Metro

  • Written by Forest Agostinelli, Assistant Professor of Computer Science, University of South Carolina

The field of artificial intelligence (AI) has created computers that can drive cars[1], synthesize chemical compounds[2], fold proteins[3] and detect high-energy particles[4] at a superhuman level.

However, these AI algorithms cannot explain the thought processes behind their decisions. A computer that masters protein folding and also tells researchers more about the rules of biology is much more useful than a computer that folds proteins without explanation.

Therefore, AI researchers like me[5] are now turning our efforts toward developing AI algorithms that can explain themselves in a manner that humans can understand. If we can do this, I believe that AI will be able to uncover and teach people new facts about the world that have not yet been discovered, leading to new innovations.

A humanoid robot reading a book. When machines are left to learn and solve problems through their own experience, this is called reinforcement learning. Gremlin/E+ via Getty Images[6]

Learning from experience

One field of AI, called reinforcement learning[7], studies how computers can learn from their own experiences. In reinforcement learning, an AI explores the world, receiving positive or negative feedback based on its actions.

This approach has led to algorithms that have independently learned to play chess at a superhuman level[8] and prove mathematical theorems[9] without any human guidance. In my work as an AI researcher[10], I use reinforcement learning to create AI algorithms that learn how to solve puzzles such as the Rubik’s Cube[11].

Through reinforcement learning, AIs are independently learning to solve problems that even humans struggle to figure out. This has got me and many other researchers thinking less about what AI can learn and more about what humans can learn from AI. A computer that can solve the Rubik’s Cube should be able to teach people how to solve it, too.

Peering into the black box

Unfortunately, the minds of superhuman AIs are currently out of reach to us humans. AIs make terrible teachers and are what we in the computer science world call “black boxes[12].”

An open black case Researchers have been trying for decades to understand how AIs solve problems. rockz/iStock via Getty Images Plus[13]

A black-box AI simply spits out solutions without giving reasons for its solutions. Computer scientists have been trying for decades to open this black box[14], and recent research has shown that many AI algorithms actually do think in ways that are similar to humans. For example, a computer trained to recognize animals will learn about different types of eyes and ears and will put this information together to correctly identify the animal[15].

The effort to open up the black box is called explainable AI[16]. My research group[17] at the AI Institute at the University of South Carolina is interested in developing explainable AI. To accomplish this, we work heavily with the Rubik’s Cube.

The Rubik’s Cube is basically a pathfinding problem[18]: Find a path from point A – a scrambled Rubik’s Cube – to point B – a solved Rubik’s Cube. Other pathfinding problems include navigation, theorem proving and chemical synthesis.

My lab has set up a website where anyone can see how our AI algorithm solves the Rubik’s Cube[19]; however, a person would be hard-pressed to learn how to solve the cube from this website. This is because the computer cannot tell you the logic behind its solutions.

Solutions to the Rubik’s Cube can be broken down into a few generalized steps – the first step, for example, could be to form a cross while the second step could be to put the corner pieces in place. While the Rubik’s Cube itself has over 10 to the 19th power possible combinations, a generalized step-by-step guide is very easy to remember and is applicable in many different scenarios.

Approaching a problem by breaking it down into steps is often the default manner in which people explain things to one another. The Rubik’s Cube naturally fits into this step-by-step framework, which gives us the opportunity to open the black box of our algorithm more easily. Creating AI algorithms that have this ability could allow people to collaborate with AI and break down a wide variety of complex problems into easy-to-understand steps.

How explainable artificial intelligence can help humans innovate A step-by-step refinement approach can make it easier for humans to understand why AIs do the things they do. Forest Agostinelli, CC BY-ND[20]

Collaboration leads to innovation

Our process starts with using one’s own intuition to define a step-by-step plan thought to potentially solve a complex problem. The algorithm then looks at each individual step and gives feedback about which steps are possible, which are impossible and ways the plan could be improved. The human then refines the initial plan using the advice from the AI, and the process repeats until the problem is solved. The hope is that the person and the AI will eventually converge to a kind of mutual understanding.

[Deep knowledge, daily. Sign up for The Conversation’s newsletter[21].]

Currently, our algorithm is able to consider a human plan for solving the Rubik’s Cube, suggest improvements to the plan, recognize plans that do not work and find alternatives that do. In doing so, it gives feedback that leads to a step-by-step plan for solving the Rubik’s Cube that a person can understand. Our team’s next step is to build an intuitive interface that will allow our algorithm to teach people how to solve the Rubik’s Cube. Our hope is to generalize this approach to a wide range of pathfinding problems.

People are intuitive in a way unmatched by any AI, but machines are far better in their computational power and algorithmic rigor. This back and forth between man and machine utilizes the strengths from both. I believe this type of collaboration will shed light on previously unsolved problems in everything from chemistry to mathematics, leading to new solutions, intuitions and innovations that may have, otherwise, been out of reach.

