Nouvelle artificial intelligence

Nouvelle artificial intelligence (AI) is a subfield of artificial intelligence that differs from classical or strong AI in that its aim is to produce robots with an intelligence level akin to that of insects. Developed by Massachusetts Institute of Technology (MIT) roboticist Rodney Brooks in the 1980s, novelle AI is based on the belief that the ability to function in a real-world environment is a key element of true intelligence. Nouvelle AI practitioners seek to construct intelligent agents that are capable of learning and acting on their own without any form of human intervention. These agents should ultimately be able to improve themselves over time. Brooks himself built several nouvelle AI robots, including one named Allen that could avoid stationary and moving objects and another called Herbert that completed simple tasks at the MIT AI laboratory. While nouvelle AI has thus far only produced relatively simplistic systems, advocates believe the approach might one day be the gateway to fully conscious AI.

Background

AI is the machine simulation of human intelligence processes. It is used in a variety of applications, some of which include expert systems, machine vision, speech recognition, and natural language processing. AI systems are typically designed to consume large amounts of labeled training data, analyze that data for patterns, and make predictions about future states based on the patterns it identifies.

Three main cognitive skills are involved in AI programming. These include learning processes, reasoning processes, and self-correction processes. Learning processes are focused on the acquisition of data and the creation of the algorithms, or rules, a system uses to convert raw data into actionable information. Reasoning processes are focused on the system’s ability to choose the appropriate algorithm in order to achieve the correct outcome. Self-correction processes are focused on the system’s ability to modify algorithms to produce the best possible results.

AI systems can be divided into four broad categories: reactive machines, limited memory, theory of mind, and self-aware. Reactive machines are designed to perform specific tasks and have no memory. The most famous example of a reactive machine is the IBM chess program Deep Blue that once beat chess champion Garry Kasparov. Limited memory AI systems are similar to reactive machines but are capable of learning from past experiences because they have memory. Theory of mind AI systems are theoretical future forms of AI that would be equipped with enough social intelligence to be able to understand human emotions and predict behavior. Self-aware AI systems are another type of theoretical future AI that would have consciousness and would be able to understand their own active state.

From a practical perspective, existing AI systems have clear advantages and disadvantages. On one hand, AI systems are typically skilled at doing detail-oriented jobs, capable of completing data-heavy tasks more quickly than humans, proficient at delivering consistent results, and available around the clock. On the other hand, AI systems are often quite expensive, require a high degree of technical expertise, can only do what they are specifically programmed to do, and lack the ability to generalize from one task to another.

Overview

Nouvelle AI is a specific subfield of AI that first emerged in the 1980s. Unlike traditional strong AI, or artificial general intelligence (AGI), which is focused on systems intended to achieve human-level intelligence, nouvelle AI focuses on systems that only achieve insect-level intelligence. In addition, rather than accepting the traditional AI approach of relying on internal models of reality, nouvelle AI ideology holds that the ability to function in a real-world environment is central to true intelligence.

Nouvelle AI was first developed by MIT roboticist Rodney Brooks. Born in 1954, Brooks was an Australian native who first took an interest in computers while studying mathematics at Flinders University in Adelaide, South Australia. He quickly became an expert in working with mainframe computers and subsequently used that experience to embark on a career in AI. Brooks joined the MIT Mobile Robotics Laboratory in 1984 and there took an innovative approach to developing AI. Up to that point, AI developers typically took a top-down approach to AI that involves giving a computer an internal representation of the environment in which it operates. Dissatisfied with this method, Brooks instead argued for a bottom-up approach that focused more on action and behavior. This ultimately led him to develop nouvelle AI. In the years that followed, Brooks co-founded iRobot, the company that created the Roomba, and several other robotics businesses. He also served as director of the MIT Artificial Intelligence Research Laboratory.

Brooks’ nouvelle AI revolves around the creation of intelligent agents that are capable of learning and acting on their own. It operates on the premise that AI should be able to work independently and improve itself over time. According to nouvelle AI philosophy, intelligence emerges from simple behaviors as an agent interacts with the real world. In other words, a robot powered by nouvelle AI and programmed with simple behaviors such as the ability to move towards a moving object and avoid obstacles could eventually learn a more advanced behavior like chasing a moving object.

Brooks’ early insectoid nouvelle AI robots Allen and Herbert demonstrated this idea. Allen was equipped with a dozen ultrasonic sonars and a trio of independent behavior-producing modules programmed to avoid moving and stationary objects. When only these modules were activated, Allen would stay in the middle of a room until an object approached and then avoided obstacles while moving away from the object. Herbert was equipped with infrared sensors for obstacle avoidance and a laser system it used to collect three-dimensional data within a twelve-foot range. Operating in the real-world environment of the MIT AI lab, Herbert regularly searched for and collected empty soda cans. This apparently goal-oriented behavior emerged from the interaction of around fifteen different simple behaviors. Although neither Allen, Herbert, nor any of the other similar robots Brooks later created came anywhere near the complexity of real insects, they all clearly illustrated the potential of nouvelle AI.

Bibliography

Burns, Ed. “What is Artificial Intelligence? (AI).” TechTarget, 2023, www.techtarget.com/searchenterpriseai/definition/AI-Artificial-Intelligence. Accessed 28 Mar. 2023.

Copeland, Jack. “What is Artificial Intelligence?” AlanTuring.net, May 2000, www.alanturing.net/turing‗archive/pages/Reference%20Articles/what‗is‗AI/What%20is%20AI11.html. Accessed 28 Mar. 2023.

Frankenfield, Jake. “Artificial Intelligence: What It Is and How It Is Used.” Investopedia, 6 July 2022, www.investopedia.com/terms/a/artificial-intelligence-ai.asp. Accessed 28 Mar. 2023.

Jeevanandam, Nivash. “The Ambitious Goal to Make Robots Act & Look More Real.” Analytics India Magazine, 1 Oct. 2021, analyticsindiamag.com/embodied-ai-nouvelle-ai. Accessed 28 Mar. 2023.

“Nouvelle AI.” A.I. for Anyone, 2023, www.aiforanyone.org/glossary/nouvelle-ai. Accessed 28 Mar. 2023.

“Rodney Brooks Facts & Worksheets.” Kidskonnect, 2023, kidskonnect.com/people/rodney-brooks. Accessed 28 Mar. 2023.

Schroer, Alyssa and Andreas Rekdal. “Artificial Intelligence.” BuiltIn, 3 Mar. 2023, builtin.com/artificial-intelligence. Accessed 28 Mar. 2023.

Shadbolt, Nigel. "From So Simple a Beginning: Species of AI." Daedalus, vol. 15, no. 2, 1 May 2022, pp. 28-42, doi.org/10.1162/daed‗a‗01898. Accessed 14 Nov. 2024.

“What Is Artificial Intelligence (AI)?” IBM, 2023, www.ibm.com/topics/artificial-intelligence. Accessed 28 Mar. 2023.