Research My IBM Log in
Conference paper

Learning Convolutional Neural Networks with Deep Part Embeddings

Abstract

We propose a novel concept of Deep Part Embeddings (DPEs), which can be used to learn new Convolutional Neural Networks (CNNs) for different classes. We define DPE as a neuron of a trained CNN along with its network of filter activations that is interpretable as a part of a class that the neuron contributes to. Given a new class mathcal{C}, we explore the idea of combining different DPEs that intuitively constitute mathcal{C}, from trained CNNs (not on mathcal{C}), into a network that learns the class mathcal{C} with few training samples. An important application of our proposed framework is the ability to modify a CNN trained on n classes to learn a new class with limited training data without significantly affecting its performance on the n classes. We visually illustrate the different network architectures and extensively evaluate their performance against the baselines.

Semiconductors Artificial Intelligence Quantum Computing Hybrid Cloud About Publications Blog Events Careers Contact Research Topics People Projects Newsletter X LinkedIn YouTube RSS Contact IBM Privacy Terms of use Accessibility