Are you sick and tired of spending all your time at work creating dank memes? Does your boss become frustrated when you don’t have a new meme ready every hour? Well, worry no more, because Stanford University students are using their tuition money properly and have finally taught A.I. how to create dank memes! There’s no need to fear the uprising of Artificial Intelligence, because these two have gifted computers the power of comedy. Researchers, Abel L. Peirson V and E. Meltem Tolunay, sought to find a way to create quality, diverse, and original dank memes in a fraction of the time and, oh goodness, they did not disappoint!
It was no easy task feeding this computer over 400,000 memes, but this beautiful piece of technology was well worth it. You see, creating a dank meme is more than some guy on his laptop in his parents’ basement. It’s art. This form of art is celebrated worldwide as a, “unit for carrying cultural ideas, symbols, or practices.”  In a world fraught with political debates, high expectations, and low-flow toilets, memes bring people together. In the scientific paper released by the duo, they stated their most grappling task during this project was to truly encapsulate humor, something “people of all ages and cultures respond to.”  In order to successfully capture this true comedy, they tested different image capturing models to ensure correct match-up’s
between the text captions and photographs. Once the initial memes were created, the researchers asked a plethora of human volunteers to give feedback on the ‘dankocity’ (a.k.a. coolness) of the memes. Now, that is what we call totally legit science!
To generate the extensive data set of 400,000 images, captions, and caption triplets, they created 2600 pairs of image-caption combinations. After the combinations were produced, they were assigned with approximately 160 more captions as viable meme options. That’s some dank progress! If you wish to learn more about the process of creating this brilliant program, the link to their research paper, “Dank Learning”, can be found here. Within the research paper, the two go more in depth regarding the encoder and decoder method they used to create the memes, as well as, the actual algorithms they used. In addition, they include a website link to their GitHub account where they provide memes from their initial test runs, different data collected during the project, and the program that they created. All that is left is just to sit back, relax, and make some memes.
In various news outlets, debates have already sparked on what the two scientists’ original plan was. Was it intended solely for computer knowledge or was it their mission to create the most important web application of this era? ¯\_(ツ)_/¯ That’s all we can truly say.