We are conducting research on multimodal IoT, which collects time-series multimodal data, and multimodal AI, which combines features from videos and images with environmental and other data for machine learning. Particularly in the agricultural field, even if IoT could collect time-series data over several years, pre-processing to account for imbalance and time-series is very important. Since the available data is limited, we are also developing data augmentation, data systhesis, and data generation with generative AI for such time-series and image data using evaluation metrics for feature similarity for domain adaptation.

In any case, we are working on research and development of general-purpose technologies related to time-series multimodal AI/IoT, forming multiple groups to handle raw multimodal data. The knowledge and experience gained through the research and development of each technology can be applied to various fields. We hope that you will open up new fields from the aspect of computer science.