Hardware-Aware Efficient Training of Deep Learning Models

Ghouthi BOUKLI HACENE · Vincent Gripon · Fran├žois Leduc-Primeau · Vahid Partovi Nia · Fan Yang · Andreas Moshovos · Yoshua Bengio

To reach top-tier performance, deep learning architectures usually rely on a large number of parameters and operations, and thus require to be processed using considerable power and memory. Numerous works have proposed to tackle this problem using quantization of parameters, pruning, clustering of parameters, decompositions of convolutions, or using distillation. However, most of these works aim at accelerating only the inference process and disregard the training phase. In practice, however, it is the learning phase that is by far the most complex. There has been recent efforts in introducing some compression on the training process, however, it remains challenging. In this workshop, we propose to focus on reducing the complexity of the training process. Our aim is to gather researchers interested in reducing energy, time, or memory usage for faster/cheaper/greener prototyping or deployment of deep learning models. Due to the dependence of deep learning on large computational capacities, the outcomes of the workshop could benefit all who deploy these solutions, including those who are not hardware specialists. Moreover, it would contribute to making deep learning more accessible to small businesses and small laboratories. Indeed, training complexity is of interest to many distinct communities. A first example is training on edge devices, where training can be used to specialize to data obtained online when the data cannot be transmitted back to the cloud because of constraints on privacy or communication bandwidth. Another example is accelerating training on dedicated hardware such as GPUs or TPUs.

Chat is not available.
Timezone: America/Los_Angeles »