Big Data Digest produced
Author: Song Xin instrument
In the future we may love our robots do? You should choose to change security or privacy invasion rejection of all life technology? In the future there will be weapons of mass destruction autonomy AI do?
Before the arrival of AI era, in addition to technological advances, more and more of these problems are being instituted. For this, AI Research Institute of Tsinghua University professor and academician Tang Jie Zhang Bo Tsinghua University jointly launched "AI Time" science debate, hope in the form of debate, discuss the conflict between artificial intelligence and human future, to explore the future of artificial intelligence.
The end of May, in a safe and ethical theme of the first event held in Tsinghua Science Park. Event invitations to the AI Research Institute of Tsinghua University academician Zhang Bo, president of Hillhouse AI School of Renmin University of China Wen Jirong and Sohu network products, technical director Yang Tian, who gives a wonderful discussion of these issues.
Fall in love with artificial intelligence?
"Her" stills
Discussion from the movie "her" to start the film about a love story between man and artificial intelligence. Actor Theodore accidental access to the latest artificial intelligence system OS1-- Samantha, she has a charming voice, gentle but humorous. Theodore and Samantha evolved from human friendship is a strange love for some not understood secular.
When the AI more progress, it may produce its own consciousness, and even human development out of feelings? 2018, an artificial intelligence drawn portrait auction of $ 432,500 out of the high price , whether mechanical algorithm behind the existence of the soul? Whether AI has been able to create their own and appreciate beauty?
Several scholars believe that currently do not need to worry about that, just science fiction movie, the future of artificial intelligence to generate emotion and the will is still very far away from us. The artificial intelligence is defined as the science fiction can help replace the tools of thought and even humans, but in fact we are still in the initial stage of artificial intelligence. "It's still restricted to artificial intelligence to complete the assigned task, and from the emotional self-awareness or to achieve the" Her "as the current technology is also difficult to achieve." Professor Wen Jirong explained.
Explore the nature of human love story or love, what is love, what is the United States, this is actually a philosophical question. But it considers these to the development of artificial intelligence is undoubtedly necessary, Academician Zhang Bo, philosophy should walk in front of science and technology. If we are not sufficient for your brain to understand, how to imitate the human brain in order to create artificial intelligence it?
Dr. Yang Tian think these trainers as artificial intelligence, we really do not know themselves what is the expression of emotions, how to pass, and now the application of artificial intelligence is in fact using an optimization problem to approximate one like this expression of the reality of the problem.
Obsessed with the pursuit of
The gap between the ideal and the reality does not obstruct our enthusiasm, although the distance of artificial intelligence mature within the foreseeable future, but many researchers are forward scored on this road tirelessly, always ready to wake fourth industrial revolution bell. Now artificial intelligence research in the outbreak of the May release of "China's new generation of artificial intelligence Development Report 2019" show national strategic layout further escalation of artificial intelligence, artificial intelligence is looking toward a new pattern of global distribution from a few big powers.
The idea of thinking machines church let us fascinated, whom we seek, what artificial intelligence will change? Where it's coming in?
Academician Zhang Bo now, although currently the rapid development of artificial intelligence, but can do more than people in some way, it is impossible to fully than people. AI can play chess, you can play games, but are certain restrictions in a large number of learning, under repeated training, it is still behind the people at work. We take the philosophy of behaviorism defined genre, if the machine exhibited behavior like people, we believe it has the intelligence. So we are doing is trying to make the machine learn to imitate human behavior.
Professor Wen Jirong proposed today that all discussions are based on our understanding of artificial intelligence, what is artificial intelligence, why he wrote a program called artificial intelligence, would not I? Is defined herein teach, from both the input and output to determine: If the output determination procedure is predictable, it is not artificial intelligence; If the output is blurred, with uncertainty is, the accuracy may be measure, such as face recognition 95%, it is artificial intelligence. This means that if a program able to cope with diversity and variability of input environment, while output also received relatively high precision, is artificial intelligence. "Inaccurate things, the results have a strong tolerance occasion is the future of artificial intelligence will be successful place."
