Tag Archives: chinese
Last week, Eric Schmidt, chairman of Alphabet, predicted that China will rapidly overtake the US in artificial intelligence…in as little as five years.
Last month, China announced plans to open a $10 billion quantum computing research center in 2020.
Bottom line, China is aggressively investing in exponential technologies, pursuing a bold goal of becoming the global AI superpower by 2030.
Based on what I’ve observed from China’s entrepreneurial scene, I believe they have a real shot of hitting that goal.
As I described in a previous tech blog, I recently traveled to China with a group of my Abundance 360 members, where I was hosted by my friend Kai-Fu Lee, the founder, chairman, and CEO of Sinovation Ventures.
On one of our first nights, Kai-Fu invited us to a special dinner at Da Dong Roast, which specializes in Peking duck, where we shared an 18-course meal.
The meal was amazing, and Kai-Fu’s dinner conversation provided us priceless insights on Chinese entrepreneurs.
Three topics opened my eyes. Here’s the wisdom I’d like to share with you.
1. The Entrepreneurial Culture in China
Chinese entrepreneurship has exploded onto the scene and changed significantly over the past 10 years.
In my opinion, one significant way that Chinese entrepreneurs vary from their American counterparts is in work ethic. The mantra I found in the startups I visited in Beijing and Shanghai was “9-9-6”—meaning the employees only needed to work from 9 am to 9 pm, 6 days a week.
Another concept Kai-Fu shared over dinner was the almost ‘dictatorial’ leadership of the founder/CEO. In China, it’s not uncommon for the Founder/CEO to own the majority of the company, or at least 30–40 percent. It’s also the case that what the CEO says is gospel. Period, no debate. There is no minority or dissenting opinion. When the CEO says “march,” the company asks, “which way?”
When Kai-Fu started Sinovation (his $1 billion+ venture fund), there were few active angel investors. Today, China has a rich ecosystem of angel, venture capital, and government-funded innovation parks.
As venture capital in China has evolved, so too has the mindset of the entrepreneur.
Kai -Fu recalled an early investment he made in which, after an unfortunate streak, the entrepreneur came to him, almost in tears, apologizing for losing his money and promising he would earn it back for him in another way. Kai-Fu comforted the entrepreneur and said there was no such need.
Only a few years later, the situation was vastly different. An entrepreneur who was going through a similar unfortunate streak came to Kai Fu and told him he only had $2 million left of his initial $12 million investment. He informed him he saw no value in returning the money and instead was going to take the last $2 million and use it as a final push to see if the company could succeed. He then promised Kai-Fu if he failed, he would remember what Kai-Fu did for him and, as such, possibly give Sinovation an opportunity to invest in him with his next company.
2. Chinese Companies Are No Longer Just ‘Copycats’
During dinner, Kai-Fu lamented that 10 years ago, it would be fair to call Chinese companies copycats of American companies. Five years ago, the claim would be controversial. Today, however, Kai-Fu is clear that claim is entirely false.
While smart Chinese startups will still look at what American companies are doing and build on trends, today it’s becoming a wise business practice for American tech giants to analyze Chinese companies. If you look at many new features of Facebook’s Messenger, it seems to very closely mirror TenCent’s WeChat.
Interestingly, tight government controls in China have actually spurred innovation. Take TV, for example, a highly regulated industry. Because of this regulation, most entertainment in China is consumed on the internet or by phone. Game shows, reality shows, and more will be entirely centered online.
Kai-Fu told us about one of his investments in a company that helps create Chinese singing sensations. They take girls in from a young age, school them, and regardless of talent, help build their presence and brand as singers. Once ready, these singers are pushed across all the available platforms, and superstars are born. The company recognizes its role in this superstar status, though, which is why it takes a 50 percent cut of all earnings.
This company is just one example of how Chinese entrepreneurs take advantage of China’s unique position, market, and culture.
3. China’s Artificial Intelligence Play
Kai-Fu wrapped up his talk with a brief introduction into the expansive AI industry in China. I previously discussed Face++, a Sinovation investment, which is creating radically efficient facial recognition technology. Face++ is light years ahead of anyone else globally at recognition in live videos. However, Face++ is just one of the incredible advances in AI coming out of China.
Baidu, one of China’s most valuable tech companies, started out as just a search company. However, they now run one of the country’s leading self-driving car programs.
Baidu’s goal is to create a software suite atop existing hardware that will control all self-driving aspects of a vehicle but also be able to provide additional services such as HD mapping and more.
Another interesting application came from another of Sinovation’s investments, Smart Finance Group (SFG). Given most payments are mobile (through WeChat or Alipay), only ~20 percent of the population in China have a credit history. This makes it very difficult for individuals in China to acquire a loan.
