Researchers have developed a new framework for deep neural networks that allows artificial intelligence (AI) systems to better learn new tasks while “forgetting” less of what they have learned regarding previous tasks.
The researchers have also demonstrated that using the framework to learn a new task can make the AI better at performing previous tasks, a phenomenon called backward transfer.
“People are capable of continual learning; we learn new tasks all the time, without forgetting what we already know,” says Tianfu Wu, an assistant professor of electrical and computer engineering at NC State and co-author of a paper on the work. “To date, AI systems using deep neural networks have not been very good at this.”
“Deep neural network AI systems are designed for learning narrow tasks,” says Xilai Li, a co-lead author of the paper and a Ph.D. candidate at NC State. “As a result, one of several things can happen when learning new tasks. Systems can forget old tasks when learning new ones, which is called catastrophic forgetting. Systems can forget some of the things they knew about old tasks, while not learning to do new ones as well. Or systems can fix old tasks in place while adding new tasks – which limits improvement and quickly leads to an AI system that is too large to operate efficiently. Continual learning, also called lifelong-learning or learning-to-learn, is trying to address the issue.”
“We have proposed a new framework for continual learning, which decouples network structure learning and model parameter learning,” says Yingbo Zhou, co-lead author of the paper and a research scientist at Salesforce Research. “We call it the Learn to Grow framework. In experimental testing, we’ve found that it outperforms previous approaches to continual learning.”
To understand the Learn to Grow framework, think of deep neural networks as a pipe filled with multiple layers. Raw data goes into the top of the pipe, and task outputs come out the bottom. Every “layer” in the pipe is a computation that manipulates the data in order to help the network accomplish its task, such as identifying objects in a digital image. There are multiple ways of arranging the layers in the pipe, which correspond to different “architectures” of the network.
When asking a deep neural network to learn a new task, the Learn to Grow framework begins by conducting something called an explicit neural architecture optimization via search. What this means is that as the network comes to each layer in its system, it can decide to do one of four things: skip the layer; use the layer in the same way that previous tasks used it; attach a lightweight adapter to the layer, which modifies it slightly; or create an entirely new layer.
This architecture optimization effectively lays out the best topology, or series of layers, needed to accomplish the new task. Once this is complete, the network uses the new topology to train itself on how to accomplish the task – just like any other deep learning AI system.
“We’ve run experiments using several datasets, and what we’ve found is that the more similar a new task is to previous tasks, the more overlap there is in terms of the existing layers that are kept to perform the new task,” Li says. “What is more interesting is that, with the optimized – or “learned” topology – a network trained to perform new tasks forgets very little of what it needed to perform the older tasks, even if the older tasks were not similar.”
The researchers also ran experiments comparing the Learn to Grow framework’s ability to learn new tasks to several other continual learning methods, and found that the Learn to Grow framework had better accuracy when completing new tasks.
To test how much each network may have forgotten when learning the new task, the researchers then tested each system’s accuracy at performing the older tasks – and the Learn to Grow framework again outperformed the other networks.
“In some cases, the Learn to Grow framework actually got better at performing the old tasks,” says Caiming Xiong, the research director of Salesforce Research and a co-author of the work. “This is called backward transfer, and occurs when you find that learning a new task makes you better at an old task. We see this in people all the time; not so much with AI.”
The Latest on: Continual learning for artificial intelligence
via Google News
The Latest on: Continual learning for artificial intelligence
- Artificial intelligence set to revolutionize cell imaging methodson December 8, 2019 at 3:05 am
As such, there is a continuous need to develop and refine cell imaging techniques ... described the novel method developed by his team called Artificial-Intelligence Microscopy Screening (AIMS). This ...
- Artificial intelligence will affect Utah more than other states, new study sayson December 7, 2019 at 9:08 pm
In this file photo from May 2, 2019, Cadet Cheyenne Quilter works with a virtual reality character named “Ellie” at the U.S. Military Academy at West Point, N.Y. Artificial intelligence is spreading ...
- Baidu Leads the Way in Innovation with 5,712 Artificial Intelligence Patent Applicationson December 6, 2019 at 10:39 pm
“Baidu retained the top spot for AI patent applications in China because of our continuous research and investment in developing AI ... The report showed that Baidu is the patent application leader in ...
- Changing education: Artificial Intelligence mobile roboton December 6, 2019 at 8:41 am
It is part of the Montour School District’s Artificial ... in grades K-4 are learning AI through computer science using Google’s AI Experiment programs, such as Teachable Machines, Pattern Radio, ...
- Accelerate Your Machine Learning and Artificial Intelligence Journey Using These DevOps Best Practiceson December 3, 2019 at 11:00 am
According to Gartner, "Technology innovation leaders are keen to apply DevOps principles for AI and ML projects, but they often struggle with architecting a solution for automating end-to-end ML ...
- New Artificial Intelligence and Machine Learning Platform for Reconciliation, Matching and Exception Management Operations Introduced by Broadridgeon December 2, 2019 at 11:12 pm
LONDON and NEW YORK, Dec. 3, 2019 /PRNewswire/ -- Broadridge Financial Solutions, Inc. (NYSE: BR), a global Fintech leader and part of the S&P 500® Index, today announced the launch of Broadridge Data ...
- Global Artificial Intelligence Market Scope 2019-2028 || Segmentation And Regional Studyon December 2, 2019 at 3:15 am
The research says Artificial Intelligence market has uncovered rapid growth in the ongoing and past years and is going to grow with a continuing development in the future years ... Segmentation on ...
- With Great Power Comes Great Responsibility: Artificial Intelligence in Bankingon December 1, 2019 at 10:04 pm
Artificial intelligence (AI), whether in the form of robotic process automation (RPA ... The challenges in explaining the use of data and continuous learning by AI processes to members of the public ...
- Innovations in Deep Learning, Artificial...on November 25, 2019 at 8:15 am
DUBLIN, Nov. 25, 2019 /PRNewswire/ -- The "Innovations in Deep Learning, Artificial Intelligence, IoT Security, Endpoint Security, Network Security, and Unified Data Security" report has been added to ...
- Seattle Seahawks Select AWS as Its Cloud, Machine Learning, and Artificial Intelligence Provideron November 25, 2019 at 6:00 am
SEATTLE--(Business Wire)--Today, Amazon Web Services, Inc. (AWS), an Amazon.com company (NASDAQ: AMZN), announced that AWS is now a cloud, machine learning (ML), and artificial intelligence (AI) ...
via Bing News