RECCON: Emotion Entailment Model¶

Overview¶

The RECCON Emotion Entailment model was proposed in Recognizing Emotion Cause in Conversations by Soujanya Poria, Navonil Majumder, Devamanyu Hazarika, Deepanway Ghosal, Rishabh Bhardwaj, Samson Yu Bai Jian, Pengfei Hong, Romila Ghosh, Abhinaba Roy, Niyati Chhaya, Alexande Gelbukh and Rada Mihalcea.

The abstract from the paper is the following:

Recognizing the cause behind emotions in text is a fundamental yet under-explored area of research in NLP. Advances in this area hold the potential to improve interpretability and performance in affect-based models. Identifying emotion causes at the utterance level in conversations is particularly challenging due to the intermingling dynamic among the interlocutors. To this end, we introduce the task of recognizing emotion cause in conversations with an accompanying dataset named RECCON. Furthermore, we define different cause types based on the source of the causes and establish strong transformer-based baselines to address two different sub-tasks of RECCON: 1) Causal Span Extraction and 2) Causal Emotion Entailment.

Link to the paper
Link to the dataset
Link to the original github

Getting started¶

The model pretrained on the RECCON data can be loaded and accessed with the following code:

from sgnlp.models.emotion_entailment import (
    RecconEmotionEntailmentConfig,
    RecconEmotionEntailmentModel,
    RecconEmotionEntailmentTokenizer,
    RecconEmotionEntailmentPreprocessor,
    RecconEmotionEntailmentPostprocessor,
)

config = RecconEmotionEntailmentConfig.from_pretrained(
    "https://storage.googleapis.com/sgnlp/models/reccon_emotion_entailment/config.json"
)
model = RecconEmotionEntailmentModel.from_pretrained(
    "https://storage.googleapis.com/sgnlp/models/reccon_emotion_entailment/pytorch_model.bin",
    config=config,
)
tokenizer = RecconEmotionEntailmentTokenizer.from_pretrained("roberta-base")
preprocessor = RecconEmotionEntailmentPreprocessor(tokenizer)
postprocess = RecconEmotionEntailmentPostprocessor()

input_batch = {
    "emotion": ["happiness", "happiness"],
    "target_utterance": ["Thank you very much .", "Thank you very much ."],
    "evidence_utterance": [
        "How can I forget my old friend ?",
        "My best wishes to you and the bride !",
    ],
    "conversation_history": [
        "It's very thoughtful of you to invite me to your wedding . How can I forget my old friend ? My best wishes to you and the bride ! Thank you very much .",
        "It's very thoughtful of you to invite me to your wedding . How can I forget my old friend ? My best wishes to you and the bride ! Thank you very much .",
    ],
}
input_dict = preprocessor(input_batch)
raw_output = model(**input_dict)
output = postprocess(raw_output)
print(output)
# [0, 1]

Input¶

The input data needs to be a dictionary with the following keys:

Key	Meaning
emotion	Emotion of the target utterance
target_utterance	Utterance whose emotion cause we are interested in
evidence_utterance	Potential evidence utterance for causing emotion in target utterance
conversation_history	All utterances from the beginning of the conversation till and including the target utterance

The values need to be a list of str and the list need to be of the same length across all keys. Refer to the original paper for more details of the inputs required.

Output¶

The output returned from RecconEmotionEntailmentPostprocessor instance is a list of int.

1 indicates that the evidence_utterance at the corresponding index caused the corresponding emotion in the target_utterance, while 0 indicates that the evidence_utterance at the corresponding index did not cause the corresponding emotion in the target_utterance.

The logits can be accessed from the raw output returned from the model.

Training¶

Dataset Preparation¶

Prepare the training and evaluation dataset in the format that is the same as the RECCON dataset in the authors’ repo. You can refer to the sample dataset here. Use the dataset with context.

Config Preparation¶

Create a copy of the config file. Update the following parameters: x_train_path, x_valid_path and train_args/output_dir. For the other parameters, you can either use the default values or modify it. You can refer to an example of the config file here.

Configuration key	Description
model_name	Pretrained model to use for training
x_train_path	Folder path to training data
x_valid_path	Folder path to validation data
max_seq_length	Maximum length of input sequence
train_args/output_dir	Folder path for model weights and checkpoints

You may refer to the other train_args parameters here.

Running Train Code¶

Import train() and parse_args_and_load_config() function. Set the path to the config file as the argument for the parse_args_and_load_config() function. Run train() on the config.

import json
from sgnlp.models.emotion_entailment import train
from sgnlp.models.emotion_entailment.utils import parse_args_and_load_config

cfg = parse_args_and_load_config('config/emotion_entailment_config.json')
train(cfg)

Evaluating¶

Dataset Preparation¶

Prepare the test dataset in the format that is the same as the RECCON dataset in the authors’ repo. You can refer to the sample dataset here. Use the dataset with context.

Config Preparation¶

Create a copy of the config file. Update the following parameters: eval_args/trained_model_dir , eval_args/x_test_path and results_path. For the other parameters, you can either use the default values or modify it. You can refer to an example of the config file here.

Configuration key	Description
model_name	Pretrained model to use for training
max_seq_length	Maximum length of input sequence
eval_args/trained_model_dir	Folder path for trained model weights
eval_args/x_test_path	Folder path of test data
eval_args/results_path	Folder path to save the test result
eval_args/per_device_eval_batch_size	Batch size for prediction
eval_args/no_cuda	Whether to use cuda for prediction

Running Evaluation Code¶

Import evaluate() and parse_args_and_load_config() function. Set the path to the config file as the argument for the parse_args_and_load_config() function. Run evaluate() on the config.

import json
from sgnlp.models.emotion_entailment import evaluate
from sgnlp.models.emotion_entailment.utils import parse_args_and_load_config

cfg = parse_args_and_load_config('config/emotion_entailment_config.json')
evaluate(cfg)