“We are at the beginning of a major digital transformation, similar to the internet in the 90’s or mobile in the 2000s.” Learn about the state of Computer Vision in 2022 from Emrah Gultekin, CEO
Emrah Gultekin:
Hi everyone. I’m Emrah Gultekin, the CEO of Chooch AI. Welcome to the Infinite Vision Summit. The big idea here is computer vision. The journey of computer vision. We’re making computers see and understand the same way humans do.
Emrah Gultekin:
When we embarked on this mission several years ago, computer vision was confined to basic functions like image tagging and motion detection. But as we began to dig in, we discovered a universe of possibilities that would impact how enterprises work and how we conduct our own lives.
Emrah Gultekin:
The possibility of deploying computer vision, everywhere, the process of copying human vision into machine as a company, and as an ecosystem, basically putting a set of eyes onto whatever you want. See, computer vision is a subset of what is truly happening today. We’re at the beginning of a major transformation, like what happened in the nineties with the internet, or mobile in the two thousands. This is probably a lot bigger and will change how we work and how we live forever. Again, computer vision is a scaling of visual intelligence, never before seen in our lifetimes.
Emrah Gultekin:
Imagine being able to take a person’s visual experience and knowledge and clone it throughout camera streams and videos, live and historic. Imagine the insights and real information you can gather. You’re already aware of the developments with self-driving cars, with the metaverse, with smart glasses, AR/VR, with digital twins. There’s a major element of computer vision in all of these. Those are the more fancy cases you hear about. We hear less about the more mundane uses like computer vision and safety security, patient monitoring, human interactions, defect detection, manufacturing, logistics, retail, geospatial, and the list goes on and on and on. Basically everything.
Emrah Gultekin:
In the world of smart devices, one thing we do very well is deploy ML models to existing infrastructure. This transforms cameras into sensors that can detect defects, dangers or anomalies, and anything that can be detected visually, in any spectrum of light.
Emrah Gultekin:
See every new digital product in the future will have an element of computer vision embedded in them and it will not be trivial. What does this mean practically? It means less lives lost, less defects, less risk. Better outcomes for all of us. See from microscopes to street cameras, from cars to satellite, we add intelligence to any stream and keep those ML applications updated.
Emrah Gultekin:
These are some of the things we do to augment human vision but it’s much bigger than that. We’re structuring all the previously unstructured video and image data, whether live or historic. This has never really been done before by machines. Of course, the mission and transformations are massive and there are many challenges ahead of us. I think these challenges are threefold. We’ve got technical challenges, we’ve got adoption challenges, and we’ve got some cultural challenges.
Emrah Gultekin:
Some technical challenges we face in the ecosystem is building robust data sets, dynamic model training at scale, inference engines and analytics insights, with this newly incoming and generated data. The availability of various forms of hardware is also an issue. These are known challenges and many companies are trying to solve for specific parts of them. On the other side, we’ve got adoption challenges and partially are caused by technical difficulties but mainly caused by mixed results, in first time discovery projects.
Emrah Gultekin:
See, computer vision has historically been a sideshow, not a core function or a core capability within an enterprise. But advanced computer vision requires this capability to sit at the core of the company’s information technology strategy. That shift is slowly happening and in fact is inevitable.
Emrah Gultekin:
Of course, there are also significant cultural and ethical challenges we face, as we move into very new levels of efficiency. Some of these challenges include workforce related transformations. For example, new jobs are emerging, like data generators, annotation experts, model trainers, and ML ops engineers. These new jobs require retraining workforces, which takes time. They’re also challenges around how ethical AI can be deployed, including issues surrounding privacy and bias. We’re all learning in this process.
Emrah Gultekin:
There will definitely be more challenges future and we just can’t foresee today, but let’s focus on the benefits. The benefits we all can enjoy, like more efficiency in the workplace, less stress on humans, and more output for everyone to consume. We believe we can all benefit, at some level, from these advancements. We don’t believe in an efficiency, equity trade-off in this particular context.
Emrah Gultekin:
The point of all technologies that give us more efficiency, so we have time to do the things we want to do like build something of interest, learn a new skill, learn a new language, travel, discover new ideas and spaces, or just spend time with friends and family. This has always been the case throughout history. It’s nothing new.
Emrah Gultekin:
A technology advancements generally creates positive externalities when handled well. As these technologies become more widespread, our companies and our communities will become safer, more transparent and more equal, if the insights are used the right way.
Emrah Gultekin:
Of course their potential downsides to any new technology and we need to be cognizant of those challenges and try to foresee and rectify any negative impacts as much as possible. For example, something we deal with on a daily basis is bias and data sets, bias and model testing and deployment, the inference engine, bias and analytics. These are important challenges we face today as practitioners. That’s one reason human and machine feedback need to be provided in real time, to retrain models on the fly, when necessary.
Emrah Gultekin:
These are challenges we can solve. We can solve them collectively. Here’s an interesting thought, on a different macro scale, to challenge our pace of adoption with these new advancements. If you’ve been following demographic trends, you’ve already noticed the world is getting older. With aging populations in all of the industrialized and most of the developing world, we need to find ways to become way more efficient in order to create more output. Why? Because of decreasing workforces and increasing amount of people moving into retirement. The economics only add up with technology interventions. That’s just another reason to embrace these new products. The machines don’t get tired. Don’t need breaks. Don’t need to sleep but they do need maintenance, upgrades, and guidance.
Emrah Gultekin:
Let’s not replace humans with machines but let’s augment humans with machines, so we are able to increase our output and serve our companies, our communities, and our fellow humans.
Emrah Gultekin:
Well, thanks for listening to all that. Welcome again to the Infinite Vision Summit. We’ve got some great speakers here today. Please enjoy. The future’s yours.
