REMIND Your Neural Network to Prevent Catastrophic Forgetting

10/06/2019
by   Tyler L. Hayes, et al.
34

In lifelong machine learning, a robotic agent must be incrementally updated with new knowledge, instead of having distinct train and deployment phases. Conventional neural networks are often used for interpreting sensor data, however, if they are updated on non-stationary data streams, they suffer from catastrophic forgetting, with new learning overwriting past knowledge. A common remedy is replay, which involves mixing old examples with new ones. For incrementally training convolutional neural network models, prior work has enabled replay by storing raw images, but this is memory intensive and not ideal for embedded agents. Here, we propose REMIND, a tensor quantization approach that enables efficient replay with tensors. Unlike other methods, REMIND is trained in a streaming manner, meaning it learns one example at a time rather than in large batches containing multiple classes. Our approach achieves state-of-the-art results for incremental class learning on the ImageNet-1K dataset. We also probe REMIND's robustness to different data ordering schemes using the CORe50 streaming dataset. We demonstrate REMIND's generality by pioneering multi-modal incremental learning for visual question answering (VQA), which cannot be readily done with comparison models. We establish strong baselines on the CLEVR and TDIUC datasets for VQA. The generality of REMIND for multi-modal tasks can enable robotic agents to learn about their visual environment using natural language understanding in an interactive way.

READ FULL TEXT

page 1

page 11

research
09/16/2018

Memory Efficient Experience Replay for Streaming Learning

In supervised machine learning, an agent is typically trained once and t...
research
08/04/2023

Class Incremental Learning with Self-Supervised Pre-Training and Prototype Learning

Deep Neural Network (DNN) has achieved great success on datasets of clos...
research
04/20/2020

Generative Feature Replay For Class-Incremental Learning

Humans are capable of learning new tasks without forgetting previous one...
research
09/04/2019

Lifelong Machine Learning with Deep Streaming Linear Discriminant Analysis

When a robot acquires new information, ideally it would immediately be c...
research
08/14/2020

RODEO: Replay for Online Object Detection

Humans can incrementally learn to do new visual detection tasks, which i...
research
09/21/2022

Continual VQA for Disaster Response Systems

Visual Question Answering (VQA) is a multi-modal task that involves answ...
research
08/30/2022

BioSLAM: A Bio-inspired Lifelong Memory System for General Place Recognition

We present BioSLAM, a lifelong SLAM framework for learning various new a...

Please sign up or login with your details

Forgot password? Click here to reset