Intel officially joined the ranks of Google, Microsoft, Facebook, and other tech companies in the artificial intelligence arena by unveiling its first AI chip.
On Monday, the well-known CPU builder announced the arrival of Loihi, Intel’s AI chip reportedly designed to mimic human brain functions. Loihi, named after a subsea volcano in Hawaii, consists of 128 computing cores.
Each core is comprised of around 1,024 artificial neurons which give the AI chip a total of over 130,000 neurons and 130 million synaptic connections. Apparently, this makes Loihi more complicated than some small-brained animals but is still billions of neurons away from a real human brain.#Intel unveiled its first #AI chip, Loihi! Said its potential is limitless! Click To Tweet
A human brain approximately has over 80 billion neurons. So, doing the math, Loihi has to have over 78 million computing cores to function like a real brain. However, developing such high-capacity chip might take researchers years to accomplish.
Loihi: The AI Chip Designed to Mimic the Brain
Instead of relying on the potentials of raw computing, Loihi makes use of a yet to be proven kind of technology called ‘neuromorphic‘ that was said to be designed with inspiration from the human brain.
For years, Intel spent time and effort exploring the possibilities neuromorphic tech could bring. In fact, in 2012, the company designed a concept of neurochip that has valves which only respond to the spin of an electron and memristors as permanent storage. So, instead of ‘logic gates,’ this neurochip uses spiking neurons as its fundamental computing unit.
In essence, a human brain uses pulses or spikes to relay information to different parts of the body. Then, it strengthens frequent connections and stores the changes in local synapse interconnections. That said, brain cells couldn’t function alone because each neuron’s activity has a direct effect on other neurons.
Apparently, learning and intelligence could only be achieved if these groups of cells work together. In a blog post published by Dr. Michael Mayberry, Intel’s Corporate Vice President and Managing Director of Intel Labs, he explained:
“We believe AI is in its infancy and more architectures and methods – like Loihi – will continue emerging that raise the bar for AI. Neuromorphic computing draws inspiration from our current understanding of the brain’s architecture and its associated computations. The brain’s neural networks relay information with pulses or spikes, modulate the synaptic strengths or weight of the interconnections based on timing of these spikes, and store these changes locally at the interconnections. Intelligent behaviors emerge from the cooperative and competitive interactions between multiple regions within the brain’s neural networks and its environment.”
By mimicking this brain function, Loihi is capable of speeding up machine learning without the need for too much horsepower. It could also accelerate decision making while autonomously solving societal and industrial problems using learned experiences over time.
Right now, Loihi is still a research project. However, Mayberry said that a pinkie-sized first full version of the AI chip would be fabricated on November. Some academics and research institution would get to test the self-learning chip next year. Intel has already produced two prototypes while working on the full version. According to Mayberry, if Loihi successfully sparks interest, the company could get it to the market in two years time.
The Potentials of the Loihi AI Chip
As per Intel, AI chips like Loihi have limitless potential benefits. They said:
“One example provides a person’s heartbeat reading under various conditions – after jogging, following a meal or before going to bed – to a neuromorphic-based system that parses the data to determine a “normal” heartbeat. The system can then continuously monitor incoming heart data in order to flag patterns that do not match the “normal” pattern. The system could be personalized for any user.”
Furthermore, researchers at Intel also believe that the chip could also be used in current devices that need to learn in real time. These devices may include autonomous drones, cars, cameras, and traffic stop lights.
“The self-learning capabilities prototyped by this test chip have enormous potential to improve automotive and industrial applications as well as personal robotics – any application that would benefit from autonomous operation and continuous learning in an unstructured environment. For example, recognizing the movement of a car or bike,” Mayberry said.
Researchers already put to the test the prototype AI chip. Some of its features include:
- Fully asynchronous neuromorphic many core mesh that supports a wide range of sparse, hierarchical and recurrent neural network topologies with each neuron capable of communicating with thousands of other neurons.
- Each neuromorphic core includes a learning engine that can be programmed to adapt network parameters during operation, supporting supervised, unsupervised, reinforcement and other learning paradigms.
- Fabrication on Intel’s 14 nm process technology.
- A total of 130,000 neurons and 130 million synapses.
- Development and testing of several algorithms with high algorithmic efficiency for problems including path planning, constraint satisfaction, sparse coding, dictionary learning, and dynamic pattern learning and adaptation.
While this is the first neuromorphic chip of Intel, it is not the first time that that concept of mimicking brain neurons has been done. A similar AI chip is currently being developed by IBM which it calls TrueNorth. The TrueNorth chip is comprised of 4,096 cores and 5.4 million transistors, while only utilizing 70 milliwatts of power.
The chip could simulate a million neurons and 256 million synapses-way larger than the Loihi prototype.
While no one knows what the future holds for Loihi, Intel’s growing interest in artificial intelligence and machine learning is undeniable. Something that could potentially pave the way for a new era of processors powered by artificial intelligence to arise.