HMTL: Heterogeneous Modality Transfer Learning for Audio-Visual Sentiment Analysis

Multimodal sentiment analysis is an important research area that predicts speaker’s sentiment tendency through features extracted from textual, visual and acoustic modalities. The central challenge is the fusion method of the multimodal information. A variety of fusion methods have been proposed, but few of them adopt end-to-end translation models to mine the subtle correlation between modalities. Enlightened by recent success of Transformer in the area of machine translation, we propose a new fusion method, TransModality, to address the task of multimodal sentiment analysis. We assume that translation between modalities contributes to a better joint representation of speaker’s utterance. With Transformer, the learned features embody the information both from the source modality and the target modality. We validate our model on multiple multimodal datasets: CMU-MOSI, MELD, IEMOCAP. The experiments show that our proposed method achieves the state-of-the-art performance.

Integrating color cues to improve multimodal sentiment analysis in social media

With the increasing popularity of video sharing websites such as YouTube and Facebook, multimodal sentiment analysis has received increasing attention from the scientific community. Contrary to previous works in multimodal sentiment analysis which focus on holistic information in speech segments such as bag of words representations and average facial expression intensity, we propose a novel deep architecture for multimodal sentiment analysis that is able to perform modality fusion at the word level. In this paper, we propose the Gated Multimodal Embedding LSTM with Temporal Attention (GME-LSTM(A)) model that is composed of 2 modules. The Gated Multimodal Embedding allows us to alleviate the difficulties of fusion when there are noisy modalities. The LSTM with Temporal Attention can perform word level fusion at a finer fusion resolution between the input modalities and attends to the most important time steps. As a result, the GME-LSTM(A) is able to better model the multimodal structure of speech through time and perform better sentiment comprehension. We demonstrate the effectiveness of this approach on the publicly-available Multimodal Corpus of Sentiment Intensity and Subjectivity Analysis (CMU-MOSI) dataset by achieving state-of-the-art sentiment classification and regression results. Qualitative analysis on our model emphasizes the importance of the Temporal Attention Layer in sentiment prediction because the additional acoustic and visual modalities are noisy. We also demonstrate the effectiveness of the Gated Multimodal Embedding in selectively filtering these noisy modalities out. These results and analysis open new areas in the study of sentiment analysis in human communication and provide new models for multimodal fusion.

Implicit emotions are often expressed through implicit and weak clues between modalities due to the lack of explicit emotional feature words, representing a significant challenge for multimodal sentiment analysis. In order to improve implicit emotion recognition, this paper proposes a multimodal sentiment analysis method that integrates KAN and the modal dynamic fusion mechanism. This method first introduces the KAN structure to construct a modal feature encoder to enhance the emotional expression ability of features. Then, the emotional contribution weight of each modality is calculated using the difference between the unimodal and multimodal sentiment scores, and the cross-attention mechanism guided by the main modality is used for feature fusion. Experiments on four datasets, CH-SIMS, CH-SIMSv2, MOSI, and MOSEI, show that the proposed method significantly outperforms the mainstream model in multiple indicators, especially when dealing with samples with implicit or ambiguous emotional expressions. The results verify the effectiveness of enhancing feature encoding capabilities and utilizing modal asymmetry information in implicit sentiment analysis.

The Multimodal Sentiment Analysis Challenge (MuSe) 2024 addresses two contemporary multimodal affect and sentiment analysis problems: In the Social Perception Sub-Challenge (MuSe-Perception), participants will predict 16 different social attributes of individuals such as assertiveness, dominance, likability, and sincerity based on the provided audio-visual data. The Cross-Cultural Humor Detection Sub-Challenge (MuSe-Humor) dataset expands upon the Passau Spontaneous Football Coach Humor (Passau-SFCH) dataset, focusing on the detection of spontaneous humor in a cross-lingual and cross-cultural setting. The main objective of MuSe 2024 is to unite a broad audience from various research domains, including multimodal sentiment analysis, audio-visual affective computing, continuous signal processing, and natural language processing. By fostering collaboration and exchange among experts in these fields, the MuSe 2024 endeavors to advance the understanding and application of sentiment analysis and affective computing across multiple modalities. This baseline paper provides details on each sub-challenge and its corresponding dataset, extracted features from each data modality, and discusses challenge baselines. For our baseline system, we make use of a range of Transformers and expert-designed features and train Gated Recurrent Unit (GRU)-Recurrent Neural Network (RNN) models on them, resulting in a competitive baseline system. On the unseen test datasets of the respective sub-challenges, it achieves a mean Pearson's Correlation Coefficient (ρ) of 0.3573 for MuSe-Perception and an Area Under the Curve (AUC) value of 0.8682 for MuSe-Humor.

