COVID-19 Response COVID-19 Response: Source manufacturers and distributors that provide COVID-19 medical supplies ico-arrow-default-right
Welcome to Thomas Insights-we release the latest news and analysis every day to keep readers informed of the latest industry trends. Register here to send the hot news of the day directly to your inbox.
The routine is now familiar. From the beginning of the day, lying on the sofa and wandering around, ready for ice cream, remote control or mouse click on the preferred streaming site, maybe watching a show that everyone is talking about or re-acquainting with an old lover.
Nowadays, streaming TV and movie services have been deeply rooted in daily affairs. Crazy watching is a common activity on weekends, and colloquial dating terms call them “Netflix and chill”. People log in to Netflix, Hulu and other streaming services all day long: when preparing in the morning; time-lapse in the evening
Despite the large number of users and the continuous use of streaming media services, such sites continue to build a completely unique experience for each user, mainly through the use of artificial intelligence (AI) to predict the complex mathematical equations of user behavior, that is, algorithms.
Streaming media giant Netflix has managed to weave algorithms to predict almost all interactions on the site. In fact, it can control almost all the content that users see when they log in to watch the program, including (but not limited to): page structure, genre lines, popular videos that users may see, the order in which they watch videos, and the content displayed image.
According to a 2013 Wired magazine article, although the team does have human support-a team of 40 freelancers who have tagged content and gestured for more than 800 engineers-but in terms of suggestions, The workload of the algorithm is huge.
Algorithms are essential to the site, and Netflix reportedly even collected data on how users click and browse to better serve its members. Xavier Amatriain told Science and Technology Publications: “We know the games you’ve played, what you’ve searched for, or the ratings you got, as well as the time, date and device. We even track user interactions, such as browsing or scrolling behavior. All this data is fed Among several algorithms, each algorithm is optimized for different purposes.”
According to Netflix research, the Computer Research Association was written by Carlos Uribe-Gomez and Neil Hunt in 2015. The streaming site estimates that if the average user can’t find a valuable streaming title within 60 to 90 seconds, they will Will lose interest. To reduce this loss and increase engagement, the algorithm tracks several markers to best customize the viewer’s experience.
“Our data shows that depending on the day of the week, the time of the day, the device and sometimes even the location, the viewing behavior will vary,” Amatriain told Wired magazine. The algorithm will work accordingly to best foresee and provide to viewers, thus solving the daily or even hourly experience that Netflix gives users.
This efficiency helps Netflix attract consumers for a longer period of time and maintain its consumer base. According to VentureBeat, this compiled algorithm system can help companies “make more than $1 billion a year in terms of retention alone.” The system is described as the “core” of streaming services because it helps alleviate “abandoning our services in exchange for other entertainment options.”
Netflix uses a variety of algorithms to control and personalize the user experience. On any given Netflix homepage, there are about 40 genre rows, of which there are about 75 titles. All these rows are selected by the algorithm and summarized by the page generation algorithm.
The category line is determined by the personal video classifier (PVR). In the type row, the “Top N” ranking uses short-term trends to find the best matches for users. The similarity between the videos informs viewers of options that may be of interest to them based on previously viewed titles.
“By looking at the metadata, you can find various similarities between the screenings,” Gomez-Uribe explained to WIRED, pointing out the similarities in ratings and time periods to help determine the recommended content. “You can also view user behavior-browse, play, search.”
Based on statistical data or even images that the viewer may be most interested in, the evidence selection algorithm specifies what the viewer can see. Gomez Uribe told Wired magazine: “Placement is important. The closer the title is to the first position in the line, the more likely it is to play. The higher the page on a line, the more It may produce playback.” Even the internal search engine has been algorithmically optimized and can generate 20% of viewers.
Compared with streaming media giant Netflix, Hulu’s system relies less on the Al algorithm, and more on the connections between people. At the end of 2019, they launched a more enhanced recommendation system that records the content and time users watched and a more complete search engine. But when it comes to video trends that may interest users, Hulu still relies on connections between people.
“We know that today’s connected consumers want a profoundly personalized experience when watching TV. At Hulu, we have always adopted a unique recommendation method. Product Management Director Jason Wong said: “This is human curation, empowering and empowering viewers. The perfect combination of algorithms, these algorithms complete the personalized experience we provide to Hulu subscribers. ”
Although the algorithm helps the site, the impact of human-computer interaction on Hulu is far greater than its rival Netflix. The streaming site said in a statement to The Verge: “At Hulu, we believe that the best search and discovery experience is built on three key pillars-our editors can find and highlight relevant content in time, our The recommendation algorithm can understand the audience’s preferences and features that allow us to listen to the audience and give them more control.”
However, even if all the data is collected and organized, artificial intelligence is only human. Many users do not rely on the programs they suggest at all, but only play content recommended by friends and family.
Sometimes users even think that the algorithm will make life harder, just like users who share an account in a relationship and have to deal with sharing suggestions after a breakup.
Jesus Diaz wrote the following article for Fast Company: “I don’t care how effective the company’s algorithm is-according to my personal experience, it doesn’t work. A machine always Can’t completely replace personal taste and exploration based on human interaction.” He continued in the column: “We hate algorithms telling us where to go, who to listen to, and what to watch. Your machine expects that my chance of watching Frozen is 98%. I will never actually, I have not found any examples of algorithms, but there is a clever suggestion that surprised me.”
Even if Gomez-Uribe accepts WIRED, the algorithm proposal has certain limitations. “I watched the French thriller “Tell No One” more than a year ago. I have been looking for similar movies. The winner of the content team said this is the only one in the world.
Although nothing can replace the buzz of TV shows that everyone is talking about, algorithms will undoubtedly affect what people watch when streaming.
Ordinary viewers may not even think of how their favorite streaming media sites work, but algorithm-driven artificial intelligence is constantly assisting their host sites to master the exquisite swan-like dance: running smoothly on the surface while intensely paddling to control and create And maintain the unique adventure experience of each consumer. When someone logs on to and from get off work or shrinks on the couch with Gilmore Girls couch, they are using their personal artificial intelligence to relax.
We are using the power of the platform to solve the massive shortage of key supplies. If your company can help provide consumables, functions or materials for N-95 masks and Tyvek suits, please contact us.
Copyright © 2020 Thomas Publishing Company. all rights reserved. Please refer to the terms and conditions, privacy statement and California non-tracking notice. The website was last modified on October 13, 2020. ThomasRegister® and ThomasRegional® are part of Thomasnet.com. Thomasnet is a registered trademark of Thomas Publishing Company.
Post time: Oct-13-2020
