- Notifications
You must be signed in to change notification settings - Fork1
[ECAI 2023] Official implementation of "FATRER: Full-Attention Topic Regularizer for Accurate and Robust Conversational Emotion Recognition"
License
ludybupt/FATRER
Folders and files
| Name | Name | Last commit message | Last commit date | |
|---|---|---|---|---|
Repository files navigation
[ECAI 2023] Official Pytorch implementation of "FATRER: Full-Attention Topic Regularizer for Accurate and Robust Conversational Emotion Recognition"[paper]
Full-attention topic regularizer(FATRER) introduces an emotion-relatedglobal view when modeling the local context in a conversation. Ajoint topic modeling strategy is introduced to implement regularization from both representation and loss perspectives. To avoid overregularization, FATRER drops the constraints on prior distributions that exist in traditional topic modeling and perform probabilistic approximations based entirely on attention alignment. Experiments showthat FATRER obtain more favorable results than state-of-the-artmodels, and gain convincing robustness.
- [2023-10-02]: FARTER will be presented orally(Video) in ECAI2023(Programme) Technical Session 1 at 09:30 AM in Room S4A and Poster Session 1 at 11:15 AM in Hall S3B(Poster).
- [2023-07-15]: FARTER has been accepted byECAI 2023 (Paper 223).
- Python 3.9.12
- Pytorch 1.10.1+cu113
pipinstll -rrequirements.txt
- IEMOCAP/MELD/EmoryNLP/EmoryNLP
- FARTER-Multi:
# trainpythonmain.pyconf/FATRER_multi.yaml#train and conduct attack(U+C) based on PWWS(per 50 epoch):pythonmain.pyconf/FATRER_multi_pwws_attack.yaml#train and conduct attack(U+C) based on TextFooler(per 50 epoch):pythonmain.pyconf/FATRER_multi_textfooler_attack.yaml#train and conduct attack(U+C) based on TextBugger(per 50 epoch):pythonmain.pyconf/FATRER_multi_textbugger_attack.yaml
- FARTER-Multi(without topic-oriented regularization):
# trainpythonmain.pyconf/FATRER_multi_wo_topic.yaml
- FARTER-Single:
# trainpythonmain.pyconf/FATRER_single.yaml
- FARTER-Single(without topic-oriented regularization):
#trainpythonmain.pyconf/FATRER_single_wo_topic.yaml
- DialTRM(Baseline):
#trainpythonmain.pyconf/Baseline.yaml
- VAE(topic-oriented)
#train VAE(Laplace)pythonmain.pyconf/VAE_Laplace.yaml#train VAE(Dirichlet)pythonmain.pyconf/VAE_Dirichlet.yaml#train VAE(Gamma)pythonmain.pyconf/VAE_Gamma.yaml#train VAE(LogNormal)pythonmain.pyconf/VAE_LogNormal.yaml
- FARTER-Multi:
pythonmain.pyconf/FATRER_multi_MELD.yaml
- FARTER-Single:
pythonmain.pyconf/FATRER_single_MELD.yaml
- FARTER-Multi:
pythonmain.pyconf/FATRER_multi_EmoryNLP.yaml
- FARTER-Single:
pythonmain.pyconf/FATRER_single_EmoryNLP.yaml
- FARTER-Multi:
pythonmain.pyconf/FATRER_multi_DailyDialog.yaml
- FARTER-Single:
pythonmain.pyconf/FATRER_single_DailyDialog.yaml
Please cite the following paper if you find this code useful in your work.
@article{mao2023fatrer, title={FATRER: Full-Attention Topic Regularizer for Accurate and Robust Conversational Emotion Recognition}, author={Mao, Yuzhao and Lu, Di and Wang, Xiaojie and Zhang, Yang}, journal={arXiv preprint arXiv:2307.12221}, year={2023}}MIT license