Multimodal Aspect-Based Sentiment Analysis (MABSA) is an essential task in sentiment analysis that has garnered considerable attention in recent years. Typical approaches in MABSA often utilize cross-modal Transformers to capture interactions between textual and visual modalities. However, bridging the semantic gap between modalities spaces and addressing interference from irrelevant visual objects at different scales remains challenging. To tackle these limitations, we present the Multi-level Textual-Visual Alignment and Fusion Network (MTVAF) in this work, which incorporates three auxiliary tasks. Specifically, MTVAF first transforms multi-level image information into image descriptions, facial descriptions, and optical characters. These are then concatenated with the textual input to form a textual+visual input, facilitating comprehensive alignment between visual and textual modalities. Next, both inputs are fed into an integrated text model that incorporates relevant visual representations. Dynamic attention mechanisms are employed to generate visual prompts to control cross-modal fusion. Finally, we align the probability distributions of the textual input space and the textual+visual input space, effectively reducing noise introduced during the alignment process. Experimental results on two MABSA benchmark datasets demonstrate the effectiveness of the proposed MTVAF, showcasing its superior performance compared to state-of-the-art approaches. Our codes are available at https://github.com/MKMaS-GUET/MTVAF.

We present a case study as part of a work-in-progress project about multimodal sentiment analysis on historic German plays, taking Emilia Galotti by G. E. Lessing as our initial use case. We analyze the textual version and an audio version (audio book). We focus on ready-to-use sentiment analysis methods: For the textual component, we implement a naive lexicon-based approach and another approach that enhances the lexicon by means of several NLP methods. For the audio analysis, we use the free version of the Vokaturi tool. We compare the results of all approaches and evaluate them against the annotations of a human expert, which serves as a gold standard. For our use case, we can show that audio and text sentiment analysis behave very differently: textual sentiment analysis tends to predict sentiment as rather negative and audio sentiment as rather positive. Compared to the gold standard, the textual sentiment analysis achieves accuracies of 56% while the accuracy for audio sentiment analysis is only 32%. We discuss possible reasons for these mediocre results and give an outlook on further steps we want to pursue in the context of multimodal sentiment analysis on historic plays.

Inferring Sentiments from Supervised Classification of Text and Speech cues using Fuzzy Rules

Modality translation-based multimodal sentiment analysis under uncertain missing modalities

Multimodal sentiment analysis (MMSA) is a research method that extracts effective information from heterogeneous modal information. Then, MMSA processes the multimodal data and performs sentiment analysis. Along with big data and machine learning development, multimodal sentiment analysis has become a hot research direction in multimodal learning and natural language processing. Although various feature extraction methods and information fusion methods have been continuously proposed, challenges exist in MMSA research. First, in terms of feature extraction, pre-trained models trained with many data sets can obtain higher quality features, but research on how to use these feature extraction methods to extract the best features is still needed. Currently, the more popular feature fusion methods do not focus on the interaction between multiple modal information and the retention of basic information. To overcome these problems a multimodal sentiment analysis model utilizes text features as core modal features, using video and audio modal features as auxiliary modal features, multimodal feature modality attention mechanism to extract the intrinsic connection between different modalities. The attention mechanism uses the features of video modality and audio modality as the focus and then enhances the text modality with the fusion of video modality and modality. To improve the quality of extracted features, this method chooses the transfer learning training method and uses the pre-trained model for processing. This research uses the CMU-MOSI dataset to test the proposed method. Experimental results show that the performance of the proposed model in emotion score prediction and emotion classification tasks exceeds traditional methods and baseline methods.

In the fields of artificial intelligence (AI) and natural language processing (NLP), sentiment analysis (SA) has become increasingly popular. Demand for automating user sentiment analysis of goods and services is rising. Videos, as opposed to just text, are becoming more and more common online for sharing opinions. This has made the use of various modalities in SA, known as Multimodal Sentiment Analysis (MSA), a significant field of study. MSA uses the most recent developments in deep learning and machine learning at several phases, such as sentiment polarity detection and multimodal feature extraction and fusion, with the goal of reducing error rates and enhancing performance. Multiple data sources, such as text, audio, and video, are used into MSA to improve sentiment classification accuracy. Using cutting-edge deep learning algorithms, this work integrates text, audio, and video characteristics to examine multimodal sentiment analysis. After outlining a framework for feature extraction, fusion, and data pre-processing, we assess the framework's performance against industry-standard benchmarks.

