Specifically, Hashimoto Kanna participated in a Japanese program called 1-Oku 3000-Mannin no SHOW Channel . When participating in this program, the guests will be challenged in a field that has never been done before. It could be said that the guests – including Hashimoto Kanna had to take on the challenges without any prior practice. And the challenge for Hashimoto Kanna in the program is to pull the tissue out of the box within 1 minute. Hashimoto Kanna’s challenge is to pull a tissue out of a box for 1 minute. Draws 157 tissues. Hashimoto Kanna was also shocked when he broke the previous Guinness record of 140. Challenge to play and get a real Guinness record! As a rule, Hashimoto Kanna puts his left hand behind his back and only uses his right hand to pull out a tissue. She challenged twice. For the first time, Hashimoto Kanna drew 126 pieces. But the second time, she quickly pulled out 157 tissues. Meanwhile, the Guinness record in this field is 140 pieces, so Hashimoto Kanna automatically becomes the new Guinness world record holder for taking a tissue out of a box. Hashimoto Kanna is currently the holder of the Guinness World Record for pulling out a tissue. Hashimoto Kanna is a famous female idol in Japan, fondly called “a once-in-a-thousand-year idol”. Hashimoto Kanna has been involved in showbiz since childhood. At the age of 14, her pure and sweet beauty in a performance became famous throughout Japan; Since then, Hashimoto Kanna has been advertising, acting… Hashimoto Kanna is fondly called the “One Thousand Year Idol” of Japan. Super muddy acting in “Couple of Cards”. And extremely funny, cute in “Gintama”. As she matures, Hashimoto Kanna becomes more beautiful and talented. She participated in many famous movies such as Worm Card Pair Cặp g ( Kyou kara Ore wa !), Gintama , Kingdom … Not only is she beautiful, she is also impressed by her ability to perform muddy comedy but is still very lovely.

Last month, Microsoft released Windows 10 21H1 feature update, also known as the May 2021 Update, to users as an optional update displayed when manually performing Check for updates in Windows Update. In the latest announcement, Microsoft said that they have started rolling out the feature to automatically upgrade Windows 10 2004 devices to Windows 10 21H1 to train their machine learning algorithms. “We have also begun the first phase of rolling out machine learning (ML) training targeting devices running Windows 10 version 2004,” Microsoft said on the Windows 10 21H1 announcements and issues page. for automatic update to Windows 10 version 21H1. We will continue to train our machine learning through all stages to intelligently deploy new versions of Windows 10 and deliver a smooth update experience.” Microsoft announced that they are adjusting the availability for this update in the coming weeks and that it may not be available to all users. For those who want to install Windows 10 21H1 right away, they can click Check for updates in Windows Update, then click the Download and install option when displayed. The new feature update is also available to deploy using the Media Creation Tool or through WSUS.

It turns out that behind it is a long story, to know the story we have to go back many years ago, when the “snooze alarm” button was born, watchmakers were in trouble when they cannot be fine-tuned so that the gears match to delay exactly 10 minutes. And then, watchmakers had 2 options of 9 minutes or 11 minutes, to this day many people are still talking about their choice. However, many people think that 9 minutes is a reasonable amount of time because when going to 10 minutes, people have already entered a deep sleep and make this second awakening make you more sleepy, or simply is that waking up 1-2 minutes earlier is better than “trying to sleep” more. There is also a lot of controversy that Apple’s choice of “Snooze” time of 9 minutes is to honor the first watch model with this “alarm delay” feature. This electric watch model is called 7H241, of Telechron watch company – USA, integrated with an “alarm delay” button with the default time of 9 minutes.

Compilation | Xiong Dabao

Editor | Li Shuiqing

Zhixi News reported on April 14 that at the beginning of this month, researchers at the University of Waterloo in Canada updated the first open source high-resolution wearable camera image database for human motion scenes. On this basis, AI and wearable cameras can be used to allow exoskeleton robots to walk autonomously.

