A Review of Popular Deep Learning Models
This post is a concise overview of a few of the more interesting popular deep learning models to have appeared over the past year. Get up to speed and try a few of the models out for yourself.
By Jason Toy, Founder of Somatic.
In the financial services industry, deep learning models are being used for "predictive analytics," which have helped improve forecasting, recommendations, and risk analysis. HR departments are using AI-based tools to help streamline the process of talent acquisition and management. Facebook has just developed a Bot store to help businesses take advantage of chat bots that can act as customer service representatives for sales or with simple troubleshooting. As deep learning algorithms become increasingly prevalent across industries, deep learning models are also becoming more accessible to people outside of mathematics, engineering and robotics. We hope this showcase can inspire you to see what is possible.
Neural Talk is a vision-to language model that analyzes the contents of an image and outputs an English sentence describing what it “sees.” In the example above, we can see that the model was able to come up with a pretty accurate description of what ‘The Don’ is doing.
Computer sees: a man in a suit and tie standing in front of a wall
This model was trained using pictures from Flickr and captions that were generated by crowdsourcers on Amazon’s Mechanical Turk. The model looks for related images in its training data and examines the captions to synthesize what is occurring in the input image. Try the demo here.
If you have ever used Instagram or Snapchat, you are familiar with using filters that alter the brightness, saturation, contrast, and so on of your images. Neural style, a deep learning algorithm, goes beyond filters and allows you to transpose the style of one image, perhaps Van Gogh’s “Starry Night,” and apply that style onto any other image. In the example above, we see the style of a pencil sketch applied to the selfie of a young man. You can play around with the demo here.
Neural Style is one of the first artificial neural networks (ANNs) to provide an algorithm for the creation of artistic imagery. This exciting technology could be used as a valuable tool for creative professionals to test new ideas and concepts, saving a great deal of time and maximizing efficiency.
Inception is a model that allows you to see what it looks like when a computer “dreams” (they seem to have a lot of nightmares!) The model’s hierarchy structure can be customized and flipped upside down so that the computer is “tricked” into believing that there are hundreds of a particular object inside of an image. Then, the model enhances all of those objects in the image. The result looks something like a bad acid trip. Play around with inception here.
Bio: Jason Toy is a serial entrepreneur and engineer, who founded socmetrics, truelens, and is the former CEO of filepicker and CTO of backchannelmedia. He builds machine learning models, researches artificial intelligence, and starts companies. He is the founder of Somatic.
About: Somatic is a deep learning platform that aims to bring deep learning to the masses. They offer pretrained artificial intelligence models so that you can easily integrate them into your existing mobile apps, web applications, or anywhere else.