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
- CognitiveScale’s Cortex Certifai Wins at the Global Annual Achievement Awards for Artificial Intelligenceon November 18, 2019 at 4:29 pm
the world’s first automated scanner for black-box AI models that detects and scores vulnerabilities in most types of machine learning and statistical models, has won at the Global Annual Achievement ...
- How To Get Your Résumé Past The Artificial Intelligence Gatekeeperson November 18, 2019 at 6:34 am
Many enterprise businesses use Artificial Intelligence (AI) and machine learning tools to screen résumés when recruiting and hiring ... including data from outside your company and even your industry.
- Human–machine partnership with artificial intelligence for chest radiograph diagnosison November 18, 2019 at 2:18 am
It is important to note that these user inputs are not discrete votes, but continuous ... swarm artificial intelligence algorithm, generally took between 15 and 60 s. No swarm failed to reach an ...
- Using Artificial Intelligence to Revolutionize Art Discoveryon November 17, 2019 at 10:38 am
It promises to transform art discovery the way Shazam transformed music discovery. ArtPI is the first public API designed and optimized for art. It uses AI (artificial intelligence) and deep learning ...
- One way for the Pentagon to prove it’s serious about artificial intelligenceon November 15, 2019 at 3:12 pm
Department of Defense officials routinely talk about the need to more fully embrace machine learning and artificial intelligence, but one leader in the Marine Corps said ... it’s the result of ...
- New legal support firm pitches artificial intelligence to review documentson November 14, 2019 at 5:18 pm
A new legal services company has developed an artificial intelligence program to support lawyers during time-consuming ... One of the technologies used by iNof8 Legal, however, is electronic discovery ...
- HRS Uses Augmented Artificial Intelligence to Improve Corporate Hotel Rate Projectionson November 6, 2019 at 10:10 pm
HRS, the leading Global Hotel Solutions end-to-end technology provider in business travel, today introduced innovative hotel rate projection technology powered by Augmented Artificial Intelligence ...
- Due date delay for Application of Artificial Intelligence/Machine Learning Tools for NASA Science (RFI)on November 6, 2019 at 10:54 am
FDL is an applied artificial intelligence (AI) research accelerator leveraging the newest developments in AI and Machine Learning (ML) technologies from academia and the private sector and applying ...
- Machine Learning and Artificial Intelligence: Not New Concepts for the Data Science Practitioneron November 5, 2019 at 2:43 pm
Artificial intelligence (AI) has simply accelerated this process. Virtually every industry has been impacted by AI and certainly data science is no exception. Yet, we may also inquire how does machine ...
- Q&A: The Promise and Pitfalls of Artificial Intelligence and Personalized Learningon November 5, 2019 at 2:00 pm
But artificial intelligence has the potential to do much more, because true AI technologies are in a state of continuous learning, developing better strategies and tactics as they analyze more data.
via Bing News