References

  1. ^ drive cars (www.nytimes.com)
  2. ^ synthesize chemical compounds (arxiv.org)
  3. ^ fold proteins (doi.org)
  4. ^ detect high-energy particles (doi.org)
  5. ^ AI researchers like me (scholar.google.com)
  6. ^ Gremlin/E+ via Getty Images (www.gettyimages.com)
  7. ^ called reinforcement learning (en.wikipedia.org)
  8. ^ play chess at a superhuman level (doi.org)
  9. ^ mathematical theorems (papers.nips.cc)
  10. ^ an AI researcher (scholar.google.com)
  11. ^ solve puzzles such as the Rubik’s Cube (doi.org)
  12. ^ black boxes (doi.org)
  13. ^ rockz/iStock via Getty Images Plus (www.gettyimages.com)
  14. ^ decades to open this black box (papers.nips.cc)
  15. ^ to correctly identify the animal (doi.org)
  16. ^ explainable AI (proceedings.neurips.cc)
  17. ^ My research group (cse.sc.edu)
  18. ^ pathfinding problem (en.wikipedia.org)
  19. ^ AI algorithm solves the Rubik’s Cube (deepcube.igb.uci.edu)
  20. ^ CC BY-ND (creativecommons.org)
  21. ^ Sign up for The Conversation’s newsletter (theconversation.com)

Authors: Forest Agostinelli, Assistant Professor of Computer Science, University of South Carolina

Read more https://theconversation.com/how-explainable-artificial-intelligence-can-help-humans-innovate-151737

Metropolitan republishes selected articles from The Conversation USA with permission

Visit The Conversation to see more

Entertainment News

Gonna Tell You A Bit About Me & My Journey

There was a time that I didn’t like seeing myself in a mirror. I lacked confidence. I was shy. Then, I found modeling. When I first started modeling, being in front of the camera didn’t feel natural and I was nervous. Like most things – practice ...

Thorn Castillo - avatar Thorn Castillo

#1 Chart-Topping Band GREYE Comes Full Circle With “I Don’t Mind”

Six years after releasing their debut album Providence, Greye’s evolution from progressive Americana upstart to trailblazing indie rock powerhouse is complete. Pandemic and lockdown be damned, the well-traveled multi-generational Daytona Beach, F...

News Co - avatar News Co

AN INAUGURATION AND AMERICAN GREED’S SINGLE “TOGETHER” START AMERICA ON A PATH TOWARD'S HEALING

From COVID-19 To the Shocking Events of January 6, This Is Michael Mesey’s Song of Hope to A Weary Nation and World This Inauguration Day is a chance to turn the page from a difficult, divisive era and embrace a more hopeful spirit of unity m...

Michael Mesey - avatar Michael Mesey

EMODULARI Releases The Hard-Driving, Intensely Visceral “Craniotomy”

Caught up in the maelstrom of bad to worse news in the New York Times one morning, Eli Soiefer spotted an uplifting article offering respite from the crazy. It was a piece about an open-brain surgical procedure where they keep the patient awake t...

News Co - avatar News Co

AV SUPER SUNSHINE “Super Cool”

“Super Cool” is more than the name of rising international hitmaker AV Super Sunshine’s latest infectious 80’s techno inflected single. It’s the dizzying, freewheeling world he’s been living in these past months since the release of his previous ...

News Co - avatar News Co

Angela Predhomme - “Warmest Wish for You”

What could possibly help us heal from the exceptionally trying year of 2020? Singer-songwriter Angela Predhomme hopes that an inspiring new holiday song might give us a little solace. “Warmest Wish for You” is soul-tinged, feel-good song with a ...

News Co - avatar News Co

Metropolitan Business News

Do Directories Still Help SEO?

Directories were once one of the main staples of SEO and they were definitely in existence before search engines took over. Directories were once the main way that people navigated the world wide we...

News Co - avatar News Co

Shipping Container FAQs

If you are looking to rent or purchase a shipping container, you probably have a few questions. We have selected some of the most common questions and answered them here. We hope that you find this ...

News Co - avatar News Co

Is Duplicate Content an SEO Myth?

Any marketer you speak to about duplicate content is concerned about a “duplicate content penalty” but they are probably not very experienced with SEO. Google has specific guidelines on duplicate co...

News Co - avatar News Co

5 Key Factors To Starting A Website

Every business needs a website to be successful. However, this is a challenge for many due to the lack of coding and designing skills. The good news is, you do have a way around this—hire a deve...

News Co - avatar News Co

So You Want To Be A Project Manager? Here's What You Must Do

Project management is often a popular choice for people looking to make a mark in the organizations they work for. It's a great choice as well. If you hate monotony, love getting people and res...

News Co - avatar News Co

How to Deliver the Perfect Elevator Pitch

Do you struggle with explaining your career with professional or personal contacts, and you want to make a lasting impression? Whether you're looking for a job or building up your network, a wel...

Adam Jacobs - avatar Adam Jacobs

Writers Wanted


News Co Media Group

Content & Technology Connecting Global Audiences

More Information - Less Opinion