2016/08/26 by Valentina Gregori, Gregori, Valentina, Barbara Caputo +1
Engineering · Neuroscience · #EEG and Brain-Computer Interfaces #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Muscle activation and electromyography studies #Neuroscience and Neural Engineering
paper · pdf · doi:10.48550/arxiv.1608.07536
openalex publication_date 2016/08/26 · openalex created_date 2022/10/03 · openalex updated_date 2026/07/28
Non-invasive myoelectric prostheses require a long training time to obtain\nsatisfactory control dexterity. These training times could possibly be reduced\nby leveraging over training efforts by previous subjects. So-called domain\nadaptation algorithms formalize this strategy and have indeed been shown to\nsignificantly reduce the amount of required training data for intact subjects\nfor myoelectric movements classification. It is not clear, however, whether\nthese results extend also to amputees and, if so, whether prior information\nfrom amputees and intact subjects is equally useful. To overcome this problem,\nwe evaluated several domain adaptation algorithms on data coming from both\namputees and intact subjects. Our findings indicate that: (1) the use of\nprevious experience from other subjects allows us to reduce the training time\nby about an order of magnitude; (2) this improvement holds regardless of\nwhether an amputee exploits previous information from other amputees or from\nintact subjects.\n