《未來(lái)機(jī)器人智能能否戰(zhàn)勝人類(lèi)智能.ppt》由會(huì)員分享，可在線(xiàn)閱讀，更多相關(guān)《未來(lái)機(jī)器人智能能否戰(zhàn)勝人類(lèi)智能.ppt（29頁(yè)珍藏版）》請(qǐng)?jiān)谘b配圖網(wǎng)上搜索。

1、1/29 Robotic Intelligence VS Human Intelligence Trista 2011-5-24 2/29 Part 1 Part 2 Part 3 Part 4 Evolution of Robots In Movies /10 stages /causes & results Insights /Say Yes: Strong AI /Say No: Weak AI Our Point of View /Tool Properties & Subordination /Cannot Self-evolve /No Intentionality /No

2、 Sociality The Way Out /We need not /We need /Man-Machine Symbiotic System 3/29 Evolution of Robots In the Movies Causes Science frictions: Exaggeration of threats High-techs: Moores Law,Nanotechnology, Quantum Calculations Works: Interaction robot, Cook robot, Talent robot, Sexual robot Resul

3、ts Attitude to robots: Love yet fear, desire yet reject. Worries: Ultra-intelligent robots will be a threat to the existence of human civilization. 4/29 Whether will the robotic intelligence surmount the human intelligence or not in the future ?? Your standpoints? 5/29 Say Yes : Hu

4、go de Garis Ray Kurzweil Stephen William Hawking 6/29 Hugo de Garis: Father of the artificial cerebrum His prophecy: Super AI will bring war and replace humans as the dominant species. He is crazily believe in ultra- intelligent machines, whose thinking and learning abilities are trillion s

5、quares times of humans as he thinks. And those machines will eradicate human beings as easily as kill mosquito, without any sorry. Its similar to a kind of robot worship religious idea. 7/29 Ray Kurzweil: Inventor & Computer scientist The turning point of information technology will arrive in 20

6、45. Then artificial intelligence will surpass the human brain, completely changing the significance of humanity. Humans and machines will merge into a super human, and under the help of the developing technology, humans can obtain eternal life. Idealism world outlook Both human thoughts and relig

7、ious beliefs can be calculated by mathematics. 8/29 Stephen William Hawking : The greatest scientist If the extreme complex chemical molecular can act in the human body and come into being intelligence, then so can the rather complex electronic circuits do. Propose to modify the human gene, other

8、wise we may be surmounted by the computer. 9/29 Strong AI They think that humans can create the intelligent machines with self-awareness, truly capable of reasoning and solving problems. Under the external environmental stimulus, they can think and accumulate unceasingly. Through self-study,

9、they can enhance their own intelligence. Representative figure: Turing Turing Test If a computer can make tester believe that it is a person, namely the computer possesses intelligence. 10/29 Say No: Weak AI They think that humans can produce the intelligent machi

10、nes endowed with some kinds of intelligence, which is fixed, cannot be changed, reacting to some designated external environmental stimulus according to programs. It seems to be some kind of intelligence, but without self-awareness and consciousness. Moreover, let alone with the true ability to reas

11、oning and solving problems. Representative figure: John R.Searle Chinese Room Model 11/29 1. Chinese characters are sent into the room. Chinese Room 2. According to operating rules, change these input Chinese characters into some other Chinese characters. 3. Sent the new Chinese characters out o

12、f the room. 你懂中文嗎 ? 懂，當(dāng)然懂 ! Chinese Room Model Doesnt know Chinese at all. 12/29 Our Point of View: Humanism The function of the human brain will continue to be understood, thus robotic intelligence will continue the approach to human intelligence. And it can exceed human intelligence on the loca

13、l scale , such as precision fast computing power, a powerful storage capacity, etc. However, based on dynamic analysis, the development and evolution of human intelligence are also constantly.Robotic Intelligence as the principal object of human intelligence product, its role and function is subj

14、ect to the constraints of human intelligence. So it must be less than human intelligence. 13/29 Talking Points a) Tool properties & Subordination The design principles and motivation of robots are to complement and supplement human intelligence rather than as a will to replace human intelligence.

15、 The human allot those junior human intelligence tasks to robots, and spend more time on senior creative activities. On the initiative of creating, people show the irreplaceable advantages. Small Story: The source of Robot 14/29 Examples ABB industry robots Classification by role: welding, grin

16、ding, coating, handling, sorting, assembly, packaging, etc. Properties: accurate, fast, flexible, efficient, stable. make up for deficiencies on the human physical labor. Famous advertising: truly right move , power and productivity However: The motion of end-eff

17、ectors lack of closed-loop control as human visual system, which causes some blindness on reaching destinations more or less. Compare with their fixed jobs, we human deal with the senior tasks: eliminate errors, product design and development. 15/29 NASA Spirit Mars Rover Honorary but onerous task

18、 Looking for possible life on Mars. 16/29 Why cannot the human land on Mars till now? Due to the constrains of human physiological structure, the environment on Mars is a great challenge to mankind, such as weightlessness, intense radiation, large temperature difference between day and night.

19、Undeniable that robots play as powerful tools in assisting the human exploring Mars. However, compared to spirit s workdrilling rock,take photos and some other kinds of collecting work, we human deal with the senior worka series of sophisticated analysis and scientific researches of the photos sent

20、back. New Scientist :To answer the problem about life on Mars,the only way is the human person to set foot on Mars to have a look. 17/29 Sleep-Wake Intelligent Robot The robot can act as an actor, presents peoples activities in the dream, which can help experts analyze dreams. Robots can make

21、 use of EEG(腦電圖 ), MRI (核磁共振 ) to capture and simulate dream space , but cannot make up sweet dream to invite people in. It cannot be true to perform Inception. 18/29 b) Cannot self-evolve really, but much similar to system-upgrade Although some robots perform some analogous human intelligence,

22、 such as learning, inference and self-tuning. But after all, these are owed to the corresponding algorithm endowed by engineers. ANN Whether can ANN function well or not largely depends on its weights. Furthermore,the key part of gaining appropriate weights is the learning aigrithm,which developed

23、 by the human,namely,limited by human intelligence. When humans come up with better algorithms, robots will show higher intelligent behaviors. Robotic Intelligence materialization of Human Intelligence 19/29 Deep Blue VS Mr Kasparov In 1997,Mr Kasparov lost to Deep Blue by only a round of chess ga

24、me in the first Human-Machine Competition. In 2003,Mr Kasparov competed with Deep Blue II again. This time, Kasparov tried to break the routine, taking some tactics contrary to his previous style of chess, which almost made Deep Blue II fail. After the second game, the two programmers supplemented

25、 to the programs again. In results, Kasparov and Deep Blue II broke even. Thinking On Man vs Machine 20/29 Two Tips: After all, this was a competition between human and human intelligence essentially, for which Deep Blue assembled two engineers intelligence at least. Since 1997, Deep Bl

26、ue defeated Kasparov, IBM and FIDE (國(guó)際棋聯(lián) ) found out the huge business opportunity that Man vs Machine contained.And they earned a windfall. Deep Blue won the chess game, but humans won the final in the market game. Deep Blue may never understand this. 21/29 c) Without Intentionality (consciousness,

27、 perception) Robots have no self-consciousness. they imitate the input and output process of human intelligent behaviors only in the form, not aware of their behaviors. For all kinds of objects, tangible or intangible, robots regard them as different logic symbols merely, rather than knowing thei

28、r true meanings as we human beings. watermelon toothbrush 22/29 Example Robot Hai Bao Expo star 23/29 Language Interaction: Mastering more than 3,000 daily dialogues, greeting people with enthusiasm in six languages (Chinese, Japanese, Korean, English, French, German ) and making self-in

29、troduction. The so called language interaction is actually a kind of key- words-match work for Hai Bao. He need not and does not understand the talking contents that he indeed performs. What if you read poems to him? Will it become you re whistling jigs to a milestone? After all,poetry appreciati

30、on(especially Chinese poems) is a kind of senior human intelligence. 24/29 Facial Expression Recognition Rich facial expressions base. Corresponding reaction to tourists by recognizing their facial expression and voice. You smile, he smiles, too. Robot cannot perceive the true meaning of a s

31、mile.What he do is extracting feature parameters of human facial expression and voice,followed by logical match jobs. A sincere smile and a hypocritical smile may make no difference to robot.But we human are surely fed up with the latter one. 25/29 d) Without Socialitykey problem Intelligence is

32、developed through learning, training and thinking unceasingly in society. In terms of dialectical materialism(辯證唯物 主義觀 ), it is an product of the development of objective human society. Current research on robots Tend to solve specific problems, and has made an achievement, but without human intel

33、ligence. Try to create human intelligent robots by imitating the hardware structure of human brain . The differences between human and robots, not the technology, but the sociality. But both are ignoring the humans biological and social nature. 26/29 Story Time from Endhiran Figure: Chitti, a

34、 trapped girl, the girls mother , the masses & media Event: a fire disaster Topics: Indian traditional civilization, social values (Self-respect VS Life) Thinking on the story: The wisdom of human brain does not built up on brain mechanism alone. Though Chitti has a powerful ne

35、ural scheme, its still a humanoid, not human. It doesnt understand shame and emotion, which need to be accumulated in society. Its a unbridgeable gulf between robotic intelligence and human intelligence. 27/29 The way out: People-oriented, man-machine harmony We dont need: A provided conclusion.

36、Simply affirming and denying make no sense. Fear of the extinction of human beings. Even if the technical issues allow, our government our ethics wont, just like the way they treat nuclear weapons and clone man. We need: The combing of concept & the expression of faith. Grasp the future world

37、with vivid subject consciousness. Find the way to develop human intelligence certainly in the uncertain reality. To create better life for humanity. 28/29 Man-Machine Symbiotic System Complement with each other Partnership Human play a leading role Lets fight against disease & disater together!! 29/29 Special thanks : Dear Pro. Cai & Xie Bin