Elamur Research⌕
☰
home / papers / 2026 / AFourP at SemEval-2026 Task 2: Predicting Variation in Emotional Valence and Arousal over Time from Ecological Essays

AFourP at SemEval-2026 Task 2: Predicting Variation in Emotional Valence and Arousal over Time from Ecological Essays

Paper · 2026 · Shrika Thota, Lakshmi Priya Swaminatha Rao, Shivaanee Sk et al.

Summary

The AFourP team took part in Subtask 1 of SemEval-2026 Task 2, which asks systems to estimate continuous Valence (scale −2 to 2) and Arousal (scale 0 to 2) scores for diary-like texts. The texts were gathered through Ecological Momentary Assessment between 2021 and 2024. They range from narrative paragraphs to bare lists of feeling words.

The system is deliberately simple. It fine-tunes RoBERTa-base end to end and adds one linear layer that maps the [CLS] vector to two outputs. Training minimizes MSE with AdamW. Each entry is scored on its own, without using author identity or time order.

On the official test set the system reached Pearson-based composite scores of .679 for Valence and .466 for Arousal, ranking 4th. It clearly beat the organizers' linear(BERT) baseline. Ablations show that fine-tuning the encoder is essential and that shorter inputs work slightly better.

The authors argue that ignoring per-user emotional baselines and temporal context is the system's main weakness. They propose user embeddings and hierarchical sequence models as next steps.

Key points

  • Data: the training split has 2,764 texts from 137 users, of which 51.8% are word lists and 48.2% are prose. The test split has 1,737 texts from 91 users, with 46 users appearing in both splits. Per-user mean valence varies widely across users (SD 0.781).
  • Setup: RoBERTa-base (~125M parameters) with byte-level BPE truncated to 256 tokens. Training used a learning rate of 2×10⁻⁵, batch size 8 and 3 epochs on a Tesla T4. Validation MSE was stable across three random seeds.
  • Ablations on an 80/20 validation split (composite r):
    • linear head: 0.6626
    • 2-layer MLP head: 0.6490
    • frozen encoder: 0.1400
    • 128-token limit: 0.6785, the best result
  • Official test results: .679 (V) and .466 (A), 4th place. The organizers' linear(BERT) baseline scored .557 and .299.
  • Future work: the authors suggest learned user embeddings, hierarchical temporal encoders, training on label deltas, contrastive objectives and uncertainty estimation. Code is released on GitHub (shrikathota/SemEval2026-Task2).

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓Sri Sivasubramaniya Nadar College of EngineeringChennai · India10 works · 9 citations

 Fields, topics and concepts

📚 Shelves 5The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksPsychology · 1,552 worksExperimental and Cognitive Psychology · 870 worksSocial Psychology · 306 works

🧭 Topics 1Research subjects and languages the work is about.Sentiment & emotion · 972 works

Cite

Cite as Shrika Thota, Lakshmi Priya Swaminatha Rao, Shivaanee Sk, Thirumurugan Ra, Vishal Muralidharan and Dhannya Santhakumari Madhavan. 2026. AFourP at SemEval-2026 Task 2: Predicting Variation in Emotional Valence and Arousal over Time from Ecological Essays. Proceedings of the 20th International Workshop on Semantic Evaluation (2026).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.semeval-1.217

@inproceedings{thota2026afourp,
  title = {{AFourP at SemEval-2026 Task 2: Predicting Variation in Emotional Valence and Arousal over Time from Ecological Essays}},
  author = {Shrika Thota, Lakshmi Priya Swaminatha Rao, Shivaanee Sk, Thirumurugan Ra, Vishal Muralidharan and Dhannya Santhakumari Madhavan},
  booktitle = {Proceedings of the 20th International Workshop on Semantic Evaluation (2026)},
  year = {2026},
  doi = {10.18653/v1/2026.semeval-1.217},
  url = {https://doi.org/10.18653/v1/2026.semeval-1.217}
}

References

9 references, 3 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Sven Buechel and Udo Hahn. 2017. EmoBank: Studying the impact of annotation perspective and representation format on dimensional emotion analysis. In Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Volume 2, Short Papers, pages 578–585, Va… find
  2. Lukas Christ, Shahin Amiriparian, Manuel Milling, Ilhan Aslan, and Björn Schuller. 2024. Modeling emotional trajectories in written stories utilizing transformers and weakly-supervised learning. In Findings of the Association for Computational Linguistics: ACL 2024, pages 7144–7159, Bangkok, Thailan… find
  3. Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. BERT: Pre-training of deep bidirectional transformers for language understanding. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies,…
  4. C. Hutto and Eric Gilbert. 2014. Vader: A parsimonious rule-based model for sentiment analysis of social media text. Proceedings of the International AAAI Conference on Web and Social Media, 8(1):216–225. find
  5. Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, and Veselin Stoyanov. 2019. Roberta: A robustly optimized bert pretraining approach. Preprint, arXiv:1907.11692.
  6. Saif M. Mohammad and Peter D. Turney. 2013. Crowdsourcing a word-emotion association lexicon. Computational Intelligence, 29(3):436–465. find
  7. James A. Russell. 1980. A circumplex model of affect. Journal of Personality and Social Psychology, 39:1161–1178. find
  8. Saul Shiffman, Arthur A. Stone, and Michael R. Hufford. 2008. Ecological momentary assessment. Annual Review of Clinical Psychology, 4:1–32. find
  9. Nikita Soni, H. Andrew Schwartz, Ryan L. Boyd, Phi Long Bui, Syeda Mahwish, August Håkan Nilsson, Adithya V Ganesan, Lyle Ungar, Niranjan Balasubramanian, and Saif M. Mohammad. 2026. SemEval-2026 task 2: Predicting variation in emotional valence and arousal over time from ecological essays. In Proce…

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/thota-2026-afourp-semeval-2026-task-2/