June 19, 2018
Since its inception, Neuromation has been working with creative professionals to find innovative use cases for artificial intelligence. Our goals are to improve the quality of work output while streamlining and simplifying the creative process across a range of use cases; and to use cutting edge tools to allow designers, makers and builders to expand the realm of what is possible.
Our first collaboration on this front comes with an ambitious startup called Let’s Enhance. Let’s Enhance seeks to dramatically improve the capability for upscaling the resolution on old, low-res or blurry photos. Through its partnership with Neuromation, Let’s Enhance employs Deep Convolutional Neural Networks trained on a dataset of real and synthetic images, allowing them to learn typical features of objects so the application can add extra details that were not present in the original. This moves beyond what was possible with the previous state of the art — bicubic interpolation — which is employed by industry leading platforms like Adobe’s Photoshop.
But this is only the beginning of the potential for AI in creative industries.
Adobe, which has long been the premiere creator of professional creative tools, has begun incorporating AI face recognition technology in its products, allowing for what it calls face-aware features that apply effects and filters to images of faces. You could argue that this technology was first used in consumer smartphone applications and on platforms like Snapchat for making photos look better and accurately applying layers to video. This illustrates how this technology can move from professional to consumer, but also from consumer to professional as AI-powered ease of use finds applications across the board.
Over the past decade, the advertising industry has seen an explosion of technological development as adtech and real-time bidding have revolutionized online advertising. The entire industry has become more data driven and technology dependent — from audience segmentation and intent attribution to fraud detection and sentiment analysis.
In this area, IBM has adapted it’s Watson technology for optimizing programmatic media buying, a technology it is calling “Cognitive Bid Optimizer”.
And as we have seen in companies like Facebook, AI-powered solutions are also seeing great success in predicting fraud as well as identifying malicious content, both of which are extremely important to responsible and image-sensitive brands.
Now, however, artificial intelligence technology has the potential to revolutionize even the most creative aspects of the industry.
Automation of labor-intensive and time consuming menial work can free up designers and creative directors to create new images and experiences. Automation can also save significant costs, stretching a company’s marketing budget. This process has happened before, of course, as PC’s simplified work which had previously been done by hand. AI now takes these efficiency gains further by automating processes previously carried out by computer design technicians, not only realizing time and cost savings but also allowing for work that is significantly more detailed and accurate then had previously been possible.
Another aspect in which artificial intelligence can assist with design is by analyzing user interaction with an application or website using machine learning and to recommend modifications that would allow the site to be easier to use and more responsive. Companies such as AirBnB and WeWork are experimenting with this technology, as are giants like Google and IBM.
Finally, there has been a lot of attention given to purely generative design technologies such as those achieved with the application of GAN’s, or generative adversarial networks. These neural network models, when trained on sets of images can learn to ‘hallucinate’ completely new yet visually similar images. These images can have an otherworldly or dream-like quality, but as we have seen with Let’s Enhance, when applied to small areas of images they can be extremely effective in filling in blanks that were not present in the original data.
Fashion design is another area where generative design technology is being experimented with. Amazon, in cooperation with Lab126, a research studio in San Francisco is now experimenting with AI-designed clothing. After learning about a given style of fashion from a series of photos, the AI can create a completely new yet related design.
A core belief at Neuromation is that goal of artificial intelligence should not be to replace people, but to assist us and augment our capabilities. We seek to build tools, instruments and vehicles for creative exploration and discovery. The creative tools that we have seen to date and that most excite us may use artificial intelligence to make work easier and to expand what is possible, but they still require the use of creative individuals who can make important aesthetic decisions. We look forward to continuing to work with these inspiring creatives and developing better and better tools for them using artificial intelligence. As we have seen with ecommerce and website creation platforms, this can also put cutting edge and responsive design within the reach of smaller companies who previously have not been able to afford dedicated creative professionals. Artificial intelligence is an important next step in this ongoing evolution of creative industries.
by Angus Roven
Senior Analyst, Neuromation