Math @ Duke
|
Publications [#356190] of Vahid Tarokh
Papers Published
- Le, CP; Soltani, M; Dong, J; Tarokh, V, Fisher Task Distance and Its Application in Neural Architecture Search, vol. abs/2103.12827
(March, 2021)
(last updated on 2023/06/01)
Abstract: We formulate an asymmetric (or non-commutative) distance between tasks based
on Fisher Information Matrices, called Fisher task distance. This distance
represents the complexity of transferring the knowledge from one task to
another. We provide a proof of consistency for our distance through theorems
and experiments on various classification tasks from MNIST, CIFAR-10,
CIFAR-100, ImageNet, and Taskonomy datasets. Next, we construct an online
neural architecture search framework using the Fisher task distance, in which
we have access to the past learned tasks. By using the Fisher task distance, we
can identify the closest learned tasks to the target task, and utilize the
knowledge learned from these related tasks for the target task. Here, we show
how the proposed distance between a target task and a set of learned tasks can
be used to reduce the neural architecture search space for the target task. The
complexity reduction in search space for task-specific architectures is
achieved by building on the optimized architectures for similar tasks instead
of doing a full search and without using this side information. Experimental
results for tasks in MNIST, CIFAR-10, CIFAR-100, ImageNet datasets demonstrate
the efficacy of the proposed approach and its improvements, in terms of the
performance and the number of parameters, over other gradient-based search
methods, such as ENAS, DARTS, PC-DARTS.
|
|
dept@math.duke.edu
ph: 919.660.2800
fax: 919.660.2821
| |
Mathematics Department
Duke University, Box 90320
Durham, NC 27708-0320
|
|