AiThority Interview Series with Adi Pinhas, CEO and Co-Founder at Brodmann17

Companies across all industries are using AI to make their operations and services more efficient – from healthcare to banking and retail.

How did you start in this space? What inspired you to found Brodmann17?

My co-founders and I have been developing deep learning since the early days of 2012. Back then, like everyone else, we developed deep learning solutions for the cloud. We saw how effective and powerful the cloud was and we knew that the next frontier would be to take deep learning to edge devices.

From day one, our research at Brodmann17 has focused on AI for the edge. To do this, we have had to reinvent deep learning. Optimizing existing open source technologies, which is the common practice today, can only get you so far in terms of increasing speed and efficiency. To get 95% cost reduction, you need to go back to basics and invest new deep learning architectures.

Our solution is pure software, offering game-changing deep learning perception technology so efficient that it can run on any hardware, including low-power processors.

Tell us about Brodmann17’s ADAS capabilities? Would you be directing the company into Autonomous Vehicles in the near future? 

At Brodmann17, we are developing Advanced Driver Assistance Systems (ADAS) and automated driving technology that exceeds current state-of-the-art accuracies, while consuming just a fraction of the resources. This enables these features to expand from premium vehicles to reach mass market vehicles.

A vehicle might be perceived as an expensive solution with an endless amount of power. But that is not the case; the amount of power for every subsystem in a vehicle is limited, and power creates heat that needs to be dissipated, so size and cost is always an issue. The ability to take off 95% of the cost is therefore key. Shortening timelines and bringing ADAS quicker and more safely to the mass market will save lives.

Could you explain Deep Learning in the simplest of ways?

Deep learning is a field inside the larger domain of AI and machine learning. Machine learning is when a machine can learn from data on its own; the data and the examples provided are ‘writing’ the ‘code.’ Deep learning is an algorithm that does this very well and this breakthrough has driven the AI boom in recent years.

How do you see the raging trend of including ‘AI in everything’ impacting businesses?

Companies across all industries are using AI to make their operations and services more efficient – from healthcare to banking and retail. AI-based solutions have proven to be superior to anything that was created previously. Businesses will need to keep up with the pace of the global giants and adapt. One option for them is to work with companies like ours, who already provide breakthrough solutions today.

What are the biggest challenges and opportunities for AI companies in dealing with rising technology prices?

Pricing presents a challenge to bringing AI solutions to mass market products. However, it is also an opportunity to streamline the technology as much as possible and to use our expertise to push our AI capabilities to the limit. While other AI companies are struggling with the costs of bulky, expensive and power-hungry hardware, we’ve avoided this issue by developing a deep learning perception solution that can run even on low-power processors, enabling the same high-quality results without the expense.

Where do you see AI/Machine learning and other smart technologies heading beyond 2025?

Within the automotive sector, we see AI pushing the performance and accuracy of ADAS and automated driving forward. ADAS is projected to reach 80% adoption by 2025, as drivers are becoming increasingly interested in using the technology to ensure safe driving and comfort.

In parallel, legislators have already started  mandating it. As a result – and as cars become increasingly automated – automakers must meet the growing demand for highly-reliable perception technologies. We believe that in seeking to meet this need, automakers and their Tier-1 suppliers will increasingly look to more practical deep learning products that are hardware-agnostic and cost-efficient.

How do you consume information on AI/ML/DL and related topics? (what/where do you read/learn about AI/ML/DL)

Our team of experts delves through about 20 academic articles a week to continually learn and stay up-to-date with the rapid developments and progress of the AI community. A huge amount of money is invested in the industry and the current pace of development is very fast.

What makes deploying AI so hard?

While AI, and deep-learning specifically, has met every benchmark and challenge in recent years, it is still new to everyone. It’s important to test it in every possible scenario, including extreme cases,  to see if its behaviour is still as predicted and meets expectations. There is still a lot to learn.

Thank you, Adi! That was fun and hope to see you back on AiThority soon.