Multimodal aspect-based sentiment analysis (MABSA) is an essential task in the field of sentiment analysis, which still confronts several critical challenges. The first challenge is how to effectively capture key information within both image and text features to enhance the recognition and understanding of complex sentiment expressions. The second challenge is how to achieve cross-modal alignment of multi-granularity text features and image features. The third challenge is how to narrow the semantic gap between image modality and text modality through effective cross-modal feature fusion. To address these issues, a framework that leverages external knowledge and multi-granularity image and text features (EKMG) is proposed. Firstly, an external knowledge enhanced semantic extraction module is introduced to fuse external knowledge with image features and text features, thereby capturing the key information from texts and images. Secondly, we design a multi-granularity image-text contrastive learning module. This module initially introduces a graph attention network and a novel cross-modal fusion mechanism to align image features and text features at multiple granularities. Additionally, the module employs an image-text contrastive learning strategy to narrow the semantic gap between different modalities. Experimental results on two public benchmark datasets demonstrate that EKMG achieves significant performance improvements compared to state-of-the-art baseline models.

Multimodal sentiment analysis faces a number of challenges, including modality missing, modality heterogeneity gap, incomplete datasets, etc. Previous studies usually adopt schemes like meta-learning or multi-layer structures. Nevertheless, these methods lack interpretability for the interaction between modalities. In this paper, we constructed a new dataset, SM-MSD, for sentiment analysis in social media (SAS) that differs significantly from conventional corpora, comprising 10K instances of diverse data from Twitter, encompassing text, emoticons, emojis, and text embedded in images. This dataset aims to reflect authentic social scenarios and various emotional expressions, and provides a meaningful and challenging evaluation benchmark for multimodal sentiment analysis in specific contexts. Furthermore, we propose a multi-task framework based on heterogeneous graph neural networks (H-GNNs) and contrastive learning. For the first time, heterogeneous graph neural networks are applied to multimodal sentiment analysis tasks. In the case of additional labeling data, it guides the emotion prediction of the missing mode. We conduct extensive experiments on multiple datasets to verify the effectiveness of the proposed scheme. Experimental results demonstrate that our proposed scheme surpasses state-of-the-art methods by 1.7% and 0 in accuracy and 1.54% and 4.9% in F1-score on the MOSI and MOSEI datasets, respectively, and exhibits robustness to modality missing scenarios.

Multimodal sentiment analysis and emotion recognition represent a major research direction in natural language processing (NLP). With the rapid development of online media, people often express their emotions on a topic in the form of video, and the signals it transmits are multimodal, including language, visual, and audio. Therefore, the traditional unimodal sentiment analysis method is no longer applicable, which requires the establishment of a fusion model of multimodal information to obtain sentiment understanding. In previous studies, scholars used the feature vector cascade method when fusing multimodal data at each time step in the middle layer. This method puts each modal information in the same position and does not distinguish between strong modal information and weak modal information among multiple modalities. At the same time, this method does not pay attention to the embedding characteristics of multimodal signals across the time dimension. In response to the above problems, this paper proposes a new method and model for processing multimodal signals, which takes into account the delay and hysteresis characteristics of multimodal signals across the time dimension. The purpose is to obtain a multimodal fusion feature emotion analysis representation. We evaluate our method on the multimodal sentiment analysis benchmark dataset CMU Multimodal Opinion Sentiment and Emotion Intensity Corpus (CMU-MOSEI). We compare our proposed method with the state-of-the-art model and show excellent results.

Recent Transformer-based contextual word representations, including BERT and XLNet, have shown state-of-the-art performance in multiple disciplines within NLP. Fine-tuning the trained contextual models on task-specific datasets has been the key to achieving superior performance downstream. While fine-tuning these pre-trained models is straight-forward for lexical applications (applications with only language modality), it is not trivial for multimodal language (a growing area in NLP focused on modeling face-to-face communication). Pre-trained models don't have the necessary components to accept two extra modalities of vision and acoustic. In this paper, we proposed an attachment to BERT and XLNet called Multimodal Adaptation Gate (MAG). MAG allows BERT and XLNet to accept multimodal nonverbal data during fine-tuning. It does so by generating a shift to internal representation of BERT and XLNet; a shift that is conditioned on the visual and acoustic modalities. In our experiments, we study the commonly used CMU-MOSI and CMU-MOSEI datasets for multimodal sentiment analysis. Fine-tuning MAG-BERT and MAG-XLNet significantly boosts the sentiment analysis performance over previous baselines as well as language-only fine-tuning of BERT and XLNet. On the CMU-MOSI dataset, MAG-XLNet achieves human-level multimodal sentiment analysis performance for the first time in the NLP community.