Related references
Note: Only part of the references are listed.
Article
Computer Science, Artificial Intelligence
Kai Han et al.
Summary: Transformer, a deep neural network with a self-attention mechanism, has been initially used in natural language processing and is now gaining attention in computer vision tasks. Transformer-based models perform as well as or even better than convolutional and recurrent neural networks in various visual benchmarks. This paper reviews vision transformer models, categorizes them based on different tasks, and analyzes their advantages and disadvantages. The discussed categories include backbone network, high/mid-level vision, low-level vision, and video processing. Efficient methods for applying transformer in real device-based applications are also explored. The challenges and further research directions for vision transformers are discussed as well.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2023)
Article
Computer Science, Software Engineering
Jie M. Zhang et al.
Summary: This paper provides a comprehensive survey of techniques for testing machine learning systems and analyzes trends and challenges in ML testing, offering promising research directions for the future.
IEEE TRANSACTIONS ON SOFTWARE ENGINEERING
(2022)
Article
Acoustics
Bagus Tris Atmaja et al.
Summary: This paper presents a survey on bimodal emotion recognition which combines acoustic and linguistic information. It reviews five components of bimodal SER and presents major findings from commonly used datasets. The survey also proposes future research directions in this field.
SPEECH COMMUNICATION
(2022)
Proceedings Paper
Acoustics
Andreas Triantafyllopoulos et al.
Summary: This study investigates the utilization of linguistic information in speech emotion recognition using pre-trained neural networks. The findings suggest that transformer models can effectively leverage linguistic information to improve valence predictions.
Article
Computer Science, Artificial Intelligence
Kusha Sridhar et al.
Summary: Predicting valence from speech is an important and challenging problem. This study proposes an unsupervised approach to adapt the speech emotion recognition system to the target speakers in the test set. By searching for speakers with similar acoustic patterns and creating an adaptation set, the models can be personalized. The results show significant improvements in valence prediction using these unsupervised approaches.
IEEE TRANSACTIONS ON AFFECTIVE COMPUTING
(2022)
Article
Computer Science, Artificial Intelligence
Anton Batliner et al.
Summary: With the rapid development of artificial intelligence, ethical considerations have gained increasing attention. However, there is still insufficient focus on ethical issues in the field of computational paralinguistics. This article provides an overview of ethics and privacy, describes the field of computational paralinguistics and its applications, and proposes guidelines for good practice, establishing a foundation for ethical standards in the field.
IEEE TRANSACTIONS ON AFFECTIVE COMPUTING
(2022)
Article
Engineering, Electrical & Electronic
Jiaee Cheong et al.
Summary: The increasing prevalence of facial analysis technology has raised concerns about bias in the tools. Despite efforts to address bias, understanding, investigating, and mitigating bias in facial affect analysis remain underexplored areas. This work provides an overview of bias definitions, measures of fairness, algorithms, and techniques in facial affective signal processing, while discussing opportunities for further research.
IEEE SIGNAL PROCESSING MAGAZINE
(2021)
Article
Computer Science, Artificial Intelligence
Jean Kossaifi et al.
Summary: Natural human-computer interaction and audio-visual human behaviour sensing systems are more important than ever, as digital devices become increasingly integral to our lives. The SEWA database provides a valuable resource with over 2000 minutes of audio-visual data from 398 individuals representing six cultures, aiding research in affective computing and automatic human sensing.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2021)
Proceedings Paper
Audiology & Speech-Language Pathology
Shu-wen Yang et al.
Summary: Self-supervised learning is important in advancing research in NLP and CV, but there is a lack of a similar setup for speech processing. The SUPERB benchmark is introduced to evaluate the performance of a shared model across various speech tasks, with a focus on utilizing representations learned from SSL. Results show promising generalizability and accessibility of SSL representations across SUPERB tasks.
Proceedings Paper
Audiology & Speech-Language Pathology
Leonardo Pepino et al.
Summary: The study proposes a transfer learning method for speech emotion recognition, utilizing features extracted from pre-trained models to recognize emotions in speech. Evaluations are conducted on two standard emotion databases, comparing different feature extraction techniques and model architectures.
Proceedings Paper
Audiology & Speech-Language Pathology
Wei-Ning Hsu et al.
Summary: In this paper, we explore the self-supervised learning of speech representations and find that using target domain data during pre-training significantly improves performance. Pre-training in different domains enhances the generalization performance on domains not seen during training.
