Shulou(Shulou.com)06/02 Report--

On March 18, the "Human-oriented Institute of Industrial Intelligence" (HAI), one of the directors of Li Feifei, finally completed its official highlight moment after its launch. The official website, which was officially launched, also updated two blog posts a few days ago, one of which is a detailed introduction to HAI The other is a brief history of artificial intelligence in Stanford's HAI edition, which covers information such as artificial intelligence memorabilia, the layout of global artificial intelligence start-ups, the popularity of artificial intelligence, advances in artificial intelligence research and education, as well as Stanford's breakthroughs in artificial intelligence and artificial intelligence courses.

The reason why this picture is prefixed with "Stanford HAI version" is because this picture is full of "Stanford" halo. Although this picture is not detailed enough, it is enough for you to get a glimpse of the history of artificial intelligence and the current progress.

Let's take a look at it one by one.

Memorabilia of artificial intelligence

In 1955, at the Dartmouth Conference, John McCarthy, a respected computer scientist at that time, first proposed the concept of "artificial intelligence". After that, the meeting was also seen as a symbol of the official birth of artificial intelligence, and the proposer, John McCarthy, was also known as the "father of artificial intelligence".

In 1963, John McCarthy, the father of artificial intelligence, founded the Stanford artificial Intelligence Laboratory (SAIL,Stanford Artificial Intelligence Laboratory). The laboratory has experts in many fields, including robotics, computer vision, machine learning, image processing, natural language processing and so on. Representatives include a large number of people who are famous in the field of artificial intelligence, such as Christopher Manning, Wu Enda, Li Feifei and so on.

In 1966, MIT computer scientist Joseph Weizenbaum developed the first natural language processing program, ELIZA, which is a chat robot that simulates Rogers psychotherapy.

In 1967, E.A. Feigenbaum of Stanford University led the development of the first "expert system"-DENRAL, which changed the research of artificial intelligence from reasoning algorithm to knowledge.

In 1969, Arthur Bryson (Arthur Bryson) and Yu-Chi Ho (Yu-Chi Ho) proposed the most important algorithm in the field of machine learning-back propagation algorithm (Backpropagation). This algorithm can be used in multi-layer artificial neural networks, and the development of deep learning from 2000 to now is inseparable from its inspiration.

In 1973, the Stanford International Research Institute (Stanford Research Institute, SRI) developed the first mobile robot using artificial intelligence-Shakey.

The Stanford Cart, invented by Hans Moravec as a graduate student at Stanford University in 1979, automatically crossed a room full of chairs and drove back and forth for five hours without human intervention. Stanford Cart is the equivalent of an early self-driving car.

In 1988, the IBM Watson Research Center published Statistical methods of Machine Translation (A STATISTICAL APPROACH TO LANGUAGE TRANSLATION), which heralded the transformation of rule-based machine translation methods to probability-based methods, and reflected a broader shift from "machine learning" to statistical analysis based on known examples.

In 1991, the World wide Web invented by Tim Berners-Lee (Tim Berners-Lee) was launched for the first time.

In 1997, Deep Blue, developed by IBM, became the first computer program to defeat human chess champion Garry Kasparov.

In 1998, Kenneth Sarisbury (Kenneth Salisbury), a professor at Stanford University, published a patent for surgical robots (robotic surgery).

In 2005, Stanford University professor Sebastian Thrun and Stanford University AI Lab invented the first self-driving car to complete the 132-mile Mojava desert route, winning the DARPA Super Challenge (DARPA Grand Challenge).

In 2009, the ImageNet project led by Li Feifei created a database of 15 million photos covering 22000 items. This project and a series of subsequent work have affected the development of the whole field of computer vision.

In 2010, Apple launched Siri, an artificial intelligence assistant software built into Apple's iOS system.

In 2011, Watson, a natural language question and answer computer developed by IBM, appeared on the veteran American puzzle show "Jeopardy!" To defeat the human race.

In 2012, Jeff Dean and Andrew Ng published an experimental report showing 10 million untagged images to a large neural network and found that the network could recognize the image of cats.

In 2014, Amazon launched the smart speaker Echo and the intelligent voice assistant Alexa.

In 2016, AlphaGo, developed by Google DeepMind, beat go world champion Lee se-dol.

Layout of global artificial intelligence start-up companies

Citing survey data from Asgard CB Insights, a world-renowned venture capital research institution, HAI says that at present, more than 3600 artificial intelligence start-ups have emerged in the world, of which 1393 are in the United States, accounting for about 40%, ranking first; China has 383, accounting for about 11%, ranking second; while the United Kingdom has the largest number of AI companies among European countries, accounting for 245, accounting for about 7%; Israel has 362, accounting for about 10% There are only 131 in Canada, accounting for about 4%.

At the same time, Asgard CB Insights collected data on the number of artificial intelligence startups owned by American cities, including San Francisco with 596, New York with 180 and Boston with 102, followed by Los Angeles, Washington, Austin and Seattle with 73, 36, 36 and 35 respectively. In addition, according to Pricewaterhousecoopers,2018, American startups raised a total of $9.3 billion from investment companies.

In addition, with regard to the contribution of artificial intelligence to the global economy, a report released by Gartner predicts that artificial intelligence will contribute $15.7 trillion to the global economy by 2030, a figure that will exceed the current combined economies of China and India.

Progress in the Popularization of artificial Intelligence

With regard to the popularity of artificial intelligence, HAI mainly collects data on the number of times that the US Congress mentions "artificial intelligence" and the number of scientific and technological terms mentioned by investors of listed companies.

With the development of artificial intelligence, the frequency of mentioning "artificial intelligence" in the US Congress is also increasing day by day. Before 2017, the number of references to "artificial intelligence" in the US Congress was very small, with a maximum of less than 20 times. However, the number of references soared to nearly 100 in 2017, which also shows from the side that countries have raised the layout of artificial intelligence to the level of national strategic layout.

As for the number of scientific and technological terms mentioned by investors of listed companies, before 2011, the most frequently mentioned term was "cloud computing"; from 2011 to 2016 it was "big data"; from 2016 to 2017, "machine learning" was comparable to "artificial intelligence"; and after 2017, "artificial intelligence" was mentioned the most. These data reflect the development trend in the field of science and technology.

Research and Educational Progress of artificial Intelligence

In terms of artificial intelligence research and education, HAI cited AI index's statistics on the number of academic papers published and the number of university courses registered as "introductory courses in artificial intelligence". In terms of the number of academic papers published, the number of academic papers published on the subject of "artificial intelligence" has increased more than eightfold since 1996, while from 2012 to 2017, the number of courses registered as "introductory courses in artificial intelligence" in university courses has increased by as much as 500%. This undoubtedly shows the great enthusiasm of the academic circles for the research of "artificial intelligence".

Stanford's breakthrough in artificial intelligence

As for the breakthrough achievements in the field of artificial intelligence, HAI focuses on the research achievements of Stanford University, which is representative in the field of artificial intelligence, including:

Stanley, a self-driving car invented by AI Lab of Stanford University, won the championship of DARPA Super Challenge (DARPA Grand Challenge).

ImageNet led by Li Feifei, a professor at Stanford University, changed the course of machine learning / AI and led the era of deep learning.

A total of 18 members of Stanford University's AI Lab won the ACM Turing Award.

Stanford University in conjunction with Stanford Medical School has developed the first supercomputer for medical AI

Stanford International Institute (Stanford Research Institute, SRI) has developed the first real mobile robot.

Stanford University released Stanford CoreNLP, a leading open source natural language processing toolkit, in 2010.

Stanford University has developed a humanoid diving robot, OceanOne, which has a tactile feedback function and can explore the underwater world with high fidelity.

Stanford artificial Intelligence course

In 2018, Stanford's artificial intelligence-related courses covered 54 research topics, including computer vision, natural language processing, advanced robotics and computational genomics. Among them, the most popular course is "CS 221: principles and Technologies of artificial Intelligence" launched by Wu Enda.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.