Just having a surf before crashing and don't recall this article posted but...we find so many hey
We get our mention with GF towards the end. These guys think there's a positive future too
Updates on the semiconductor industry, technology breakthroughs, and the latest news on electronic component market trends and chip design.
www.sourcengine.com
The Future of AI Search and Neuromorphic Semiconductors Gain Success - March 3, 2023
Innovation is racing ahead. Now that the chip shortage is winding down and components are increasing in availability, researchers are picking up where they left off. Recent breakthroughs in neuromorphic semiconductors put these brain-like chips on the fast track to further tech integration. With artificial intelligent search features and chatbots, the smarter the chips power them, the more accurate they become.
Though, there might be storm clouds on the horizon for artificial intelligence (AI). Despite the recent popularity of artificial intelligence programs like
ChatGPT, further development might be hampered. This hinges on the decision of the U.S. Supreme Court’s two cases involving search algorithms. The result of which might not be fully realized until years into the future.
Neuromorphic Semiconductor Device Success and its Future
Chips are becoming more human at every turn. Dr. Yong-hun Kim and Dr. Jeong-Dae Kwon at the Surface & Nano Materials Division at the Korea Institute of Materials Science (KIMS) and their team have “
successfully developed the world’s first neuromorphic semiconductor device with high-density and high-reliability by developing a thin film of lithium-ion battery materials.” The new technology was produced by combining ultra-thin lithium ions with two-dimensional nanomaterials. The results of this research were published in late 2022 within
ACS Applied Materials & Interfaces.
What is a neuropathic semiconductor? To start, neuro is about the human body’s nervous system, including the brain, and its synaptic connections within its nerves. Neuropathic semiconductors have synapses and neurons that are like the human brain. With the ability to process and memorize information, the synaptic device receives signals from neurons, translates them by connection strength (synaptic weight), and stores the information. Each signal and the corresponding synaptic weight can enable pattern recognition.
This pattern recognition was used in further developing the neuropathic semiconductor. The device was implemented with an artificial neural network where its learning pattern uses the synaptic device to establish handwriting pattern recognition. Its artificial intelligence (AI) semiconductor was proven to have a
handwriting recognition rate of 96.77%.
Dr. Kim and Dr. Kwon’s research team plans to conduct follow-up research on other low-power artificial intelligence devices, like their neuromorphic semiconductors and wearable edging devices. “Our next-generation neuromorphic semiconductor device does not require CPU and memory,” the research team said. “It can simultaneously process and store information and learns and recognizes images such as handwriting patterns. It is expected to be applied to various low-power artificial intelligence devices such as world-class neuromorphic hardware systems, haptic devices, and vision sensors.”
This development comes as GlobalFoundries (GF) aids in the creation of next-gen vision and computing technologies like neuropathic semiconductors. BrainChip, the world’s first company to develop ultra-low-power, event-based neuromorphic AI IP, utilized GF technology to advance its neuromorphic tech further. GF’s low-leakage FD SOI platform was used for BrainChip’s AKD1500 chip. This collaboration will empower further advances in neuromorphic technology over the coming years, paired with continual research by teams like KIMS.
By partnering with large original chip manufacturers (OCMs), mass production of these smart, human-brain-like devices will become more commonplace in technology. When it does, it's sure to exciting growth in areas such as AI and machine learning.