Robots are often celebrated for their ability to efficiently handle repetitive tasks.
Just like an information security analyst, HR teams can leverage intelligent software to handle routine tasks like scheduling interviews, managing benefits, and providing answers to common HR questions.
However, it is quite different for a robot to perform repetitive tasks consistently, as that is the main purpose of robotic process automation. On the other hand, it becomes a whole new challenge when the machine is faced with a completely unfamiliar situation each time.
Imagine a warehouse robot that is specifically designed to locate and retrieve items from shelves. Now picture that robot being instructed to find that item in a messy environment. Can the machine effectively evaluate the new challenges at hand?
Is it possible for a worker robot to learn from its mistakes?
Robots are celebrated for their ability to excel at repetitive tasks.
Just like an information security analyst, HR teams can utilize intelligent software to handle routine tasks like scheduling interviews, managing benefits, and providing answers to common HR questions.
However, it is quite different for a robot to perform repetitive tasks consistently, as that is the main purpose of robotic process automation. On the other hand, it becomes a whole new challenge when the machine is faced with completely unfamiliar situations each time.
Imagine a warehouse robot that is specifically designed to locate and retrieve items from shelves. Now picture that robot being instructed to find that object in a messy environment. Can the machine effectively evaluate the new challenges at hand?
Discover more: Exploring the harmonious coexistence of humans and robots in the workforce
Intelligent robots are becoming increasingly prevalent
This new type of robot being developed by computer scientists at the University of Leeds is based on two aspects of artificial intelligence. There are two main areas of focus: automated planning and reinforcement learning.
Initially, the robotic software visually assesses the environment in which the machine will operate. Considering the challenges at hand, the machine carefully plans its approach to successfully complete the task.
However, not all worker robots possess the necessary level of sophistication to effectively navigate real-world obstacles. Additionally, the researchers incorporated reinforcement learning into their approach.
The process involves the robot going through a series of trial and error, continuously learning from its mistakes. Actually, there were a total of 10,000. Accidentally causing disruptions, clumsily toppling objects, or inadvertently letting the target item slip from one’s grasp. With the aim of comprehending the most effective course of action.
According to Dr. Matteo Leonetti, who was involved in the study, robots have demonstrated their ability to reason in activities such as playing chess against grandmasters.
According to him, robots lack the agility and mobility that humans excel in.
According to Dr. Leonetti, the human brain has developed these physical skills through evolution and extensive practice. We are implementing that concept in the upcoming generation of robots.