Research On Dialog Emotion Recognition Method Based On Multi-Feature Fusion

Currently,social media platforms allow users to universally express their opinions and sentiments due to their convenience and openness.As one of the most common ways of communication,dialogue contains rich information and sentiment expression of participants.Emotion Recognition in Conversations(ERC)plays a central role in the field of NLP.In the past few years,the task gained attention from the NLP community due to the increase of public availability of dialogue data.It can be used to analyze dialogues that take place on social media or other scenes and provide support for downstream tasks,such as dialogue response generation.It can also aid in analyzing dialogues in real times,which can be public opinion monitoring,interviews,psychological consulting and more.Recent research efforts,such as DAG-ERC,attempt to model the informative dependencies based on both adjacent and distant utterances in a conversation,and achieve SOTA performances.However,the emotion preference of topics and speakers in a conversation haven’t been well explored.We argue that different conversations usually correspond to different topics which would have specific emotion preferences.Meanwhile,different speakers of conversations would also have distinct emotion preference.To address this issue,we propose a novel model named Topic-Aware Recurrent Coupled Speaker Model(TA-RCSM)for emotion recognition in conversations.Specifically,we develop a topicaware speaker feature extraction module to learn the topic-aware speaker representations,which will be injected into their corresponding utterance representations learned by a DAG propagation layer.Moreover,as there is a strong correlation among the emotion labels of each utterance in a conversation,we further treat the ERC task as a sequence labeling task and employ a CRF layer to guide the representation learning of utterances.We carry out extensive experiments on four benchmark datasets,and results demonstrate that our proposed approach significantly outperforms state-of-the-art baselines.Dialogue sentiment classification and dialogue act recognition are two sub-tasks in dialogue systems that aim to predict the sentiment and act label of each utterance in a dialogue.These two tasks are influenced by multiple factors and closely related.However,existing models do not make reasonable use of the explicit and implicit information contained in a dialogue,such as speaker information,temporal information,and label information,and simply or coarse-grained modeling the interaction of two tasks.To solve the above problems,this thesis proposes a new multi-task learning model,namely Speakeraware Cross-task Co-interactive Graph Network(SA-CCGN).The model first captures speaker-aware sentiment and act cues along with the time to generate speaker-aware utterance representations,and then adequately models information propagation within a conversation and information interaction between tasks through a cross-task co-interactive graph network,where information propagation of a conversation is modeled by constructing a directed acyclic graph,and after each graph propagation,appropriate interaction between two tasks is performed using the co-interactive layer.Finally,the label information is introduced,i.e.,differentiation and correlation between labels,which can constrain the model training when decoding.Specifically,in the multi-loss decoder,the supervised contrastive learning loss is used to make the learned representation of different labels more differentiated and the conditional random field loss is used to constrain the generation of adjacent label sequences,then the final sentiment and act label of each utterance are obtained.In order to prove the effectiveness of the model in this thesis,experiments were conducted on the two public two-way dialogue datasets,and we compare our proposed method with a variety of state-of-the-art methods.Experimental results on two public datasets show that our model outperforms the current state-of-the-art joint model Co-GAT,with an improvement of 4.57% and 3.33% in F1 scores for the dialogue sentiment classification task and 2.15% and 0.63% in F1 scores for the dialogue act recognition task on the two datasets,respectively,while reducing the number of parameters and memory usage by about 1/2.The performance of SA-CCGN on two public datasets exceeds the best results in the known literature.Experiments show that this method can effectively utilize dialogue information,and has obvious advantages in dialogue sentiment classification task and dialogue act recognition task compared to previous methods.