Proceedings Paper
Computer Science, Artificial Intelligence
Mariana Rodrigues Makiuchi et al.
Summary: The study proposed a novel cross-representation speech model and a CNN-based text emotion recognition model, addressing the issues of overfitting and learning based on superficial cues in emotion recognition tasks. By combining speech-based and text-based results using score fusion, the method surpassed current works on speech-only, text-only, and multimodal emotion recognition on the IEMOCAP dataset.
2021 IEEE AUTOMATIC SPEECH RECOGNITION AND UNDERSTANDING WORKSHOP (ASRU)
(2021)
Article
Acoustics
Wei-Ning Hsu et al.
Summary: HuBERT is a self-supervised approach for speech representation learning, leveraging an offline clustering step to align target labels for BERT-like prediction loss and applying prediction loss only over masked regions to force the model to learn a combined acoustic and language model. Experimental results on various fine-tuning subsets of Librispeech and Libri-light benchmarks demonstrate the superiority of the HuBERT model.
IEEE-ACM TRANSACTIONS ON AUDIO SPEECH AND LANGUAGE PROCESSING
(2021)
Proceedings Paper
Acoustics
Andreas Triantafyllopoulos et al.
Summary: With the rise of deep learning, deep knowledge transfer has become one of the most effective techniques for achieving state-of-the-art performance using deep neural networks. The choice of pre-training task and acoustic condition differences between datasets influence the effectiveness of transfer learning in speech emotion recognition. Layers closer to the input show more adaptation during transfer learning, explaining the need to fine-tune all layers in previous works.
2021 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP 2021)
(2021)
Proceedings Paper
Acoustics
Mao Li et al.
Summary: This study investigates how unsupervised representation learning on unlabeled datasets can benefit speech emotion recognition. The experiment results show that using the contrastive predictive coding method can significantly improve emotion recognition performance.
2021 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP 2021)
(2021)
Proceedings Paper
Acoustics
Raghavendra Pappagari et al.
Summary: Data augmentation is a widely used strategy for training robust machine learning models, and CopyPaste, a novel augmentation procedure proposed in this study, has shown significant improvements in speech emotion recognition. By concatenating utterances with different emotions, the model performance can be enhanced, particularly in noisy test conditions.
2021 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP 2021)
(2021)
Proceedings Paper
Acoustics
Leonardo Pepino et al.
2020 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING
(2020)
Proceedings Paper
Acoustics
Andy T. Liu et al.
2020 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING
(2020)
Article
Acoustics
Qiuqiang Kong et al.
IEEE-ACM TRANSACTIONS ON AUDIO SPEECH AND LANGUAGE PROCESSING
(2020)
Article
Computer Science, Artificial Intelligence
Reza Lotfian et al.
IEEE TRANSACTIONS ON AFFECTIVE COMPUTING
(2019)
Article
Physics, Multidisciplinary
Jack Fitzsimons et al.
Review
Computer Science, Hardware & Architecture
Bjoern W. Schuller
COMMUNICATIONS OF THE ACM
(2018)
Proceedings Paper
Computer Science, Artificial Intelligence
Felix Burkhardt et al.
LANGUAGE TECHNOLOGIES FOR THE CHALLENGES OF THE DIGITAL AGE, GSCL 2017
(2018)
Proceedings Paper
Computer Science, Artificial Intelligence
Fabien Ringeval et al.
PROCEEDINGS OF THE 2018 AUDIO/VISUAL EMOTION CHALLENGE AND WORKSHOP (AVEC'18)
(2018)
Article
Robotics
Ognjen Rudovic et al.
Editorial Material
Computer Science, Artificial Intelligence
Amir Zadeh et al.
IEEE INTELLIGENT SYSTEMS
(2016)
Proceedings Paper
Computer Science, Information Systems
Karol J. Piczak
MM'15: PROCEEDINGS OF THE 2015 ACM MULTIMEDIA CONFERENCE
(2015)
Article
Computer Science, Artificial Intelligence
Myriam Munezero et al.
IEEE TRANSACTIONS ON AFFECTIVE COMPUTING
(2014)
Review
Computer Science, Artificial Intelligence
Rafael A. Calvo et al.
IEEE TRANSACTIONS ON AFFECTIVE COMPUTING
(2010)
Review
Computer Science, Artificial Intelligence
Zhihong Zeng et al.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2009)
Article
Computer Science, Interdisciplinary Applications
Carlos Busso et al.
LANGUAGE RESOURCES AND EVALUATION
(2008)
