SDSC Explores Deep Learning Multimodal Models to Enhance Ear