The way we live is changing more and more every day as a result of artificial and machine intelligence. Self-driving vehicles, manufacturing robots, and online travel agencies are just a few instances of how these technologies have impacted our daily lives. Even with the noteworthy improvements in machine intelligence over the past several years, according to JoAnn Paul, associate professor in the Bradley Department of Electrical and Computer Engineering, we’ve just started to scratch the surface.
Paul has been given a National Science Foundation grant to research how to create a computer architecture that mimics how the human brain functions to advance the field of artificial intelligence. The $400,000 project’s lone principal investigator, a professor at Virginia Tech who specialises in brain-inspired computer architecture, aspires to build the groundwork for a potentially ground-breaking understanding of real artificial intelligence.
“Artificial intelligence” is a misnomer in its present state. “No kind of computer used today is knowledgeable,” Paul stated. As they are necessary for general intelligence, “real machine intelligence must demonstrate common sense and creativity to come into existence.”
The human brain is capable of processing several incoming inputs of various qualities at once. This degree of information analysis is referred to as parallel processing, which academics have been attempting to replicate on computers for quite some time, according to a new study in Nature.
Paul’s study focuses more precisely on solving a paradox, which is another talent of the human brain.
What then is a paradox? It may be seen as the brain’s greatest degree of conflict accommodation in terms of information processing in the brain. Paul claims that the human brain uses parallelism to handle ambiguity when processing information that is in conflict. The basis for our ability to think creatively is uncertainty and ambiguity.
Paul once remarked, “Creativity is the outcome of seeing things in several, varied, even conflicting ways. “The classic illustration of a contradiction is a paradox. To understand how conflicting outputs may be tolerated and how they can continually change, we will explore algorithm-architecture solutions to issues that create them. This will build the groundwork for innovation.”
Finding basic issues to highlight relevant ideas will be one of the difficulties of this study. Paul’s study focuses on how the brain compartmentalises information but also utilises various compartments depending on the inputs or the context. This capacity could be the secret to the brain’s ability to handle conflict and ambiguity while simultaneously using such high levels of parallelism.
An example might be a dialogue between two bilingual individuals. A statement that is spoken by one person but includes words in both French and English may be processed by different parts of the brain, one for each language. Whether the statement is in French or English, each appropriate section must conclude. The correct phrase, according to Paul, is neither—nor both.
In particular, if there is phonetic ambiguity in what is being said, the listener must determine what the speaker genuinely means, according to Paul. They have to accomplish this while avoiding their language centres becoming too big and becoming ineffective. They also have to deal with conflict and uncertainty. To better comprehend the fundamental mechanisms behind actual machine intelligence, Paul aims to apply the same ideas to computer architecture that are used to analyse information in the human brain.
Algorithms continue to be a major part of how machine and artificial intelligence “make judgments,” yet adhering to these guidelines does not result in outcomes that are human-like. The majority of issues have several solutions, as most of us have discovered via trial and error.
Paul and graduate student Isaac Bettendorf will run experiments throughout the three-year award cycle that include issues with various algorithmic solutions comparable to those that a human could come across daily. These tests will contribute to the development of a high-level foundation for a new architecture that can support large levels of parallelism and resolve conflict.
A master’s degree in computer engineering is what Bettendorf is now working for. As an embedded systems engineer with the U.S. Navy in Dahlgren, he appreciates the challenge of this specific project and the fresh outlook it has provided.
“Fundamentally fresh perspectives on computers are being offered by the study that Dr Paul and I are doing. The privilege of working with her, “Bettendorf stated. “I’ve had a lot of possibilities working on technology for the Navy. I’ve created and implemented solutions utilising both integrated circuit and microcontroller unit contexts, and I’ve come into many of the constraints that [Dr. Paul and I] addressed this in our study. This has in many ways encouraged me to sit back, attempt to see the wider picture, and approach these limits from a new angle.”
Paul will also write a magazine-style piece to educate readers without a background in STEM in addition to the study. By discussing the potential uses of personal avatars and how they might eventually be making decisions for us—especially if those avatars are capable of coming up with innovative solutions to problems—the article will concentrate on the differences between computer architecture and human intelligence.
The fascinating questions, according to Paul, are the basic ones. “Applied research doesn’t thrill me as much. True research, in my opinion, offers the foundation and impetus to see things completely different rather than a tool or a solution.”
