Deep Learning Classification and model compression

Deep neural networks (DNNs) have achieved great success in many visual recognition tasks. However, existing state of the art deep neural network models are computationally expensive and memory intensive, hindering their deployment in devices with low memory resources or in applications with strict latency requirements. Therefore, a natural thought is to perform model compression and acceleration in deep networks without significantly decreasing the model performance. Developing models for inference on clients has multiple economical benefits, but it becomes difficult to match the performance of bigger architectures by simply training smaller architectures. Therefore, we have to look for solutions like Knowledge Distillation, Network Pruning, Quantization to obtain highly efficient models that can match the performance of bigger architectures.

Faculty Supervisor:

Ioannis Mitliagkas

Student:

Partner:

Jumio

Discipline:

Computer science

Sector:

Professional, scientific and technical services

University:

Université de Montréal

Program: