According to an IT industry report recently released by IDC, by 2019, 40% of global digital transformation projects will use artificial intelligence services. By 2021, 75% of commercial enterprise applications will choose artificial intelligence, and more than 90% of consumers will Interact with customer service robots, and artificial intelligence will fully penetrate the smart home.
In order to respond to the tremendous changes brought about by artificial intelligence, the Chinese government has issued a “New Generation Artificial Intelligence Development Plan.” According to the plan, by 2030, China will achieve an artificial intelligence core industry scale of more than 1 trillion yuan, stimulating the size of related industries to exceed 10 Trillions of dollars. The plan also shows that by 2030, China's artificial intelligence theory, technology and applications should reach the world's leading level overall, and become the world's leading artificial intelligence innovation center.
However, when it comes to the topic of robotics and artificial intelligence (AI), public opinion is often divided into two camps: One thinks that humans will benefit from the development of machine intelligence, and the other thinks that machine intelligence will not even destroy human society. The foundation of the project will definitely affect people's employment opportunities.
Specifically, on the one hand, many repetitive or purely transactional tasks are replaced by automation; on the other hand, some tasks that are physically demanding or heavy or even dangerous will be completed by robots. According to a survey released by the University of Oxford It shows that in the next 20 years, 47% of employment opportunities in the United States may be replaced by robotics and automation technology. However, automation will replace some jobs while also creating some job opportunities, such as more areas of technology and innovation, including training, Construction, deployment, maintenance, maintenance, development, etc. Then how do we properly view the development of intelligent robots?
The increasing integration of robots and artificial intelligence will further stimulate the birth of autonomous robots. These robots will be able to make a variety of higher-level behaviors and reactions, such as learning experiences, parsing and further generating language, thereby creating an operation that is similar to autonomous consciousness. status.
In order to eliminate the opposing views held by cognitive robots, e-Research conducted a series of investigations on the most revolutionary robot and artificial intelligence solutions developed specifically for hospitals, workshops and workplaces. Key trends to assist robotics and intelligent automation in the next five years.
1. The appearance and movement of the robot will soon be more similar to humans
Unlike what we see in movies, at present most robots don't look like humans or even humanoids. Although robotic robotic arms in most automated manufacturing fields can perform lifts, tightening or welding jobs, etc. , but it does not have human characteristics. However, this situation is changing. With the replacement of robots in hospitals, nursing homes and other more services and social scenes, the anthropomorphic characteristics of robots and a more 'humanized' response model. become more and more important.
However, it is still too early to assert whether humanoid robots will increase or alleviate some people’s concerns about robots, especially when autonomous humanoid robots begin to join the workplace of ordinary humans. The benefits of robotic workers are obvious – they Never tired, rarely make mistakes, and will not ask for a raise, but it is also this perfect feature of humanoid robots that can make real humans hostile.
Not only do robots gradually possess more human characteristics, they also rapidly gain access to human intelligence through advanced artificial intelligence. Deep reinforcement learning is one of the areas of machine learning inspired by behaviorist psychology, drawing on the positive or negative effects of animals and humans. The way to learn behaviors in the results. Using this method, a sensor-equipped robot can learn how to move forward through trial and error, and associate positive results with behaviors that cause the result.
By combining deep neural networks based on data association assessment algorithms, deep reinforcement learning enables the machine to have the ability to generate intelligent behaviors, and formulate smarter decisions by modeling human psychology, thus solving complex situations. 2016 defeated the world. AlphaGo, the defending champion of the Weiqi Tournament, is also a computer program based on trial and error and continuous analysis of previous games.
Reinforcement learning is used in the robotics field to make the machine's behavior pattern more and more similar to the conscious body, which helps deepen the application of robots in areas such as automatic driving and industrial automation. With the continuous integration of robots and other advanced technologies, they can not only perform Highly complex tasks and learning experiences will also be able to take on jobs that require high-level decision making. This is particularly useful for work environments that are dangerous or unfit for humans.
As the social technologist Sherry Turkle pointed out: 'The face of the robot can trigger our imagination: Let us believe that robots can think from the perspective of humans, and vice versa.' As the development of artificial intelligence becomes more mature, humans and robots There is an increasing possibility of implementing role interchange in various situations.
2. The Internet of Things powers robots from functional to social
Currently, most consumer-oriented robots are designed to perform purely functional tasks, transporting drones from the air to Roomba vacuum cleaners. The vast majority of robotic R&D efforts are purely out of service and do not consider communicating with people.
In the next five years, the relationship between robots and humans will change. With the combination of advanced voice and image recognition software and artificial intelligence, and being widely adopted by the robotics industry, consumer-oriented functional robots will soon become social companions. Change. With the rise of the Internet of Things, consumer robots have become more intelligent and more context aware.
3. Low-cost hardware will boost the boom of consumer robotics
Most importantly, the cost of various hardware components continues to decline, which will become a major driver for the development of consumer robotics. With the rapid development of smart phones, many common robot components, such as cameras, sensors, speech recognition modules Due to the large number of mobile phone applications, the cost has been reduced to a certain extent. Therefore, possession of a private robot is no longer a multi-million dollar patent. The common people in the world can own their own robots in the near future.
In addition, according to research published by Dmitry Grishin, head of Grishin Robotics, the world's largest robot-specific risk investment fund, the widespread adoption of low-cost electronic devices such as the Raspberry Pi and Arduino open source development boards will also promote the development of the robot market. These powerful lows Cost electronic devices allow enthusiasts to develop their own autonomous devices at a fraction of the cost of the past, thus facilitating the development of self-made robot movements.
The cost reduction will attract more people to participate in robot R&D. It is expected that a large number of startups focusing on robotics will emerge in the next five years. However, low-cost circuits may not be enough to support this momentum. Adequate and efficient power supplies are still consumer products. An important problem that robots need to solve is that many robots are still plagued by 'weight-power ratio'. In addition, in order to realize the 'humanization' of robot behavior, it is necessary to run complex artificial intelligence algorithms, which will cause mass processing work, and thus lead to At present, the research and development of cognitive robots is limited to well-equipped research laboratories.
Unfortunately, although Moore's law may help to significantly reduce the size and cost of internal circuits, battery, motor and processor still have a long way to go to make consumer robots a reality.
