [7], Emotion-Classification [48], MTOPIntent-Classification[31],IMDB-Classification[38],ToxicConversations-Classification[1],TweetSentimentExtraction-Classification[57], AmazonCounterfactual-Classification[45].”