The News: Qualcomm has been developing and experimenting with a wide range of AI initiatives inside its Qualcomm AI Research division since its creation in 2018, but the group’s work has been largely kept out of the public spotlight until now. In a recent blog post, Qualcomm AI Research pulls back the curtains on these AI research initiatives and shares its progress, coverage areas and strategies in making AI a bigger part of everything it does in the enterprise marketplace. For the full blog post click here.

Qualcomm AI Research Quietly Working to Make AI Ubiquitous in a Wide Range of Business Fields, Including Wireless, Automotive, Extended Reality, IoT and Mobile

Analyst Take: Qualcomm’s AI Research division is doing some really interesting things. They’re quiet about it, but it’s clear that their work in the fields of AI, machine learning, and other advanced computing studies is going to touch many industries.

During the four years since the creation of the Qualcomm AI Research group, its researchers, scientists and other staffers have been investigating and experimenting with the use of AI for wireless communications, automotive applications and platforms, extended reality, IoT, and mobile technologies. I’m impressed with the breadth and depth of this work, which is being done by the group’s teams in San Diego, Amsterdam, Seoul, and Markham, Ontario, Canada.

Based on the company’s ongoing commitment to innovation, this path is not a surprising one for Qualcomm. The company is focusing its researchers everywhere to find new ways to use AI to further drive and power Qualcomm products in the global marketplace. This is a core growth area for this company and its partners as well.

So, what are these Qualcomm AI Research team members working on? The assignments cover a broad range of topics, from visionary fundamental research to applied research that solves specific challenges in the market today. This includes computer vision; wireless and RF sensing; power efficiency; machine learning fundamentals; foundational AI research in quantum, geometric, and Bayesian deep learning; speech, audio, and language processing; data compression and generative modeling; personalization and federated learning; optimization and reinforcement learning; and AI compilers and algorithms.

An extensive list of research areas, to be sure, but this commitment across the board supports the company’s idea that virtually every technology can be optimized with machine learning. I believe that this approach is not hyperbole in today’s market.

This is further evidence that the Qualcomm AI Research approach of seeing AI as a ubiquitous technology that can improve other businesses, technologies, and the lives of people is a laudable one, and an indicator of how far-reaching Qualcomm’s diversified strategy is. In a world where everything is connected, or where everything will be connected, Qualcomm can be counted on to play a significant role.

The Importance of Collaboration in AI

One of the most important and beneficial parts of the Qualcomm AI research efforts is that the company’s researchers and scientists are sharing their knowledge, results, and experiments with other internal company teams via an open door policy across all seniority levels and departments. By taking this critical, collaborative step, the Qualcomm AI research group is working to remove internal silos that can keep research in multiple departments from blending and reaching critical mass and success.

A key goal of this work at Qualcomm is bringing together hardware and software experts, researchers, and engineers who will deliver commercialization at scale for their AI ideas and initiatives. I believe this is absolutely the right approach for such research.

But Qualcomm AI Research has not just been doing this work for itself. Researchers are also publishing papers at top scientific conferences to share their insights with others and have been publishing code to allow others to build on top of their results, all in the name of collaboration and information sharing for broader AI development. The group has also released new and improved versions of the AI Model Efficiency Toolkit (AIMET) and released an AIMET whitepaper which describes the quantization techniques that developers can use to successfully implement on-device AI.

All this work focuses on one of the most important goals of Qualcomm AI Research: working to enable compact neural network models and efficient silicon that will allow technology to take on more tasks successfully from humans so that humans can focus on the things that matter to them.

We’ll continue to watch as the Qualcomm AI Research innovates and broadens its initiatives, all in the name of making AI ubiquitous and more useful and look forward to good things ahead.

Disclosure: Futurum Research is a research and advisory firm that engages or has engaged in research, analysis, and advisory services with many technology companies, including those mentioned in this article. The author does not hold any equity positions with any company mentioned in this article.

Analysis and opinions expressed herein are specific to the analyst individually and data and other information that might have been provided for validation, not those of Futurum Research as a whole.