Centaur on the Rise: How a Decades-Old Paradigm is Changing the Way Major Institutions Think About AI





We’re excited to bring back Transform 2022 in person on July 19 and virtually from July 20 through August 3. Join leaders in AI and data for in-depth discussions and exciting networking opportunities. Learn more about Transformer 2022


Are you ready to be a centaur?

What is the most important part of an AI system?

Are these the terabytes of data you use to train the base model? The billions of weights and biases sitting atop your gradient tower? Network architectures meticulously designed and built over decades of hard work?

What bottleneck are we stuck in?

Are our GPUs just not powerful enough? Do we need a few smart architectural tweaks and fine points to lead us to full automation? Or maybe shoveling hundreds of millions of dollars into the data labeling money pit will tip us into the future?

Or maybe, just maybe, we thought of everything the wrong way.

Perhaps we have already shifted to a new paradigm for AI as a direct result of the meteoric rise of deep learning techniques. Perhaps the most important part of your AI system is the person using it.

power of the people

With all the emphasis on full automation and Level 5 autonomy, it almost seems foolish to focus on the person operating the system. After all, they are only temporary. However, as the adoption of AI within the enterprise continues to accelerate, we see a very different picture emerging.

In the vast majority of cases, the success of AI initiatives comes down to transparency, control and trust. Eighty-four percent of businesses still don’t trust AI, and a deep lack of specialized AI talent — compounded by the pandemic — is one of the biggest barriers to AI adoption. The modern appetite for automation can’t wait for a workforce of hundreds of thousands of AI experts who aren’t coming.

All of this points to a critical need to rethink the way we build AI systems. How do you empower the citizen data scientist and bring the next slice of AI users into the fold? We need to stop viewing them as stand-alone systems with incidental humans. The development, operation and maintenance of these systems are all fundamentally people-centric.

Enter the centaur.

After Garry Kasparov’s famous loss to Deep Blue in 1997, the world watched with anticipation, wondering what the future looked like for humans in chess. A person who does not have waiting was Garry Kasparov. In the truest expression of “If you can’t beat them, join them”, Garry Kasparov has teamed up with a chess program called Fritz 5 to become the world’s first Centaur. In 1998 he participated in the world’s first centaur chess competition against Veselin Topalov associated with ChessBase 7.0.

Even today, with two more decades of AI advancements under our belt, Centaurs compete with the best AI in the world. Given the obvious complexities of benchmarking centaurs against pure chess AI, the exact state of the art is somewhat controversial, but Garry Kasparov claimed in 2017 that there was no “no doubt” that “a human paired with a set of programs is better than playing against just the most powerful computer program for chess.

The paradox here is that human control and direction adds value even when the AI ​​is operating at blatantly superhuman levels. The assumption that a sufficiently advanced AI will eliminate the need for human beings seems false. Instead, we are now tasked with creating the proper interface for mutualism between us and the AI.

Even massive organizations, unwavering in their dedication to general artificial intelligence, have begun to adopt more holistic approaches that recognize humans as a necessary part of the process. Obvious lines of evidence include the progressive emphasis on learning by a few moves rather than learning by zero moves, the closely related rise of rapid engineering, and Microsoft’s promotion of machine learning.

Famously, OpenAI has even begun to effectively include human beings in its training architecture. In a recent paper, they significantly outperformed state-of-the-art summaries by integrating a human feedback loop directly into the structure of their experiment. It’s a many of human involvement for a domain that is supposed to automate humans.

But we shouldn’t be surprised.

History repeats and repeats

The first industrial revolution began with steam and iron, but was built with the loom and machine tools. The breakthrough was critical, but the interface with humans was the thing that changed the world.

The second industrial revolution began with steel and sparks, but was built with rail and the telegraph. Even when technology was rudimentary by our modern standards, it seeped into the critical pathways of daily life.

The third industrial revolution started with digital logic and computability, but it was built with silicon, HTML and javascript. As in all other industrial revolutions, these early advances have a timeless quality. Pong, MS Paint, Notepad – even with the drastic improvements in technology since their release, their interfaces are still relevant and influential.

The fourth industrial revolution – the AI ​​revolution – is underway. Data infrastructure, machine learning and cloud computing are key enablers, but the underlying technology will only be evident in hindsight. What East clear is that the basic interfaces that will echo through history have yet to be developed.

Despite all the incredible work we’ve done to improve the technology, the vast majority of our AI interfaces have remained unchanged for decades. We are sorely lacking the interfaces we need to enable self-driving cars, automated assistants and other theoretically revolutionary technologies.

This is our challenge. The next generation of AI problems must be centered around user experience and human cognition as much as developing and improving massive neural networks. We need to start learning from the past to recognize that these human/machine interfaces that we are building are do not an intermediate state on our way to a utopian, automated future. They are the future.

So I’ll ask again.

Are you ready to be a centaur?

I am.

Slater Victoroff is founder and CTO of Indico Data.

DataDecisionMakers

Welcome to the VentureBeat community!

DataDecisionMakers is where experts, including data technicians, can share data insights and innovations.

If you want to learn more about cutting-edge insights and up-to-date information, best practices, and the future of data and data technology, join us at DataDecisionMakers.

You might even consider writing your own article!

Learn more about DataDecisionMakers




Leave a Comment

x