EE Seminar: Juggler: Multitask Learning with Task Performance Constraints

הרישום לסמינר יבוצע באמצעות סריקת הברקוד למודל (יש להיכנס לפני כן למודל,  לא באמצעות האפליקציה)- הרישום מתחיל ב- 15:30 ומסתיים ב- 15:40

Registration to the seminar will be done by scanning the barcode for the Moodle (Please enter ahead to the Moodle, NOT by application)- Registration starts at 15:30 and ends at 15:40

Electrical Engineering Systems Seminar

Speaker: Ella Avidan

M.Sc. student under the supervision of Dr. Ilai Bistritz

Wednesday, 12th June 2025, at 15:30

Room 011, Kitot Building, Faculty of Engineering

Juggler: Multitask Learning with Task Performance Constraints

Consider multitask learning (MTL), in which $N$ models share parts of their architecture (e.g., a backbone with a head for each task). Our goal is to train the overall model so that the training loss of each task falls below a given threshold. However, we may harm others when varying the shared parameters to help one task. A weighted total loss could balance between the different tasks to achieve the target threshold. Nevertheless, the weights that correctly balance the tasks are unknown in advance. To overcome that, we propose a scheme that divides the total training time into epochs of increasing length. The scheme adjusts the weights every epoch based on the performance of the tasks at the end of the epoch. We prove that our scheme asymptotically converges to a model that satisfies the target loss constraints (if feasible), provided the learning rate, control step size, and epoch lengths are properly tuned. We experiment on deep neural networks to demonstrate that our scheme is effective even beyond our theoretical assumptions.