This is not a good time for physicists, said Oscar Boykin, a former physicist who studied physics at the Georgia Institute of Technology and, in 2002, at the University of California, Los Angeles Ph.D. in physics, Ph.D. Physicists running the Large Hadron Collider in Switzerland discovered the Higgs boson in 2013. It is well-known that the Higgs boson was first used by physicists The predicted sub-atomic particles, which everyone expected to find out after that, did not, however, have a theoretical model that would subvert the existing universe, nor change anything or bring physical Any new information for the scientist, Boykin admits: 'Physicists are excited when something really goes wrong in physics, and now it is frustrating that we are right in the middle of a period where there is not too much mistaking in physics. This is very frustrating to a physicist. "In addition, physics is not paid well enough to be a profession.
So Boykin made his way to Silicon Valley as a software engineer and for other physicists it is a good time to be a part of it.
Boykin works at Stripe, a $ 9 billion startup with a primary focus on online payment acceptance and Bojkim, who develops and runs a software system that collects data about the entire company's services, Predict these service data and analyze when fraudulent transactions occur and guard against them.As a physicist, he is well suited to this task, which requires both extreme mathematical thinking and abstract thinking.However, unlike the occupational physicist He is working in a field that offers endless challenges and endless possibilities, and more importantly, high salaries.
Silicon Valley has become a Hadron Collider if both physics and software engineering are considered as subatomic particles and Boijin works with three other Stripe physicists in December last year and General Electric When acquiring machine learning startup Wise.io, chief executive Jeff Immelt boasted that he had just caught up with a company of physicists, most notably the University of California, Berkeley The astrophysicist Joshua Bloom, an on-campus open source machine learning software used by about 70,000 data scientists around the world, is based on the Swiss physicists who worked at SLAC's National Accelerator Laboratory Developed with the help of Arno Candel, Vijay Narayanan, director of data science at Microsoft, is also an astrophysicist with several other physicists in his department.
It's all happening in Silicon Valley, and it's no coincidence, because virtually every Internet company's needs fit in with the skills that physicists possess, both structurally and technically.
natural factors
Of course, physicists have long played an important role in computer technology, just as they have played an important role in many other areas. John Mauchly, who helped design ENIAC, one of the first computers in the world, A physicist Dennis Ritchie, the father of the C programming language, is also a physicist.
But for physicists in the field of computer technology, it is now the golden age for a switch to Silicon Valley, thanks to the rise of machine learning technology, where computers begin to learn tasks by analyzing vast amounts of data.New wave of data science and artificial intelligence is exactly Suitable for physicist things.
In addition, the industry has embraced neural network software designed to mimic the structure of the human brain, but these neural networks are in fact only math applications, with the main associated disciplines being linear algebra and probability theory. Computer scientists do not necessarily accept these subjects Training, but physicists must understand the corresponding disciplines.Bueijin said: 'For physicists, the only real new thing in neural network technology is to learn how to optimize these neural networks and how to train, but for physics Home is a piece of cake, one of which is called 'Newtonian', the physicist Newton.
Chris Bishop, head of the Cambridge Research Lab at Microsoft, had been together 30 years ago when Deep Neural Networks had just shown promise in academia, allowing him to jump from physics to machine learning. 'It is quite natural for physicists to enter the field of machine learning,' he said. 'Even more natural than computer scientists.'
Challenge space
Ten years ago, Boydin said many of his friends in physics are fond of setting foot in the financial world, and the same mathematical style is also very useful for predicting market trends on Wall Street. One of the key ways is the Black-Scholes option pricing Model Black-Scholes, which is an important way to determine the value of financial derivatives, but it is Black-Scholes option pricing model to promote the financial tsunami of 2008. Now, Bojamin and other physicists have said that more Colleagues are turning to data science and other types of computer technology.
Earlier, physicists switched to leading technology companies to help develop so-called 'big data' software that can be used to process data on hundreds or even thousands of machines. On Twitter, Bojn Kim helped to develop a program called Summingbird systems, while three physicists at MIT Physics Department developed similar software for a start-up called Cloudant Physicists know how to process data - the founder of Cloudant, whose previous job was to work on large Large datasets at Hadron Colliders - developing these extremely complex systems require considerable abstraction from the mind of developers, and when these systems are built, many physicists use the data they have in their possession.
One of the key players in developing a large-scale distributed system for a company's farm during the early days of Google was Yonatan Zunger, who holds a Ph.D. in string theory from Stanford University.When Kevin Scott joined the Google advertising marketing team , Which is primarily responsible for collecting data from across Google and using it to predict which ads are most likely to receive more clicks, and for which reason Scott has recruited a large number of physicists who, unlike many computer scientists, have very limited skills Suitable for machine learning. "It's almost like laboratory science," said Scott, the incumbent chief technology officer.
Now that big data software is commonplace, Stripe uses the open-source system Bojkin helped develop, an open-source system that helps many companies' machine learning models improve their predictive power, providing a broader development for physicists in Silicon Valley In Stripe, Boykin works with many physicists such as Roban Kramer (Ph.D. in Physics), Christian Anderson (Harvard Physics) and Kelley Rivoire (MIT). They found themselves well adapted to work in Silicon Valley and paid higher here, as Bojkin said: 'The high wages are outrageous.' But they came here because of so many problems to solve.
future
Today, physicists are entering Silicon Valley companies and will take over Silicon Valley in the coming years.M machine learning will not only change the way the world analyzes data, but it will also change the way software is built.Neural networks are already reshaping image recognition, speech recognition, Machine translation, and software interface.As Chris Bishop of Microsoft said, software engineering is shifting from logic-based manual code to probabilistic and uncertainty-based machine learning models. Companies such as Google and Facebook are starting to use this new Way of thinking to re-engineer their own engineers and ultimately, other areas of the computer world will follow suit.
In other words, more and more physicists entering Silicon Valley mark major changes in the computer industry. Soon all Silicon Valley engineers will be physicists.