google training ad placement computers to be offended
by：Yufengling
2019-09-24
Mountain View, California. —
Google has been training computer systems for years to exclude copyrighted content and pornographic content from the YouTube service.
But after seeing the Coca-Cola ad
Coca-Cola, P & G and Walmart
Mart appears in racism,-
Their engineers realized that there was a blind spot in their computer model: they didn\'t understand the background.
Now, teaching computers to understand what humans are easy to master may be to calm down the big-
Their ads come along with videos from extremist organizations and other offensive information.
Google\'s engineers, product managers and policy experts are all trying to train computers to grasp the objectionable nuances of certain videos.
Advertisers may tolerate the use of race names on hips
Hop Video, for example, but it can be scary to see it used in videos of racist bald-headed groups.
The advertisement that Well bought
It has long been known that companies occasionally appear next to offensive videos, which is considered a major trouble for the YouTube business.
But the issue has become more urgent in recent weeks, as the London Times and other media have reported brands that inadvertently fund extremists through automated advertising --
A by-product of a system in which YouTube shares part of the ad sales with the creator of the content the ad is targeting.
This failure in the company\'s huge automation process has turned into an open-
Relationship nightmare.
Companies like AT&T and Johnson & Johnson say they will withdraw their ads from YouTube and Google\'s display advertising business until they can be assured that such placements will not happen again.
Consumers watch more than a billion hours a day on YouTube, making it the dominant video platform on the Internet, and it is clearly a beneficiary as ad revenue shifts from television to the Internet.
But recent issues have criticized Google for not doing enough to focus on advertisers.
This is a big problem for billions of people.
The dollar company that still earns most of its revenue through advertising.
In an interview last week, Google\'s chief business officer Philip Schindler said: \"We take this issue as seriously as we do with it . \".
\"We have been in a state of emergency.
\"In the past two weeks, Google has changed the type of video that can play ads, banning ads with hate speech or discriminatory content.
In addition, Google is simplifying how advertisers can rule out specific websites, channels and videos on YouTube and Google\'s display network.
It allows the brand to fine-
Adjust the type of content they want to avoid, such as \"sexual cues\" or \"sensational/weird\" videos.
It is also setting stricter security standards by default, so advertisers have to choose to put ads next to more provocative content.
When ads appear next to offensive content, Google creates a quick reminder.
The Silicon Valley giant is trying to reassure the world\'s second-largest company like Unilever.
The largest advertisers have consumer brands such as Dove and Jerry.
As other brands began to flee YouTube, Unilever found examples of three brands appearing on objectionable YouTube channels.
But Keith Weed, Unilever\'s chief marketing officer, decided not to withdraw its ads because there was a small proportion of ads that showed objectionable content.
According to video analytics firm OpenSlate, an average of $100,000 of YouTube activity is on more than 7,000 channels, and Unilever spends hundreds of millions of dollars on YouTube.
In addition, Google found that these ads appeared because of human error when setting the security level. Mr.
Weed said it is in Unilever\'s best interest to win concessions from Google rather than cut ties.
As part of the new measures, Google agreed to work with external companies to provide a third
Party verification about where ads appear on YouTube.
When he came up with the idea of independent verification,
Weed said Google executives acted as if he didn\'t think Google was trustworthy.
The problem, he says, is not trust, but the ability of companies to \"mark their own jobs \".
He said he thought Google would eventually agree, but \"the current situation has accelerated their plans . \"
Google\'s efforts are being watched.
Johnson & Johnson, for example, said it has resumed YouTube advertising in some countries.
Google says other companies are returning.
To a large extent, Google has not adequately addressed this issue before, as it does not have;
Examples of advertising appearing next to objectionable content rarely occur, as well as from a broader public perspective.
Google says bad video accounts for less than 1 of the top advertisers-
A few percent of their total advertising impressions.
Google is using machines to train computers.
Learning Technology
Many of its biggest breakthroughs in potential technologies such as self-driving car.
