Mardi 28 Novembre 2023
- Imprimer
- Partager
- Partager sur Facebook
- Share on X
- Partager sur LinkedIn
Affective Computing: The Endgame?
Abstract:
Roughly three decades of Affective Computing have enabled computing systems to recognize, generate, and respond to emotions of humans in audio, language, video, and further modalities. Over years, this was realized by collecting target domain data, designing features for representation, and architectures for machine learning. A first disruption of this process came with the advent of end-to-end learning of representations and architectures of machine learning approaches directly from data. Likewise, the frontiers between modalities have been increasingly washed away, as Affective Computing solutions could more and more be realized by experts of machine learning without particular knowledge in specific modalities: deep learning helped to learn the “features” and neural architecture search increasingly helps to learn the optimal hyperparameters alongside the parameters of a model. Just now, however, a second disruption emerges: The advent of large “foundation” models and their zero-shot learning ability. In other words, with large audio, language, vision, and further models, one can observe emergent Affective Computing by these models directly, without even a need for specific training data. I will discuss these disruptions and the “new age” of Affective Computing to come: our efforts may shift to prompt design, optimal fusion of large models and traditional approaches, and facing new challenges as the "nightshades“ will start to poison tomorrow’s large models. With all this comes a whole new opportunity to finally deploy fully multimodal and more accurate than ever before Affective Computing at scale
Biography:
Björn W. Schuller received his diploma, doctoral degree, habilitation, and Adjunct Teaching Professor in Machine Intelligence and Signal Processing all in EE/IT from Technical University of Munich (TUM) in Munich, Germany where he has been named Full Professor of Health Informatics. He is also Full Professor of Artificial Intelligence and the Head of GLAM at Imperial College London/UK, Full Professor and Chair of Embedded Intelligence for Health Care and Wellbeing at the University of Augsburg/Germany, co-founding CEO and current CSO of audEERING amongst other Professorships and Affiliations. Previous stays include Full Professor at the University of Passau/Germany, Key Researcher at Joanneum Research in Graz/Austria, and the CNRS-LIMSI in Orsay/France.
Björn W. Schuller is a Fellow of the IEEE and Golden Core Awardee of the IEEE Computer Society, Fellow of the BCS, Fellow of the ELLIS, Fellow of the ISCA, Fellow and President-Emeritus of the AAAC, Elected Full Member Sigma Xi, and Senior Member of the ACM. He (co-) authored 1,200+ publications (50,000+ citations, h-index=100+), is Field Chief Editor of Frontiers in Digital Health and was Editor in Chief of the IEEE Transactions on Affective Computing amongst manifold further commitments and service to the community. His 50+ awards include IEEE Signal Processing Society Distinguished Lecturer 2024.
Roughly three decades of Affective Computing have enabled computing systems to recognize, generate, and respond to emotions of humans in audio, language, video, and further modalities. Over years, this was realized by collecting target domain data, designing features for representation, and architectures for machine learning. A first disruption of this process came with the advent of end-to-end learning of representations and architectures of machine learning approaches directly from data. Likewise, the frontiers between modalities have been increasingly washed away, as Affective Computing solutions could more and more be realized by experts of machine learning without particular knowledge in specific modalities: deep learning helped to learn the “features” and neural architecture search increasingly helps to learn the optimal hyperparameters alongside the parameters of a model. Just now, however, a second disruption emerges: The advent of large “foundation” models and their zero-shot learning ability. In other words, with large audio, language, vision, and further models, one can observe emergent Affective Computing by these models directly, without even a need for specific training data. I will discuss these disruptions and the “new age” of Affective Computing to come: our efforts may shift to prompt design, optimal fusion of large models and traditional approaches, and facing new challenges as the "nightshades“ will start to poison tomorrow’s large models. With all this comes a whole new opportunity to finally deploy fully multimodal and more accurate than ever before Affective Computing at scale
Biography:
Björn W. Schuller received his diploma, doctoral degree, habilitation, and Adjunct Teaching Professor in Machine Intelligence and Signal Processing all in EE/IT from Technical University of Munich (TUM) in Munich, Germany where he has been named Full Professor of Health Informatics. He is also Full Professor of Artificial Intelligence and the Head of GLAM at Imperial College London/UK, Full Professor and Chair of Embedded Intelligence for Health Care and Wellbeing at the University of Augsburg/Germany, co-founding CEO and current CSO of audEERING amongst other Professorships and Affiliations. Previous stays include Full Professor at the University of Passau/Germany, Key Researcher at Joanneum Research in Graz/Austria, and the CNRS-LIMSI in Orsay/France.
Björn W. Schuller is a Fellow of the IEEE and Golden Core Awardee of the IEEE Computer Society, Fellow of the BCS, Fellow of the ELLIS, Fellow of the ISCA, Fellow and President-Emeritus of the AAAC, Elected Full Member Sigma Xi, and Senior Member of the ACM. He (co-) authored 1,200+ publications (50,000+ citations, h-index=100+), is Field Chief Editor of Frontiers in Digital Health and was Editor in Chief of the IEEE Transactions on Affective Computing amongst manifold further commitments and service to the community. His 50+ awards include IEEE Signal Processing Society Distinguished Lecturer 2024.
Date et Lieu
Mardi 28 Novembre 2023 à 10h00
Bâtiment IMAG - Amphithéâtre
Bâtiment IMAG - Amphithéâtre
Organisé par
Dominique VAUFREYDAZ
Responsable de l'équipe M-PSI
Responsable de l'équipe M-PSI
- Imprimer
- Partager
- Partager sur Facebook
- Share on X
- Partager sur LinkedIn