In case we didn't have enough with the hundreds of thousands of funny-few of them really funny-circulating through the network with photoshop ready to create memes that offer them their moment of glory, now comes the artificial intelligence to try to brighten our lives.And researchers at Stanford University are teaching a machine to create memes with the intention of making it the most sympathetic cradle of all social networks. The method they use, playing with the names machine learning or deep learning , it has been named dank learning (ironic learning).Everything is a party.
RelatedStarting from the classical definition, a meme is a minimal unit of cultural information disseminated by imitation.The memetics includes both mimesis and replication by repetition.From the Greek mimema (imitated), this concept was first introduced in 1976, through the British evolutionary biologist Richard Dawkins in his work 'The Selfish Gene'.
In our days, the term internet meme is an idea, concept, situation, expression and/or thought, manifested in any type of virtual media, comic, video, audio, texts, images and all kinds of multimedia construction that it is replicated through the Internet from person to person until it reaches a wide diffusion.It is usually a humorous manifestation.They are spread in forums, websites and mainly in social networks. A possible mechanism of cultural evolution has been proposed.
from democracy to memecracy
Beyond its usefulness to give us a laugh on social networks, as we already have in Nobbot, memes are a thermometer of our time: to understand a meme is to understand the time in which we live .And if they reproduce, it is because they are useful, both as information containers and tools to memorize our existence.
Back in 2013, journalist Delia Rodriguez said that we lived in a Memecracy.Five years later, an important sporting event or the blunder of a politician is enough so that even the web pages of the more "serious" newspapers ", echo echo of the best memes that flood the network, especially Twitter.
Well, as a result of the work of Standford researchers, the creation of these viral pieces would no longer be exclusive to humans: Viral images could be generated by pressing a button and not thinking much, letting an algorithm work.The one that awaits us...
hilarity test
The algorithm developed by these friendly researchers analyzed 400,000 images, labels and texts. The rest of the database created for the occasion was composed of vocabulary, semantic rules, random initialization and vectors to refine the process of artificial intelligence.The result? U nas 160 combinations of image and label with thousands of jokes in each one.To start to shake.
Subsequently, and to ensure a greater impact of the created memes, methods are used to measure grammar correction and a perplexity factor, which assesses the level of surprise or impact of the joke.We can only drool thinking about what Chiquito de la Calzada would have done if he had something like that at his disposal.
Once the meme is ready, it is evaluated by five volunteers to determine if they are able to distinguish memes created by humans and machines, and also to measure their humorous level (hilarity test).
Why create memes?
The creator of memes is born or made? Why do you dedicate your time to this activity? "Sometimes I do them at work when I find some free moment, at some break", he confesses Nobbot Unmundolibre."I was glad that people liked what I was doing.After a depression I turned to it much more.I needed laughter to get out of where I was," acknowledges this same memes generator.
But, how are these little visual success stories made? First we will know the basic tools that every meme creator should have.To make his own, Proscojoncio uses Photoshop for montages, Inkscape for illustrations and Sony Vegas for video editing.This creator of memes is dedicated, in part, professionally to it."I have a work in shifts and I have also had to become autonomous to be able to collaborate in other places and have everything in order before the Treasury."
No one seems to have the clue which are the most successful or why.It is true, as Unmundolibre acknowledges that political memes work quite well , but "every meme is like a fad: It comes , it’s a while and it’s best to appear again after a while.It’s seldom known what people want to see at any time."
Comments
Post a Comment