It also brings a huge team of people (
It refused to say how big it was)
Review the suitability of a computer-marked suspicious video.
Basically, they are training a computer to identify a shot of a woman wearing a custom sports bra and yoga leggings doing yoga poses in a sports video that is safe for advertising, not sexual cues.
Similarly, they will mark a video of a Hollywood action star waving a gun as acceptable to some advertisers, while marking similar pictures involving Islamic State gunmen as inappropriate.
Google used to use similar methods to create an automatic rating system for videos similar to movie ratings based on the suitability of content to specific audiences.
But Google is now trying to solve a different problem.
\"Computers are much more difficult in understanding the context, which is why we are actually using all of our latest and greatest machine learning abilities now to better understand this,\" Mr. Schindler said.
Armed.
Google\'s computer system decomposes the images of YouTube videos frame by frame, analyzes each image, and verifies what is safe and what is not.
They also digest what is being said, the creator\'s description of the video, and other signals to detect patterns and identify subtle clues that cause the video to be inappropriate.
The idea is to make the machine make a difficult decision in the end.
In the case that the brand believes that Google does not mark inappropriate videos, this example is fed back into the system, so it will improve over time.
Google says it has marked the number of videos that are not suitable for advertising 5 times, although it refuses to provide absolute figures on how many videos are involved.
There are over a billion videos on YouTube, 400 hours of new content per minute, 3 million ads
Channels supported on the platform
Schindler said there was no guarantee that Google would solve the problem completely.
He compared how a car company could not promise that even new tires would not fail within the first 10,000 miles.
\"No system can do 100% perfect,\" he said . \"
\"But we are doing our best to make it as safe as possible.
Dongguan Yufenglin Garment Co., Ltd. works very hard to understand your objectives, then create a program that can help you meet them.
For more tips and strategies on effective personalized shirts solutions, get your choice at Yufenglin Garment.
Your co-workers, investors and clients have busy schedules, and it can be hard to get everyone in the same place at the same time for custom t shirt design. So, it is important to create a connection between company and clients.
Google has been training computer systems for years to exclude copyrighted content and pornographic content from the YouTube service.
But after seeing the Coca-Cola ad
Coca-Cola, P & G and Walmart
Mart appears in racism,-
Their engineers realized that there was a blind spot in their computer model: they didn\'t understand the background.
Now, teaching computers to understand what humans are easy to master may be to calm down the big-
Their ads come along with videos from extremist organizations and other offensive information.
Google\'s engineers, product managers and policy experts are all trying to train computers to grasp the objectionable nuances of certain videos.
Advertisers may tolerate the use of race names on hips
Hop Video, for example, but it can be scary to see it used in videos of racist bald-headed groups.
The advertisement that Well bought
It has long been known that companies occasionally appear next to offensive videos, which is considered a major trouble for the YouTube business.
But the issue has become more urgent in recent weeks, as the London Times and other media have reported brands that inadvertently fund extremists through automated advertising --
A by-product of a system in which YouTube shares part of the ad sales with the creator of the content the ad is targeting.
This failure in the company\'s huge automation process has turned into an open-
Relationship nightmare.
Companies like AT&T and Johnson & Johnson say they will withdraw their ads from YouTube and Google\'s display advertising business until they can be assured that such placements will not happen again.
Consumers watch more than a billion hours a day on YouTube, making it the dominant video platform on the Internet, and it is clearly a beneficiary as ad revenue shifts from television to the Internet.
But recent issues have criticized Google for not doing enough to focus on advertisers.
This is a big problem for billions of people.
The dollar company that still earns most of its revenue through advertising.
In an interview last week, Google\'s chief business officer Philip Schindler said: \"We take this issue as seriously as we do with it . \".
\"We have been in a state of emergency.
\"In the past two weeks, Google has changed the type of video that can play ads, banning ads with hate speech or discriminatory content.
In addition, Google is simplifying how advertisers can rule out specific websites, channels and videos on YouTube and Google\'s display network.
