How Machine Learning makes Robots more intelligent ?

  • Prasobh V Nair
  • Published: JANUARY 11, 2019
How Machine Learning makes Robots more intelligent ?

Today, Robots are evolving fast and are getting far smarter than we had ever imagined. However, there is a big role that which Machine Learning and machine perception together play, in the overall advancement of this technology. Developments in machine perception, which is the next stage of Machine Learning, could be very fascinating for the entire field of intelligent robotics.

If you compare the modern robots with their predecessors from just a decade ago, the modern robots are extremely adaptable and intelligent. The biggest factor that has differentiated robots from humans throughout the dawn of technology, is the fact that humans can learn from their experiences. However, machines are seemingly not able to do so (well, not until the concept of Machine Learning and machine perception came about). The area of Machine Learning is rapidly undergoing an advancement as the current generation of the robots is able to learn and adapt themselves based on the input that they receive from their surroundings. This essentially removes another barrier from the road towards achieving artificial consciousness (something that which scientists dream of achieving one day, as it is apparent in Hollywood movies).

This field is referred to as ‘smart robotics’, or ‘intelligent robotics’. In this field, machine perception plays the biggest role.

This is because, in order to become intelligent and to be able to learn/adapt, robots must be able to process the new inputs that which they receive from the surroundings. Based on this input, the robots are able to conceptualize the data (which was not possible until very recently).

For such automatons, Machine Learning can be considered as the next step of Artificial Intelligence algorithms. The machine perception systems, themselves, are used together with Machine Learning systems in order to make the robots adaptable to an ever-changing environment.

There are quite a big number of robots that are already making this possible in a range of industries. For instance, a post was published by the University of California San Diego regarding a one-year-old robot that depended on machine perception. This robot is considered as the biggest miracle in the world of robotics by the authors of the post. It was pointed out that a number of advances were made by the robot, however, it still struggles to copy the complexity with which the human brain works. According to these authors, it may take a while to reach this high level of standard. However, the robot is considered a big breakthrough in the world of robotics.

There will be quite a log of companies that will rely on these intelligent robots in the future according to the founder of Kohli Ventures, Ted Kohli. He says these robots will play a big role in a number of industries in the future.

However, there are a number of challenges that Machine Learning and machine perception together face. These challenges need to be overcome by Artificial Intelligence engineers and the data scientists if they want to improve the functionalities of intelligent robots.

There are a number of ways in which scientists are trying to overcome these challenges. Early on, programmes made a big mistake by treating the early models of machine perception algorithms as a passive system. There was a lot of unwanted reliance on the biological systems upon which the algorithms were based.

The modern algorithms, however, have rather active learning and work on a multi-sensory process. With this in mind, the new robots are being developed using Machine Learning systems in conjunction with machine perception algorithms for greater intelligence.

Prasobh V Nair Chief Technology Officer
  • Last updated: Dec 4, 2018 16:05 IST
  • TOPICS: Around The World   News   Technology   Business  

SUBSCRIBE TO OUR BLOG

CONTACT