An exoskeleton robot refers to a robot that is sheathed outside the human body, also known as a “wearable robot”. The robot has the functions of sensing, control, information, fusion, mobile computing, etc. The operator can tie it to the leg to help them enhance or restore body functions.

▲ Researchers from the University of Waterloo use the team’s exoskeleton robot to complete the walking action

1. The use of exoskeleton robots: from hand control to machine autonomy

At the beginning of the use of exoskeleton robots, it is usually necessary to rely on the manual control of the operator to switch motion modes.

Brokoslaw Laschowski is one of the main developers of this research at the University of Waterloo. He said: “Every time you want to change the way you move, you need to rely on a joystick or a smartphone app. Program to operate the exoskeleton robot, from sitting to standing, standing to walking, walking on the ground to going up and down stairs. “This is a great burden for the operator.

How to make the exoskeleton robot automatically recognize when to switch the motion mode?

▲California Institute of Technology researchers use exoskeleton robots for walking demonstrations

Scientists have attached sensors to the legs, this method can detect the bioelectric signals sent from the brain to the muscles, telling them to move.

However, this method also has some problems, such as skin conductivity will be affected when the skin sweats.

2. The accuracy rate is over 70%, and the exoskeleton robot automatically recognizes the walking environment

The Laszowski team is trying a new method: install wearable cameras for users and provide visual data to the machine so that it can operate autonomously. The AI ​​software equipped with the robot can analyze this data to identify stairs, doors and other features of the surrounding environment and calculate how to make the best response.

For exoskeleton robots to realize automatic recognition in any walking environment, a large amount of data is required.

Laszowski led the ExoNet project, the first open source database of high-resolution wearable camera images of human motion scenes. The database has more than 5.6 million images of real walking environments indoors and outdoors.

▲ExoNet project introduction interface

Laszowski pointed out that the team used the data to train deep learning algorithms. Although there are huge differences in the different surfaces and objects perceived by the wearable camera, their team’s deep convolutional neural network has been able to automatically recognize different walking environments with 73% accuracy.

At the same time, Laszowski also indicated that AI’s deep learning of images will also make it dependent on traditional two-dimensional images. When outdoor lighting and distance increase, the accuracy of AI measurement will usually decrease.

Other researchers are also using AI and wearable cameras to allow exoskeleton robots to walk autonomously.

Researchers in North Carolina installed wearable cameras on their eyes or knees, allowing volunteers to walk through various indoor and outdoor environments to capture image data that exoskeleton robots might use to observe the world around them.

Edgar Lobarton, an electrical engineering researcher at North Carolina State University, said that they focus on how AI software can reduce the uncertainty caused by factors such as motion blur or overexposure of images to ensure safe operation. “We want to ensure that before AI is integrated into hardware products, we can truly rely on it.”

3. Autopilot is inspired to improve the stability of exoskeleton robots

In the future, Laszowski and his colleagues will focus on improving the accuracy of environmental analysis software to help exoskeleton robots perform real-time operations better.

Lobarton and his team also explored how to deal with the uncertainties that motion brings to the visual system.

ExoNet researchers hope to explore how AI software transmits commands to exoskeleton robots so that the robot can analyze the terrain around the user according to the system, and perform tasks such as climbing stairs or avoiding obstacles. Laszowski said that, inspired by self-driving cars, they are seeking to develop autonomous exoskeleton robots that can complete walking tasks without human input.

At the same time, Laszowski’s team is also considering the safety of the exoskeleton robots operated by the elderly and disabled. “User safety is very important. If the classification algorithm or controller of the skeletal robot makes a wrong decision, the user must always have the ability to control it.”

Conclusion: Autonomy is the future trend of bio-assisted robots

As a branch of the robotics field, exoskeleton robots have been studied since 1960, but how to use sensors to accurately obtain the intention of the human body has always been a problem to be solved.

With the support of AI and wearable cameras, the development and landing of exoskeleton robots, bio-assisted robots, will accelerate, expanding its commercial and civilian markets.

Source: IEEE Spectrum