SFG’s mobile application takes in user data (as much as the user allows) and, based on the information provided, uses an AI agent to create a financial profile with the power to offer an instant loan. This loan can be deposited directly into their WeChat or Alipay account and is typically approved in minutes. Unlike American loan companies, they avoid default and long-term debt by only providing a one-month loan with 10% interest. Borrow $200, and you pay back $220 by the following month.
Artificial intelligence is exploding in China, and Kai-Fu believes it will touch every single industry.
The only constant is change, and the rate of change is constantly increasing.
In the next 10 years, we’ll see tremendous changes on the geopolitical front and the global entrepreneurial scene caused by technological empowerment.
China is an entrepreneurial hotbed that cannot be ignored. I’m monitoring it closely. Are you?
Image Credit: anekoho / Shutterstock.com Continue reading
Chinese startup Unitree Robotics unveils Laikago quadruped robot Continue reading
Con artistry is one of the world’s oldest and most innovative professions, and it may soon have a new target. Research suggests artificial intelligence may be uniquely susceptible to tricksters, and as its influence in the modern world grows, attacks against it are likely to become more common.
The root of the problem lies in the fact that artificial intelligence algorithms learn about the world in very different ways than people do, and so slight tweaks to the data fed into these algorithms can throw them off completely while remaining imperceptible to humans.
Much of the research into this area has been conducted on image recognition systems, in particular those relying on deep learning neural networks. These systems are trained by showing them thousands of examples of images of a particular object until they can extract common features that allow them to accurately spot the object in new images.
But the features they extract are not necessarily the same high-level features a human would be looking for, like the word STOP on a sign or a tail on a dog. These systems analyze images at the individual pixel level to detect patterns shared between examples. These patterns can be obscure combinations of pixel values, in small pockets or spread across the image, that would be impossible to discern for a human, but highly accurate at predicting a particular object.
“An attacker can trick the object recognition algorithm into seeing something that isn’t there, without these alterations being obvious to a human.”
What this means is that by identifying these patterns and overlaying them over a different image, an attacker can trick the object recognition algorithm into seeing something that isn’t there, without these alterations being obvious to a human. This kind of manipulation is known as an “adversarial attack.”
Early attempts to trick image recognition systems this way required access to the algorithm’s inner workings to decipher these patterns. But in 2016 researchers demonstrated a “black box” attack that enabled them to trick such a system without knowing its inner workings.
By feeding the system doctored images and seeing how it classified them, they were able to work out what it was focusing on and therefore generate images they knew would fool it. Importantly, the doctored images were not obviously different to human eyes.
These approaches were tested by feeding doctored image data directly into the algorithm, but more recently, similar approaches have been applied in the real world. Last year it was shown that printouts of doctored images that were then photographed on a smartphone successfully tricked an image classification system.
Another group showed that wearing specially designed, psychedelically-colored spectacles could trick a facial recognition system into thinking people were celebrities. In August scientists showed that adding stickers to stop signs in particular configurations could cause a neural net designed to spot them to misclassify the signs.
These last two examples highlight some of the potential nefarious applications for this technology. Getting a self-driving car to miss a stop sign could cause an accident, either for insurance fraud or to do someone harm. If facial recognition becomes increasingly popular for biometric security applications, being able to pose as someone else could be very useful to a con artist.
Unsurprisingly, there are already efforts to counteract the threat of adversarial attacks. In particular, it has been shown that deep neural networks can be trained to detect adversarial images. One study from the Bosch Center for AI demonstrated such a detector, an adversarial attack that fools the detector, and a training regime for the detector that nullifies the attack, hinting at the kind of arms race we are likely to see in the future.
While image recognition systems provide an easy-to-visualize demonstration, they’re not the only machine learning systems at risk. The techniques used to perturb pixel data can be applied to other kinds of data too.
“Bypassing cybersecurity defenses is one of the more worrying and probable near-term applications for this approach.”
Chinese researchers showed that adding specific words to a sentence or misspelling a word can completely throw off machine learning systems designed to analyze what a passage of text is about. Another group demonstrated that garbled sounds played over speakers could make a smartphone running the Google Now voice command system visit a particular web address, which could be used to download malware.
This last example points toward one of the more worrying and probable near-term applications for this approach: bypassing cybersecurity defenses. The industry is increasingly using machine learning and data analytics to identify malware and detect intrusions, but these systems are also highly susceptible to trickery.
At this summer’s DEF CON hacking convention, a security firm demonstrated they could bypass anti-malware AI using a similar approach to the earlier black box attack on the image classifier, but super-powered with an AI of their own.
Their system fed malicious code to the antivirus software and then noted the score it was given. It then used genetic algorithms to iteratively tweak the code until it was able to bypass the defenses while maintaining its function.
All the approaches noted so far are focused on tricking pre-trained machine learning systems, but another approach of major concern to the cybersecurity industry is that of “data poisoning.” This is the idea that introducing false data into a machine learning system’s training set will cause it to start misclassifying things.
