Teach Neural Networks to identify sequences of values
First Things First
2, 5, 7, 10, 12 – and what number comes next? Mike Schilli tests whether intelligence tests devised by psychologists can be cracked with modern AI Networks.
Neural networks do great things when it comes to detecting patterns in noisy input data and assigning unambiguous results to them. If a dozen people with different handwriting enter the letters A or B in a form, a trained network can identify with almost 100 percent certainty what they wrote. Or consider pattern recognition systems for identifying the license plates of passing vehicles: Aren't these technical miracles? They extract the digits from a camera feed so that the Department of Transportation knows exactly who is going where.
Once a neural network is done learning, it always assigns the same result to the same input data, but when it comes to tasks that need to determine the next value in time-discrete value sequences, neural networks often fail to deliver perfect results, especially if the input signal is subject to variations of unknown periodicity.
In a neural network, the learning algorithm adjusts internal weights based on the training data. However, once these weights are determined, they won't change anymore at run time and thus cannot account for temporal changes in the input data, because the machine doesn't remember any previous state. Recurrent neural networks (RNNs) maintain internal connections back to the input, and thus a result can influence the next input vector, but this does not help a simple network identify temporal patterns that extend over several cycles.
[...]
Buy this article as PDF
(incl. VAT)
Buy Linux Magazine
US / Canada
UK / Australia
Subscribe to our Linux Newsletters
Find Linux and Open Source Jobs
Subscribe to our ADMIN Newsletters
Support Our Work
Linux Magazine content is made possible with support from readers like you. Please consider contributing when you’ve found an article to be beneficial.
News
-
Fedora Continues 32-Bit Support
In a move that should come as a relief to some portions of the Linux community, Fedora will continue supporting 32-bit architecture.
-
Linux Kernel 6.17 Drops bcachefs
After a clash over some late fixes and disagreements between bcachefs's lead developer and Linus Torvalds, bachefs is out.
-
ONLYOFFICE v9 Embraces AI
Like nearly all office suites on the market (except LibreOffice), ONLYOFFICE has decided to go the AI route.
-
Two Local Privilege Escalation Flaws Discovered in Linux
Qualys researchers have discovered two local privilege escalation vulnerabilities that allow hackers to gain root privileges on major Linux distributions.
-
New TUXEDO InfinityBook Pro Powered by AMD Ryzen AI 300
The TUXEDO InfinityBook Pro 14 Gen10 offers serious power that is ready for your business, development, or entertainment needs.
-
Danish Ministry of Digital Affairs Transitions to Linux
Another major organization has decided to kick Microsoft Windows and Office to the curb in favor of Linux.
-
Linux Mint 20 Reaches EOL
With Linux Mint 20 at its end of life, the time has arrived to upgrade to Linux Mint 22.
-
TuxCare Announces Support for AlmaLinux 9.2
Thanks to TuxCare, AlmaLinux 9.2 (and soon version 9.6) now enjoys years of ongoing patching and compliance.
-
Go-Based Botnet Attacking IoT Devices
Using an SSH credential brute-force attack, the Go-based PumaBot is exploiting IoT devices everywhere.
-
Plasma 6.5 Promises Better Memory Optimization
With the stable Plasma 6.4 on the horizon, KDE has a few new tricks up its sleeve for Plasma 6.5.