Introducing the first-ever self-learning chip for Intel. The company has been working on specialized architectures for accelerating classic computer platforms. The company has recently incremented in its investment in R&D in AI and neuromorphic computing.
In fact, Intel’s work on neuromorphic computing is built around decades of research and collaborations. The efforts led the company to come up with the first-ever self-learning chip, codenamed Loihi.
It is still in the prototype stage and has the capability of adapting to personalized customization as per needs. By the first half of 2018, the leading universities and research institutions, having a focus on AI advancements, will get to have the Loihi chip.
3 Top Features of the new Self-learning Chip
- The neuromorphic chip has the ability to mimic brain functions through a learning process of operating in response to environmental feedback.
- The supremely energy-efficient chip does not require traditional training for growth, as it becomes smarter over time.
- The chip is able enough to use data for learning and making inferences, thus implementing to the best of its capabilities.
Furthermore, thanks to the chip, machine learning will become faster, with enhanced efficiency but low computing power. It works in the process based on high flexibility on-chip learning, alongside training and inference on itself.
The Design
The inspiration behind the chip’s design comes from the neurons’ communication process. The model uses plastic synapses and spikes, which “can be modulated based on timing.” This makes computers self-organized and leads them to be decisive ‘based on patterns and associations.”
Inference
Intel aims at driving computing beyond desktops and servers while keeping up with the evolving pace of technology. Introducing the new self-learning chip is another major step in the ambition of the company.
