The tremendous growth in compute power and explosion of data is leading every industry to seek AI-based solutions. In this Tech.Decoded video, “The AI Opportunity – Episode 1: The Compute Power Difference,” Vice President of Intel Architecture and AI expert Wei Li shares his views on the opportunities and challenges in AI for software developers, how Intel is supporting their efforts, and where we’re heading next.
The AI software market is expected to reach US$105.8 billion per year worldwide by 2025. The explosion of data is a big reason why AI and deep learning have become so effective. Once you have a lot of data, then the compute resources necessary to process it becomes critical, but no single architecture is dominating all cases – CPUs and GPUs.
Intel takes a holistic approach to AI and machine learning in that the company provides both hardware and software to support AI from the cloud all the way to the edge.
Intel has been working on both the training side and inference side of AI. FPGAs have low latency as well as low power. But software is key to performance, especially for deep learning. The “customers” for deep learning are either data scientists or application developers. Data scientists are different from application developers as they have a high-level interface specifically designed for deep learning. As a result, Intel makes sure it has the best optimized frameworks on Intel platforms. Intel’s goal is to make it is easier for everyone to apply AI.
To learn more about Intel data science solutions click here.