Dr. Yang Tian said that the current definition of artificial intelligence, it can put a lot of things are doing good, for example, dialogue, speech synthesis. But very far away from the pass the Turing test. For now, voice-related technologies and technical aspects of the autopilot in fact, there are many practical applications from very close.
As just invented the steam engine, heard the loud bang and feel yourself being a glimpse of the people of the times of the industrial revolution a great revolution, we are witnessing perhaps the future is opening a great new era.
Worry: the pros and cons of privacy and security
Privacy and security choices facing the United States and China split on one end. May 14, San Francisco Board of Supervisors by a new ban, the city decided to ban the use of face recognition technology for all units, including the police and other government departments, launched the world's first face recognition bans city.
At the same time, China developed a comprehensive eye in the sky system, can capture the human face and match information in the first time through the application of artificial intelligence, image processing and other professional video technology. There are cameras in the classroom the students, the students time to capture every move; crossing the road accidentally ran a red light, your photos will be immediately identifiable information match, appeared on the big screen side of the road ......
Face recognition can be used to catch criminals, to protect the safety, installation of cameras in the classroom can also protect the students, to avoid similar incidents red, yellow, blue nursery tragedy. But this technology also has strong intrusive, a pair of eyes always watching you, get your information, but you never know what information will be used to do.
Zhang Bo academician spoke behind a different approach reflects the different values of the West, when the personal interests and public interests conflict, China has chosen the public interest, the Western emphasis on personal interests. How to balance the public interest and personal interests, it is a very serious, serious issues to be studied. Privacy and security we can not easily let go, to see who can do a better balance.
Professor Wen Jirong also believes that the current practice in the West are too extreme, some of the data currently collected is completely unnecessary. West because too much emphasis on privacy protection, a great impact on the development of artificial intelligence, big data industry, but China should not be simply to ignore the development of artificial intelligence industry and privacy protection.
Dr. Yang Tian believes that the current society we more or less get a recommendation from a personalized and convenient, this is in fact betrayed by the system in exchange for privacy. The presence of the camera does play a role in improving the security environment. Chinese have a saying "give the first three feet of the gods", science was not as developed in the past, when we believe they are watching every move of God above, and thus bound themselves do not do bad things. Now, we do not believe in God, and found himself without constraint, which is actually damaging to everyone. In this case, these techniques play a deterrent role, technology has become a new era of God.
Now the border is that, if you choose security, how do we ensure that the data will not be misused? A lot of things easier cross-border, how to keep our borders, is well worth careful thought.
Fear: the future of AI weapons
AI weapon is a commanding efficient, precise strike, operational automation, intelligent behavior and other characteristics of smart weapons, which can be consciously looking to identify the need to combat the goal, with the ability to distinguish natural language, it is for a thinking weapons systems.
Professor Wen Jirong express our fear of AI mainly from the fear of its autonomy, it may be out of fear of human control.
Recently, many celebrities have expressed their rejection of artificial intelligence weapons that robots will one day control humans. But in fact this argument is unreasonable, irrational at least in the short term. Professor Zhang Bo said that the will is a fundamental human war machine just manipulated tools. War no matter to what extent, are in accordance with the human will last conducted. Dangerous in people, not in the machine. The so-called unmanned aerial vehicles, ground control was also behind.
Perhaps the problem of artificial intelligence weapons that they will not control the world, but that they are very likely to be re-programmed, so anyone can be shockingly low budget, to create a highly efficient machine killing of the innocent. Machine destruction is enormous, Dr. Yang Tian question, we have not been able to allow in the case of human trials to make weapons to destroy a person?
UAV grading standards
Professor Wen Jirong think, should AI weapons grade, and is similar to the autopilot. According to the degree of processing in the face of environmental variability or diversity agent to be. Classification is to help us better understand the way a stage in which the artificial intelligence now and the next thing to do.
AI Time by a group of young people concerned about the development of artificial intelligence founder, aims to carry forward the spirit of scientific speculation, invited the public to artificial intelligence theory, algorithms, scene, nature of the problem applications to explore, debate, ideas collide, creating artificial intelligence knowledge sharing the source and gathering place.