Spring has finally arrived! It might not feel like it just yet, but in the next few weeks our city is going to blossom. Soon, our streets will be filled with colour, music, art and crowds of people searching for beautiful scenery that will (hopefully) inspire their social media platforms.
Montreal is a magnet for festivals of all kinds and since we are recognized as a creative hub filled with innovators from every background imaginable and also as an artificial intelligence powerhouse, it only makes sense that we would be fascinated by technology that can marry both.
Here are our favourite unconventional AI applications with an artistic flair:
AI: The Screenwriter
Two years ago, a long short-term memory recurrent neural network, self-styled as “Benjamin”, wrote a short science fiction film called Sunspring. Yep, let that sink in for a minute.
The director of the project, Oscar Sharp, along with his partner Ross Goodwin (an AI researcher at New York University), teamed up to develop an AI that could be credited as a screenwriter. The data came from screenplays from 80’s and 90’s sci-fi films and the technology they selected is mainly used for text recognition. The reason they chose LSTM is for its ability to predict which letters are likely to follow each other and ultimately, which words, sentences and even paragraphs are going to come next.
The AI was able to develop the structure of a screenplay, the dialogue, recognize patterns in storytelling and even outline stage directions in order to produce a 9 minute short. It’s important to remember that Benjamin wrote his story based on what had already been written, so if we think it’s a little weird…we only have ourselves to blame.
Check out the final result here:
AI: The Painter
Can machines be creative? The good people at Rutgers’ Art and Artificial Intelligence Laboratory seem to think so. Their team has created a computational creative system for art generation that is completely independent of humans during the creative process.
During the learning process however, the algorithm uses a variety of paintings throughout history in order to achieve an original end-result. The researchers believe it’s crucial to expose the algorithm to past paintings, not only because data is what algorithms learn from, but also because humans are constantly exposed to other artist’s throughout their lifetime. This exposure eventually allows an artist to develop a sense of personal style. The team took things a step further and implemented Generative Adversarial Networks (GANs) as a method to increase stylistic ambiguity in order to give the AI a chance to create art outside the confines of style norms.
Not only can the system create art that most critics can agree is aesthetically pleasing, it can also be given the freedom to create something we haven’t seen before.
At the end of the day, isn’t “generate” just another word for “create”? Let’s give AI credit where it’s due, these are exquisite:
AI: The Music Producer
A few months ago, Amper released its first album called “I AM AI”. The title wasn’t clever, it wasn’t a mind game and it doesn’t have some deeper meaning, because Amper is an artificially intelligent music composer, producer, and performer.
In the past, music technology was created with the purpose of collaborating with humans and artists in order to propel the creative process to a whole new level. It used to be that the AI would provide suggestions to the chords and melodies and the human producer would decide if he wanted to manually make the changes or not. Now, with the help of machine learning, Amper can create unique music in only a few seconds.
Check out the end result when a woman and machine collaborate to create a song:
AI: The Designer
Most fashion lovers know the Met Gala Ball. It’s an annual fundraising gala for the benefit of the Metropolitan Museum of Art’s Costume Institute in New York City. Every year, designers are inspired by the theme of the evening.
For the theme of “Fashion in an Age of Technology”, Marchesa teamed up with IBM Watson to create a one of a kind dress that was powered by artificial intelligence. The goal was to allow the gown to reflect the real-time emotions of social media users during the event.
In order to do that, Watson turned emotions into information, and created a color palette for the LED lights from that analysis. Rose signified joy, coral meant passion, aqua was excitement, lavender denoted curiosity and butter indicated encouragement.
The end result is the most high-tech dress ever created:
In the End…
Artificial intelligence is useful for more than one purpose. It’s capabilities far exceeds your phone, computer, or screen. You can find it hanging in a gallery, on a piece of paper, on a red carpet, or floating through a room in the form of a song.
The beauty of art is that it is subjective and collaborating with something new opens the possibility of creating something completely original, which is, after all, what all real artists strive to do.
Interested in starting your AI journey? Contact us today.