From the man-machine war, who will win the driverless

The green beans in "1Q84" need to go back to the physical origin to change the fate of today's destiny, but even if the human top player Ke Jie will come again, he will not see the way to defeat AlphaGo - despite the second game of man-machine war. Complex, the representatives of artificial intelligence also lamented that the machine was "forced to the limit."

Several key victories of artificial intelligence, placed in the unmanned field, also give people a glimpse of Intel's odds.

In the first game, the artificial intelligence was bursting, but there was a game of chess that made the chess world stunned. (As shown below) If the white chess red square is in the position of the black triangle of the black chess, there is a chance to escape five white children, which is a big advantage. However, artificial intelligence simply gave up, walked in the red square, and gave the key to black. Chess masters suspect that this is a bug, and the choice of artificial intelligence is the "stable" that unmanned pursuit. The final 1/4 win is also the result of the steady and dominant artificial intelligence.

In the picture, if the red square is stable, it cannot be shaken like a mountain. This stability is the key to driving a driverless car.

Since unmanned vehicles need frequent data updates and system updates, OTA (over the air download technology, that is, wirelessly upgrade the current system) will be the key to the vehicle's ability to get on the road. Not long ago, Wind River Helix CarSync, shown on Intel's Driverless Experience Day, is a wireless platform for maintaining automotive software. It features a car-to-cloud platform for efficient and secure OTA software and firmware updates with integrated encryption, authentication and integrity safeguards.

Demonstration of Wind River Helix CarSync*

Imagine a vehicle that is driving unmanned and automatically upgrades software or systems. At this point, the new system or new software pushed from the cloud must be complete and available. The intermediate link must be secure and stable, and the refreshed operation must be smooth. Moreover, the entire push and installation process must ensure that the source is reliable and free of defects. Just like the "mystery hand" of artificial intelligence mentioned above, unmanned driving must first ensure safety and stability. Any mistakes made here are unacceptable to human passengers.

At the end of the second game of the man-machine war, the two sides finally settled on the great difference: the human chess player used two hours and fourteen minutes, and the artificial intelligence took 52 minutes. This is equivalent to artificial intelligence spending 40 seconds per hand, which is the average time it takes to calculate all the changes on the board. The human chess player takes about 104 seconds per hand.

Humans need more time to make more changes. Therefore, the strategy of a game of chess, the middle of the game and how to allocate time is crucial. Time is the effort outside the board, but artificial intelligence has no sense of time.

As a result, many chess players joked, how can artificial intelligence be applied to the artificial intelligence, is it limited to artificial intelligence?

Yes, that's right!

Unmanned driving is on the road, and it is the life-saving energy problem.

In-vehicle artificial intelligence seems to be the brain of an unmanned vehicle, and it is indispensable to identify objects and plan driving plans. Intel's hybrid architecture of FPGA+CPUs delivers optimal power and performance to support artificial intelligence workloads.

Intel's FPGA+CPU hybrid architecture delivers optimal power and performance to support artificial intelligence workloads

At the most basic level, the artificial intelligence in the car uses machine vision and deep learning to help the car understand its environment. These artificial intelligence loads typically consume a lot of power and are difficult to operate with limited power in the car. By optimizing partitioned CPUs and FPGAs, Intel is able to increase energy efficiency by a factor of six, and with very low power consumption—only 23 fps/watt—can perform complex artificial intelligence algorithms in real time.

By optimizing partitioned CPUs and FPGAs, Intel is able to execute complex artificial intelligence algorithms in real time with very low power consumption.

In fact, before the man-machine battle, artificial intelligence had been on the Internet and the world of chess experts, played 60 games, no defeat. In the process, the masters of the chess world are amazed, and I don’t know where the masked heroes suddenly emerged. The Gu Li Jiu Duan even hanged 100,000 bounty, asking the omniscient netizens to reveal the answer.

Later results, everyone knows. This is the master of the ten-party and invincible, this is the artificial intelligence is doing the test.

Testing is a critical step in getting an unmanned vehicle on the road.

There will be 40 BMW 7 Series driverless cars, integrating the advantages of the BMW Group, Intel, Mobileye and Delphi. These vehicles bring together Intel® high-performance computing elements, Mobileye's automotive-grade computer vision technology, Delphi's expertise in driverless development, and BMW's industry-leading research into highly unmanned features. These jointly developed model cars are paving the way for the first highly automated production car, the BMW iNext*, which will be launched in 2021.

The first unmanned driving seminar on May 3, the first BMW 7 Series based on tripartite collaboration

The second season of the man-machine war is still the developer Professor Huang Shijie. Professor Huang worked diligently in the first man-machine battle and sat in front of the board for six hours. Can not help but sigh, as a representative of artificial intelligence, do you also need to go to the toilet?

In fact, Professor Huang is the so-called human-machine interface.

All artificial intelligence needs a human-machine interface that can communicate freely with people. Otherwise, is it that humans will not speak human words from now on, instead of mouth spit "10, 110, 1001"?

However, for unmanned driving, the human-machine interface needs to be more reliable.

Before passengers truly trust an unmanned car, they must have confidence in the basic way they interact with the vehicle. Intel is researching and developing prototypes for different human machine interfaces (HMIs) and technologies to help OEMs and Tier 1 suppliers address these trust issues. How to build trust between passengers and driverless cars, Intel's research looks at ways to harness and manage data, and how HMI can help passengers feel safe, secure, and in control.

Dai Weisen, senior vice president and general manager of Intel Corporation

The man-machine war is filled with smoke, in its heat. The unmanned smoke is filled with turmoil in its future. However, Dai Weisen, senior vice president and general manager of Intel's unmanned driving division, is confident. He made a bet on his career reputation and concluded that Intel will be successful in the driverless field. He said: "It will be my greatest achievement to let the driverless car go on the road."