{"id":57884,"date":"2024-12-18T15:30:13","date_gmt":"2024-12-18T15:30:13","guid":{"rendered":"https:\/\/www.threatshub.org\/blog\/boffins-trick-ai-model-into-giving-up-its-secrets\/"},"modified":"2024-12-18T15:30:13","modified_gmt":"2024-12-18T15:30:13","slug":"boffins-trick-ai-model-into-giving-up-its-secrets","status":"publish","type":"post","link":"https:\/\/www.threatshub.org\/blog\/boffins-trick-ai-model-into-giving-up-its-secrets\/","title":{"rendered":"Boffins trick AI model into giving up its secrets"},"content":{"rendered":"

Computer scientists from North Carolina State University have devised a way to copy AI models running on Google Edge Tensor Processing Units (TPUs), as used in Google Pixel phones and third-party machine learning accelerators.<\/p>\n

The technique, developed by NC State researchers Ashley Kurian, Anuj Dubey, Ferhat Yaman and Aydin Aysu, is a side-channel attack that measures the electromagnetic intensity of AI model use (inference) when running on TPUs, and exploits those measurements to infer model hyperparameters.<\/p>\n

Machine learning model hyperparameters<\/a> refer to values set prior to the training process that affect model training \u2013 the learning rate, the batch size, or the pool size. They’re distinct from model parameters \u2013 such as weights \u2013 which are internal to the model and are learned during training.<\/p>\n