Foreign media have published articles saying that a new trend is sweeping silicon Valley. If you look at recent startup funding, there’s one concept that’s been around for more than half a century: artificial intelligence.

“This is the hot investment area right now.” “Said Stephen Purpura of ARTIFICIAL intelligence firm Context Relevant. Context Relevant has raised more than $44 million since it was founded in 2012. More than 1,700 startups have jumped on the AI bandwagon, according to Pulpura.

New entrants to artificial intelligence believe the technology is finally catching up to what humans have come to expect of it, and that it will bring greater intelligence to computers. They want to bring new ways for humans to interact with machines, and they want to allow machines to “invade” the human world in unexpected ways.

Daniel Nadler of Kensho, another emerging A.I. startup, said, “Technically, it’s a paradigm shift to go from feeding instructions to machines to having computers automatically observe learning.” Kensho, which recently raised $15m in funding, is pursuing an ambitious goal: training computers so they can replace white-collar jobs such as financial analysts.



“We don’t call what we’re doing artificial intelligence, we call it ‘automated human intervention knowledge work. ‘” Nadler notes.

The herd mentality of investors partly explains why artificial intelligence has become one of the hottest venture capital trends after the slogan “big data” ignited millions of startup dreams. The amounts raised are relatively small, suggesting that most of the startups are in their early stages. Still, the sheer number of companies that have secured funding and the breadth of investor backgrounds speaks volumes about investor interest in AI.

In addition to some of Silicon Valley’s best-known venture capital firms, including Khosla Ventures and Greylock Partners, and tech moguls like Elon Musk and Peter Thiel, Some of the active investors in AI startups also include companies that see the technology as useful in their industries, such as Goldman Sachs.

Nadler says vc firms need to invest in the space now: Limited partners want a piece of the tech industry’s latest “next big thing.”



Value application problem

Much of the latest ai boom comes from new programming techniques for near-” intelligent “machines. Top of the list are machine learning techniques, which involve training machines to recognise patterns and make predictions by mining vast amounts of data. But as with any hot new concept that has spawned a bunch of startups, many of the companies involved risk having trouble finding profitable applications for their technology.

“A lot of A.I. platforms are like Swiss Army knives,” says Tim Tuttle, CEO of Expect Labs, which recently raised $13 million. “They can do a lot of things, but it’s not clear what the high-value applications are.”

As a result, he says, the industry is like the wild West, with entrepreneurs scrambling to apply AI to every computing problem they can think of.

Pulpura added, “I don’t think machine learning as a stand-alone technology is going to be a valuable business. A lot of these companies are going to be acquired.”

The sense that ARTIFICIAL intelligence is going to be more than just another fad stems from a broader consideration of its potential. Like “big data,” ARTIFICIAL intelligence refers not to a single technology or use, but to a solution that could have a wide range of uses.



Matt McIlwain, a partner at Madrona, a Seattle-based venture capital firm, said technologies like deep learning could help companies make smarter inferences about customers. They will be able to identify customers’ preferences and make predictions, such as when they most want to be approached and which customers are most likely not to renew their contracts, he adds.

The startups that have piled into the space face a lot of competition. The biggest advances in AI have come from big-spending tech giants like Google, IBM and Facebook. The companies haven’t disclosed exactly how much resources they’re putting into developing the technology, but have staged some public demonstrations to prove they’re ahead of the game: Google’s test to identify cats from YouTube videos, Facebook’s Deep Face system for recognizing faces, and IBM’s Watson question-and-answer system.



Giiso, founded in 2013, is a leading technology provider in the field of “artificial intelligence + information” in China, with top technologies in big data mining, intelligent semantics, knowledge mapping and other fields. At the same time, Giiso’s research and development products include editing robots, writing robots and other artificial intelligence products! With its strong technical strength, the company has received angel round investment at the beginning of its establishment, and received pre-A round investment of $5 million from GSR Venture Capital in August 2015.

However, entrepreneurs such as Mr Tuttle will pin their hopes more on packaging existing technologies for highly targeted uses than on the cutting-edge development of new technologies. Expect Labs, for example, is working on voice-activated services that allow companies to let customers do things like search online conversations.

“Big companies are trying to develop this technology to solve everything,” Tuttle said. “We’re trying to solve different problems.” Tuttle said.



Three Hot uses

The basic uses of the technology can be divided into several different areas. Thanks to advances in pattern recognition, image recognition is easier than ever. Vicarious, a start-up in this space, recently raised $72 million. It solves the captcha problem.

The same technique is used to help computers “understand” language – natural language recognition problems. This is one of the technologies behind systems such as IBM’s Watson. The third popular use of AI is to find relevance — to personalise online content and other recommendation services, or to improve the efficiency of advertising targeting.

As with promising new concepts, some of the early uses of AI are in financial markets, though participants are afraid to talk about their technology publicly because of the stakes involved.

“If your financial app works, why would you want to make it public?” Babak Hodjat, chief scientist at Sentient Technologies. His firm aims to get lots of computing power from data centres to run comprehensive simulations of financial markets: by using “evolutionary algorithms” that try to learn how markets react to different situations, it hopes to develop models to predict how markets will evolve in the future.



Putting ideas like this into practice in various fields requires a huge investment to develop AI technologies. Sentient, for example, recently raised more than $100m to apply its technology to more areas, illustrating how expensive it is to recruit, for example, the industry experts needed to “train” AI systems in many different areas.

Sentient believes the most attractive sectors are those that need to mine huge amounts of data to solve high-value problems, such as medical insurance and ecommerce. Computer security and fraud detection are also areas that many AI companies want to get into.

Pulpura, of Context Relevant, says there are other costs involved in making AI technology available for real-world use. “The key battle is not about the underlying machine learning technology, but about building the supporting systems to make it usable.” These assistive technologies include the data “pipes” needed to transmit large amounts of information, as well as control systems to ensure that AI technologies operate within acceptable business parameters, he said.

With so many startups under pressure to prove their technology is more than just a showpiece, securing funding from investors could make or break their survival in the inevitable AI shakeout.



Machine learning

Artificial intelligence, machine learning, deep learning and neural networks — all about developing machines to solve problems previously thought only to be solved by the human brain — have given rise to a range of technologies and jargon, writes Financial Times columnist Richard Waters.

As with other branches of technology, differences in what companies do best can sometimes resemble religious differences. “The name you use indicates the tribe you belong to.” Pulpura said.

Artificial intelligence carries the dream of fully human-like computer “thinking.” But attempts to decode the human mind with computer logic have not gone well.

Much of the industry’s renewed interest in ARTIFICIAL intelligence is due to machine learning, a deliberate separation from analogies to the human mind. Machine learning is a product of the falling cost of information processing and involves vast amounts of digital data that can now be collected and sent online.

As a subset of machine learning, deep learning is an important reason for the emergence of ai trends. Deep learning is based on another concept from the history of ARTIFICIAL intelligence: neural networks, software that seeks to mimic the workings of the human brain to speed up “learning”.

Jana Eggers, CEO of Nara Logics, says advances in neuroscience have led to new ideas for this biosimulation. The goal of the simulations, she added, is to “see how the human brain makes decisions and how computers can do them better.”