Presented By O'Reilly and Cloudera
December 5-6, 2016: Training
December 6–8, 2016: Tutorials & Conference
Singapore

Transfer learning and fine-tuning deep neural network models across different domains

Anusua Trivedi (Microsoft)
4:15pm–4:55pm Thursday, December 8, 2016
Chat, machine learning, and AI
Location: 328/329 Level: Advanced
Average rating: ***..
(3.00, 1 rating)

Prerequisite Knowledge

  • A basic understanding of machine learning

What you'll learn

  • Learn how transfer learning and fine-tuning strategy leads to reusability of the same convolution neural network model in different disjoint domains

Description

Anusua Trivedi proposes a method to apply a pretrained deep convolution neural network (DCNN) on different types of images datasets—a fluorescein angiographic eye image to predict diabetic retinopathy and a fashion image to predict the clothing type in that image—to improve prediction accuracy. This approach improves prediction accuracy on domain-specific image datasets compared to state-of-the-art machine-learning approaches.

Topics include:

  • A brief introduction to deep learning
  • Popular deep learning models and libraries
  • The motivation behind using deep learning models for images
  • Transfer learning and fine-tuning DCNNs
  • Deep learning models for image classification
  • Deep learning models for image tag prediction
  • Deep learning models for image caption generation
Photo of Anusua Trivedi

Anusua Trivedi

Microsoft

Anusua Trivedi is a data scientist on Microsoft’s Advanced Data Science & Strategic Initiatives team, where she works on developing advanced predictive analytics and deep learning models. Prior to joining Microsoft, Anusua was a data scientist at the Texas Advanced Computing Center (TACC), a supercomputer center, where she developed algorithms and methods for the supercomputer to explore, analyze, and visualize clinical and biological big data. Anusua is a frequent speaker at machine learning and big data conferences across the United States, including Supercomputing 2015 (SC15), PyData Seattle 2015, and MLconf Atlanta 2015. Anusua has also held positions with UT Austin and University of Utah.