Artificial intelligence is a hot spot for Silicon Valley venture investors

An article in foreign media says a new trend is sweeping the valley. Looking at the recent financing of startups, you'll find that a concept that has more than half a century of history is in hot pursuit: Artificial intelligence.

"This is a popular investment area nowadays. "Stephen purpura, Stephen Purpra of the artificial intelligence company context relevant. The context relevant has accumulated more than $44 million trillion since its inception in 2012. More than 1700 startups have joined the AI wave, according to Purpra.

The new man of artificial intelligence believes that the technology is finally catching up with what humans expect of it, which will give the computer a higher degree of intelligence. They want to bring new human-computer interaction to humans, to enable machines to "invade" the human world in unexpected ways.

Daniel Nadler, Daniel Nadler, another emerging artificial intelligence start-up, said, "Technically, it's a paradigm shift from giving machines Kensho to automatically observing learning." "Kensho recently received 15 million dollars in funding, and it is chasing an ambitious goal: training computers so that they can replace white-collar jobs such as financial analysts."

"We do not refer to what we are doing as artificial intelligence, but rather to" automated human intervention-oriented knowledge-based work. "Nadler pointed out.

The herd effect of investors, to some extent, explains why artificial intelligence has become one of the hottest trends in venture capital after the "Big Data" slogan has ignited thousands of entrepreneurial dreams. The scale of those financing is relatively small, indicating that startups are mostly in the early stages. However, the number of companies to obtain financing, the broad background of investors, fully illustrates the strong interest of artificial intelligence investors.

In addition to some of the famous VC companies in Silicon Valley (including Khosla Ventures and Greylock) and Elon, such as Ilon Masc (Musk Thiel), Thiel (Peter Tech), Some of the active investors in artificial intelligence startups also include companies that think the technology is useful in their industry, such as Goldman Sachs.

Nadler said the tradition companies now need to invest in the field: limited partnerships want a slice of the latest "next big hit" in the tech industry.

Value Application Issues

The latest AI boom has largely benefited from new programming techniques for nearly "smart" machines. The first is machine learning technology, which involves training machines to identify patterns and make predictions by digging up a lot of data. But like other new buzz ideas that have spawned a start-up, many of those involved will be at risk of finding it difficult to find profitable applications for their technology.

"A lot of AI platforms are like Swiss Army knives," said Tim Tuttle, CEO of Expect Labs, who recently voted 13 million dollars, "they can do a lot of things, but it's not clear which are high-value applications." ”

The result, he says, is that the industry is as wild as the West, where entrepreneurs scramble to apply artificial intelligence to any computing problem they can think of.

Purpra added, "I don't think machine learning as an independent technology can create a worthwhile business." Many of the companies involved will end up in acquisitions. ”

The reason why people feel that AI will be more than just another fad is a craze for its broad potential. Like "Big Data", AI refers not only to a single technology or use, but a solution that can be widely used.

Matt Mack Wayne McIlwain, a Madrona partner at the Seattle venture, says technology such as deep learning can help companies get smarter inferences about their customers ' situations. He added that they would be able to identify customers ' preferences and make forecasts, such as when the customer is most likely to be contacted and which customers are most at risk of not signing a contract.

Startups rushing into the field face huge competition. The biggest advances in artificial intelligence come from Google, IBM, Facebook and other big technology giants. Instead of revealing how much they have invested in developing the technology, the companies have made public demonstrations to show the public that they are in the lead: a test that Google has designed to identify cats from YouTube videos, the deep face system that Facebook uses to identify faces, And IBM's Watson question answering system.

However, entrepreneurs such as Tuttle will want to rely more on packaging existing technologies for targeted applications rather than on cutting-edge development of new technologies. Expect Labs, for example, is dedicated to creating voice-activated services that enable businesses to make online conversation searches through their services.

"Big companies are trying to develop this technology to solve all the problems, and we are trying to solve different problems," Tuttle said. "Tuttle said.

The three most popular uses

The basic use of the technology can be divided into several different areas. Thanks to the progress of pattern recognition, image recognition becomes easier than ever. Vicarious, a start-up company involved in the field, recently completed financing 72 million dollars. It can solve the authentication code problem.

The same technique is also used to help computers "understand" language-natural language recognition problems. This is one of the technologies behind systems such as IBM Watson. The third popular use of artificial intelligence is to identify relevance--to personalize online content and other referral services, or to improve advertising-oriented efficiency.

As with the new concept of the foreseeable future, some of the early applications of artificial intelligence are also used in financial markets, but given the benefits involved, participants are afraid to publicly preach their technology.

"If your financial application works, why should you make it public?" "Sentient Technologies chief Scientist Babak Hojart (Babak Hodjat) pointed out. His company is working to get a lot of computing power from data centers to simulate the financial markets: by using an "evolutionary algorithm" that tries to learn how markets respond to different situations, it wants to develop models to predict the future evolution of the market.

Putting ideas like this into practice in every field requires a huge investment in the development of artificial intelligence technology. Sentient, for example, recently financed more than $100 million trillion to apply its technology to more areas, explaining that the cost of recruiting "trained" AI systems to apply to industry experts in a number of different areas is very high.

Sentient that the most attractive industries are those that need to dig up huge amounts of data to address high-value issues, such as healthcare, insurance and electricity. Computer security and fraud detection are among the areas that many AI companies want to dabble in.

The Purpra of the context relevant says there are other costs to making AI technology truly useful for real-world use. "The key contention is not about the underlying machine learning technology, but about building a support system to make it available." He says these assistive technologies include the data "pipelines" needed to transmit large amounts of information, as well as control systems to ensure that AI technology operates within acceptable business parameters.

Given that many startups are under pressure to prove that their technology is more than just stunning exhibits, the ability to grab financing from investors may determine their survival in the inevitable shuffle of the AI industry.

Machine learning

Richard Waters, the FT columnist, writes that artificial intelligence, machine learning, deep learning and neural networks are about developing machines to solve problems previously thought to be Richard Watters by the human brain, which spawned a series of technical and specialized terminology.

As with other branches of technology, the differences in the best programs are sometimes like differences in religious beliefs. "The name you use denotes the tribe you belong to." "Purpra said.

Artificial intelligence carries the dream of a complete human-type computer "thinking". But the attempt to decode the human mind with computer logic is not going well.

The industry's renewed interest in artificial intelligence is largely due to machine learning-a scheme designed to draw a parallel with human thinking. Machine learning is a product of the decline in information processing costs, involving massive amounts of digital data that can now be collected and transmitted online.

As a subset of machine learning, depth learning is an important reason for the emergence of AI trends. Deep learning is another concept based on the history of artificial intelligence: neural networks, the software that seeks to simulate the workings of the human brain to speed up "learning".

Nara Logics's CEO, Yana Aigues (Jana Eggers), says the advances in neuroscience have brought new ideas to this biological simulation. She added that the aim of the simulation was "to see how the human brain is making decisions and how to make computers do better." ”

(Responsible editor: Mengyishan)