I personally like watching western movies, but now the development of machine learning is just like the development of the western region. Shafi Goldwasser, director of the Institute of Theoretical Computing at the University of California, Berkeley, delivered a speech entitled Cryptography and Machine Learning at the Worlds Top Scientists Forum (Shanghai Drop Lake), reminding us of privacy and security issues in the era of big data.
The youngest Turing Award winner has subjectively observed that machine learning and big data have contributed to power transfer worldwide. Look at the companies with the highest market capitalization in 2001, energy companies like General Electric and ExxonMobil; big data companies in 2016. No wonder they say that data is oil and data is money.
Then, what kind of security threats do ordinary people face in the era of big data? I think everything is a threat. Gold Vasey laughed when he received an exclusive interview with www.thepaper.cn in October 31st. But I think the biggest thing is how to calculate data in the case of data privacy protection.
Womens perfect discipline
Safi Godwasser (ShafiGoldwasser) surging news reporter Sun Yiyun
In 1983, Gold Vasey entered the Massachusetts Institute of Technology (MIT) and got his professorship in 1993. Her research area is computational complexity theory, cryptography and computational theory of numbers. The probability encryption developed by her and MIT colleague SilvioMicali has set the gold standard for online transaction security.
Together, they won the Turing Award in 2012. This is the highest prize in the field of computer science, and Gold Vasey is the third female winner in history.
I think computers are a wonderful subject for women who are good at mathematics. This field is relatively young, very open, and has no rules. Most computer departments in Colleges and universities have strong women, as I met some of the top female scientists in cryptography in Carnegie Mellon, MIT and Berkeley. She said.
Gold Vasey articulate clearly, but speak very fast, and show full confidence and reason when answering questions. She often switches to a better grammar in the middle of a narrative without pausing, suggesting that thinking is even faster than speaking.
In Goldwaters view, computer science is a perfect field in both academia and industry. As long as it has strong analytical and technical capabilities, programming experience is not necessary: key or mathematics. Women may start late, but math is no worse than men.
Quantum computing threatens less cryptography
Some people worry that quantum computers will challenge existing cryptosystems. Gold Vasey is quite cautious about this technology hotspot.
She thinks that quantum computer is a wonderful theory. However, people still dont know how to build a large-scale, error-correcting quantum computer, and quantum algorithm cant run on quantum computer.
As far as RSA, the mainstream public key cryptosystem is concerned, the difficulty of deciphering codes depends on the ability to factorize the maximal integers. As far as I know, the current quantum computer can decompose the maximum factor to 143, that is, 11 times 13. So its still a long way from the factorization algorithm that quantum computers can really run, and its not clear whether we can see it in our lifetime. She said.
For this reason, in the forum session the day before, when Pan Jianwei, the executive vice president of the University of Science and Technology of China and the father of the worlds first quantum communication experimental satellite, mentioned that all passwords could theoretically be cracked by classical computers or even quantum computers, Goldwasser raised questions.
Thank you for your wonderful speech. As you mentioned at first, there is no computing problem that the computer cant solve. She stood up and said.
If you theoretically get all the computing time in the world, they will theoretically be cracked, Goldwater explained to the surging journalist. But in fact, there are limitations in the computation time. If there are enough cryptographic numbers generated by progressive statements, there is no effective classical algorithm to solve them. There are four new encryption systems based on geometry and lattice problems. Even quantum computers do not have corresponding decoding algorithms.
In general, as far as we know of quantum computers, we have a secure encryption system.
In view of another hot spot in science and technology, AI, Goldwater, starting from the top American universities he has experienced, puts forward some suggestions for the development of Chinese universities. Carnegie Mellon has always been known as the global AI base, is the alma mater of Li Kaifu, Shen Xiangyang, Hong Xiaowen, Lu Qi and many other Chinese AI experts; MIT has established the worlds first AI laboratory as early as 1959.
Goldwater thinks it is important that universities not only have a strong team of artificial intelligence, but also take into account both theoretical science and hardware systems. Artificial intelligence is now very important and successful, but it needs hardware, data protection, algorithm verification, and intelligent algorithms and systems. All sides are professional, in order to make good science. I think this is a lesson that Chinese universities can learn.
I cant avoid it either.
Like many of the big names in the computer industry, Goldwater is also involved in the industry, co-founding Duality Technology to provide collaborative computing services. This is a distributed computing for private data, which can share information among different organizations without leaking information, avoiding the hidden dangers mentioned at the beginning of this article.
She cited two examples: how to ensure that the cloud calculates correctly and does not leak data when uploading data to the cloud. Second, if someone leaves an electronic medical record in the hospital, and the gene sequencing company leaves genetic data, how to integrate the two databases without handing over data to each other?
Starting from the field of cryptography, she establishes a secure cooperative computing service for multiple sources of data. Such calculations can avoid collisions between satellites in countries without knowing each others satellites, understand whether pay structures achieve gender equality without revealing personal wages, and collect large amounts of DNA samples to find genes related to schizophrenia without revealing patientsprivacy.
When finally asked how ordinary Internet users can maximize their data privacy, Goldwater said it was a very important question: Ordinary individuals are currently not getting good advice. Every day everybody is facing the account password and privacy agreement. As long as you want to use some software, you have to agree to the user privacy agreement that jumped out. Goldwater believes that individuals should strive for a variety of systems and choose software that does not require access to user privacy. I know this is not easy. Power is not in the hands of individuals, but in the hands of enterprises. They take out a service, do you love it or not? The temptation to use new technology is too great for everyone to think about what they will actually lose. She said, I cant avoid it either. Source: surging news editor: Qiao Jun Jing _NBJ11279
When finally asked how ordinary Internet users can maximize their data privacy, Goldwater said it was a very important question: Ordinary individuals are currently not getting good advice. Every day everybody is facing the account password and privacy agreement. As long as you want to use some software, you have to agree to the user privacy agreement that jumped out.
Goldwater believes that individuals should strive for a variety of systems and choose software that does not require access to user privacy. I know this is not easy. Power is not in the hands of individuals, but in the hands of enterprises. They take out a service, do you love it or not?
The temptation to use new technology is too great for everyone to think about what they will actually lose. She said, I cant avoid it either.