This could be particularly challenging for things like anti-malware systems that are constantly being updated to take into account new viruses. A related approach bombards systems with data designed to generate false positives so the defenders recalibrate their systems in a way that then allows the attackers to sneak in.
How likely it is that these approaches will be used in the wild will depend on the potential reward and the sophistication of the attackers. Most of the techniques described above require high levels of domain expertise, but it’s becoming ever easier to access training materials and tools for machine learning.
Simpler versions of machine learning have been at the heart of email spam filters for years, and spammers have developed a host of innovative workarounds to circumvent them. As machine learning and AI increasingly embed themselves in our lives, the rewards for learning how to trick them will likely outweigh the costs.
Image Credit: Nejron Photo / Shutterstock.com Continue reading
In the suburbs of China's capital, a 32-year-old engineer creates the kind of larger-than-life, shapeshifting robots that most have only seen in "Transformers" movies. Continue reading
Today by far the most commonly used robotics software is ROS, which stands for Robot Operating System. This is an open source software, and the most number of developers and robotics users are involved with this program with an ever increasing rate. It contains set of libraries, algorithms, developer tools and drivers for developing robotics projects. The first release of ROS was in 2010, and as of end of 2016, ROS has reached its 10th official release, which is called “ROS Kinetic Kame”. There are translations to 11 languages other than English, which are: German, Spanish, French, Italian, Japanese, Turkish, Korean, Portuguese, Russian, Thai and Chinese. It currently has 2000+ software libraries, which keeps increasing every year.
Many robots use ROS now, including but not limited to hobby robots, drones, educational or advanced humanoid robots, domestic robots including cleaning robot vacuums, cooking robots or telepresence robots and more, robot arms, farming robots, industrial robots, even Robonaut of NASA in space or the four legged military robots in development. A list of robots which use ROS can be found here: http://wiki.ros.org/Robots.
We were checking the Alexa.Com ranking of ROS since few years, in order to track the increase in usage, and we believe it is time to share it now, as we have enough data. The numbers on the left are dates we looked and the numbers on the right indicate the ranking of Ros.Org website from top, among all websites in the world:
May 2011: 189,000 th in the world, from top, among all other websites
April 2012: 187,900 th
January 2014: 107,821
May 2014: 112,236
September 2014: 83,875 (7219 in Canada, the country where it is most accessed)
January 2015: 83,556 (4,258 in Canada)
February 2015 : 75,680 (33185 in USA)
April 2015: 59,200 (31,334 in USA)
August 2015: 65,754 (50,132 in USA)
September 2016: 30,201 (China 5073)
This chart shows the increasing rank of ros.org among other websites in the world, which is a good indicator of its growth. The numbers on the left represent the site’s ranking from the top, among all other sites in the world. Chart Copyright: Robokingdom LLC.
As can be seen here, in May 2011, when we first checked this ranking, ROS.org was at 189,000 th place in the world from the top among all other websites in terms of unique visitors that visit the site, and it almost continuously increased its ranking. As of September 2016, it is now the 30,201st most reached website in the world, with mostly being accessed in China (5073 from top in China). Let’s not forget that even if it’s position remained the same, let alone going up, it would still mean the traffic of the site was going up, as every year there are more websites in the world which means the same ranking means better place and more traffic. The ranking of 30,201 means ROS.org is a very high traffic website in the world right now, being accessed probably by at least hundreds of thousands of people every day, with no indication of slowing down its rise yet.
The most important result of all of this, is that the use of robots is increasing, both in terms of number and type (when you look at the type of robots that use ros, as it also increases in variety all the time).
From Alexa, we were also able to see, from publicly available information, that the percentage of reach among countries for ROS.org is as follows:
South Korea 3.5%
This also shows us that in China, a lot of things are going on for robotics development right now, as it gets most of its traffic from there with 47.5%. USA then follows with 11.5% and Japan is third with 8.7%.
With ROS, any type of sensors can be controlled, including 1d/2d range sensors, 3d range finders and cameras, audio/speech recognition sensors, cameras, environmental sensors, force/torque/touch sensors, motion capture, pose estimation, power supply, RFID, and sensor interfaces.
In ros.org site, in addition to all packages, there are also extensive tutorials and a discussion board that one can ask questions and share knowledge.
ROS also has an industrial section, the version of software modified for industrial applications. It is called ROS industrial, and can be reached at: http://rosindustrial.org/. Although we see domestic robots with new abilities or advanced research projects that aim to develop capabilities of robotics every year, according to the results of a study that is shown on http://rosindustrial.org/the-challenge/ website, the abilities of industrial robots are not progressing and the abilities are restricted to welding, material handling, dispensing, coating (although we know that they do additional tasks such as packaging, inspection, labeling etc…). ROS Industrial aims to solve this challenge by providing a common skeleton to all developers, with its extensive and stronger software architecture, than other individual robotics programs.
The post Robot Operating System (ROS) continues its growth appeared first on Roboticmagazine. Continue reading