Machine perception is that aspect of digital technology that involves developing computers that can sense or “perceive” the outside world in a way that accurately mimics the five human senses – sight, sound, touch, smell, and taste — as well as taking in information in ways that humans cannot.

As you might imagine, machine perception is an integral part of machine learning and AI, particularly in AI applications that require fast decision-making based on perceiving the surrounding environment, such as autonomous driving.

By definition, perception is the process by which sensory information is captured from the world around us and then interpreted, understood, and organized to make decisions based on the input of the sensory data. In humans, that data is obtained by our sensory organs, such as our eyes, ears, skin, nose, and tongue. These specialized organic receptors transmit their information across neural pathways to the brain to organize and interpret the data they obtain.

In machine learning and AI, a variety of digital sensors are used to replicate or augment the human sense organs. These sensors then work with a complex network of hardware and software that is, in essence, a digital parallel of the nervous system to create “machine perception.”

Machine perception is a cornerstone of every AI sensory model or cognitive digital twin application. The algorithms convert the data gathered from the world into a raw model of what is being perceived by the AI or the twin.

In theory, any direct, computer-based gleaning of information from the world is a kind of machine perception. That is anything from the photoelectric sensors that automatically turn on your car’s headlights at night to how your Roomba vacuum navigates around your living room. Right now, AI and machine perception applications are in development that are designed to emulate each of the human senses, such as:

• Machine or computer vision via optical camera

• Machine hearing (computer audition) via microphone

• Machine touch via tactile sensor

• Machine smell (olfactory) via electronic nose

• Machine taste via electronic tongue

• 3D imaging or scanning via LiDAR sensor or scanner

• Motion detection via accelerometer, gyroscope, or magnetometer

• Infrared and thermal imaging sensors

From this list above, you can see how critical machine perception can be to AI applications such as medical diagnoses, as well as developing truly safe and ubiquitous autonomous vehicles. Innovation in machine perception will also pave the way for next-generation “robotic assistants” or companions.

But developing machine perception is not as easy as it may seem. Computers may be able to solve complex equations and process data vastly superior to humans; however, there are some things that humans still do a lot better than machines. Perception and acting quickly and spontaneously on data from our senses is one of them. Things we do with ease are proving to be very hard to “teach” computers. Take, for example, handwritten text. Handwriting varies greatly from individual to individual, yet, we all can pick up and read handwritten text with no problem for the most part, yet it is difficult to get AI to decern those variables in letter composition.

Similarly, a two-year-old can learn to catch a tossed ball after only a few attempts. But teaching a robot to do the same takes a lot more work. That’s because we are not even sure of the infinite combinations of data processing that almost instantaneously take place as your eyes perceive a ball coming towards you and your brain puts your hand up in time to catch it.

In the 1980s, Hans Moravec, famed member of the Robotics Institute of Carnegie Mellon University in Pittsburgh, described the paradox this way, “It is comparatively easy to make computers exhibit adult-level performance on intelligence tests or playing checkers, and difficult or impossible to give them the skills of a one-year-old when it comes to perception and mobility.”

However, such limitations will likely be more easily overcome as AI, and machine learning begin to peel away from conventional computers and enter the realm of quantum computing.

Quantum computing holds the promise of exponentially improving the way AI algorithms process, analyze and present sensory findings and predictions and may hold the key to bringing machine perception more analogous to the functionality of the human brain and nervous system.

A number of companies — startups as well as established challengers — are working to make their AI models perceive the world more as humans do, and BigRio is helping to enable much of this advancement.

How BigRio Helps Facilitate Advancement in Machine Learning

Like breakthroughs in machine perception, BigRio looks for and helps to facilitate such innovation in machine learning and AI.

We like to think of ourselves as a “Shark Tank for AI.”

If you are familiar with the TV series, then you know that, basically, what they do is hyper-accelerate the most important part of the incubation process – visibility. You can’t get better visibility than getting in front of celebrity investors and a TV audience of millions of viewers. Many entrepreneurs who have appeared on that program – even those who did not get picked up by the sharks – succeeded because others who were interested in their concepts saw them on the show.

At BigRio, we may not have a TV audience, but we can do the same. We have the contacts and the expertise to not only weed out the companies that are not ready, as the sharks on the TV show do but also mentor and get those that we feel are readily noticed by the right people in the growing AI investment community.

Because we see so many potential AI innovators, we are also ideally suited to facilitate advancements in machine learning, such as improved machine perception, that will usher in a new generation of autonomous vehicles and other smart machines.

Rohit Mahajan is a Managing Partner with BigRio. He has a particular expertise in the development and design of innovative solutions for clients in Healthcare, Financial Services, Retail, Automotive, Manufacturing, and other industry segments.

BigRio is a technology consulting firm empowering data to drive innovation, and advanced AI. We specialize in cutting-edge Big Data, Machine Learning, and Custom Software strategy, analysis, architecture, and implementation solutions. If you would like to benefit from our expertise in these areas or if you have further questions on the content of this article, please do not hesitate to contact us.