Human-in-the-loop Design with Machine Learning
Event details
| Date | 12.02.2020 |
| Hour | 09:00 › 10:00 |
| Speaker | Ms Pan Wang, Dyson School of Design Engineering and Data Science Institute of Imperial College London |
| Location | |
| Category | Conferences - Seminars |
Abstract:
Generating design via machine learning has been an on-going challenge in data-driven engineering design. Current deep learning methods have been applied to randomly generate images, such as in fashion, furniture, and product design. Such deep generative methods typically require a large number of images for image generation. However, human aspects which can play a pivotal role in a design process are not taken into account in this kind of automated design process. In this talk, I will explain how human cognitive activity works during the design process. I will present our recent approach, which is a human-in-the-loop AI design framework which shows a way to involve human cognitive factors represented by brain activity (EEG & fMRI) in the generative process, specifically, according to different cognitive mechanisms that happen during the design process. To illustrate our results, I will present three examples of using cognitive models for mental association, preference, and conceptual blending in design cases. I will also talk about the advantages of these approaches in future applications such as creative design tools, virtual reality, medical devices, future robotics by supporting decision making, scheme optimization and co-creation.
Bio:
Pan Wang is a PhD from the Dyson School of Design Engineering and Data Science Institute at Imperial College London under the supervision of Prof. Peter Childs (FREng/FIMechE/FASME) and Prof. Yike Guo (FREng/MAE). She conducts research on deep learning and brain decoding in data-driven engineering design. Her research addresses various topic relating to AI design method, human-data Interaction, data-driven engineering design, brain-computer interface, design neurocognition, and user-centred system. Her research ‘Human-in-the-loop design with machine learning’ has won the distinguished paper award ICED 2019. She has also won several top international design competitions (Winner of Electrolux Design Lab 2014). She founded a start-up company for digital health since 2015.
Generating design via machine learning has been an on-going challenge in data-driven engineering design. Current deep learning methods have been applied to randomly generate images, such as in fashion, furniture, and product design. Such deep generative methods typically require a large number of images for image generation. However, human aspects which can play a pivotal role in a design process are not taken into account in this kind of automated design process. In this talk, I will explain how human cognitive activity works during the design process. I will present our recent approach, which is a human-in-the-loop AI design framework which shows a way to involve human cognitive factors represented by brain activity (EEG & fMRI) in the generative process, specifically, according to different cognitive mechanisms that happen during the design process. To illustrate our results, I will present three examples of using cognitive models for mental association, preference, and conceptual blending in design cases. I will also talk about the advantages of these approaches in future applications such as creative design tools, virtual reality, medical devices, future robotics by supporting decision making, scheme optimization and co-creation.
Bio:
Pan Wang is a PhD from the Dyson School of Design Engineering and Data Science Institute at Imperial College London under the supervision of Prof. Peter Childs (FREng/FIMechE/FASME) and Prof. Yike Guo (FREng/MAE). She conducts research on deep learning and brain decoding in data-driven engineering design. Her research addresses various topic relating to AI design method, human-data Interaction, data-driven engineering design, brain-computer interface, design neurocognition, and user-centred system. Her research ‘Human-in-the-loop design with machine learning’ has won the distinguished paper award ICED 2019. She has also won several top international design competitions (Winner of Electrolux Design Lab 2014). She founded a start-up company for digital health since 2015.
Practical information
- General public
- Free