It allows the brand to fine-
Adjust the type of content they want to avoid, such as \"sexual cues\" or \"sensational/weird\" videos.
It is also setting stricter security standards by default, so advertisers have to choose to put ads next to more provocative content.
When ads appear next to offensive content, Google creates a quick reminder.
The Silicon Valley giant is trying to reassure the world\'s second-largest company like Unilever.
The largest advertisers have consumer brands such as Dove and Jerry.
As other brands began to flee YouTube, Unilever found examples of three brands appearing on objectionable YouTube channels.
But Keith Weed, Unilever\'s chief marketing officer, decided not to withdraw its ads because there was a small proportion of ads that showed objectionable content.
According to video analytics firm OpenSlate, an average of $100,000 of YouTube activity is on more than 7,000 channels, and Unilever spends hundreds of millions of dollars on YouTube.
In addition, Google found that these ads appeared because of human error when setting the security level. Mr.
Weed said it is in Unilever\'s best interest to win concessions from Google rather than cut ties.
As part of the new measures, Google agreed to work with external companies to provide a third
Party verification about where ads appear on YouTube.
When he came up with the idea of independent verification,
Weed said Google executives acted as if he didn\'t think Google was trustworthy.
The problem, he says, is not trust, but the ability of companies to \"mark their own jobs \".
He said he thought Google would eventually agree, but \"the current situation has accelerated their plans . \"
Google\'s efforts are being watched.
Johnson & Johnson, for example, said it has resumed YouTube advertising in some countries.
Google says other companies are returning.
To a large extent, Google has not adequately addressed this issue before, as it does not have;
Examples of advertising appearing next to objectionable content rarely occur, as well as from a broader public perspective.
Google says bad video accounts for less than 1 of the top advertisers-
A few percent of their total advertising impressions.
Google is using machines to train computers.
Learning Technology
Many of its biggest breakthroughs in potential technologies such as self-driving car.
It also brings a huge team of people (
It refused to say how big it was)
Review the suitability of a computer-marked suspicious video.
Basically, they are training a computer to identify a shot of a woman wearing a custom sports bra and yoga leggings doing yoga poses in a sports video that is safe for advertising, not sexual cues.
Similarly, they will mark a video of a Hollywood action star waving a gun as acceptable to some advertisers, while marking similar pictures involving Islamic State gunmen as inappropriate.
Google used to use similar methods to create an automatic rating system for videos similar to movie ratings based on the suitability of content to specific audiences.
But Google is now trying to solve a different problem.
\"Computers are much more difficult in understanding the context, which is why we are actually using all of our latest and greatest machine learning abilities now to better understand this,\" Mr. Schindler said.
Armed.
Google\'s computer system decomposes the images of YouTube videos frame by frame, analyzes each image, and verifies what is safe and what is not.
They also digest what is being said, the creator\'s description of the video, and other signals to detect patterns and identify subtle clues that cause the video to be inappropriate.
The idea is to make the machine make a difficult decision in the end.
In the case that the brand believes that Google does not mark inappropriate videos, this example is fed back into the system, so it will improve over time.
Google says it has marked the number of videos that are not suitable for advertising 5 times, although it refuses to provide absolute figures on how many videos are involved.
There are over a billion videos on YouTube, 400 hours of new content per minute, 3 million ads
Channels supported on the platform
Schindler said there was no guarantee that Google would solve the problem completely.
He compared how a car company could not promise that even new tires would not fail within the first 10,000 miles.
\"No system can do 100% perfect,\" he said . \"
\"But we are doing our best to make it as safe as possible.
Dongguan Yufenglin Garment Co., Ltd. works very hard to understand your objectives, then create a program that can help you meet them.
For more tips and strategies on effective personalized shirts solutions, get your choice at Yufenglin Garment.
Your co-workers, investors and clients have busy schedules, and it can be hard to get everyone in the same place at the same time for custom t shirt design. So, it is important to create a connection between company and clients.
