Today, at their annual re:Invent conference, Amazon announced a brand new Marketplace for Machine Learning. The idea is that customers can purchase production-ready ML models from an extensive library of algorithms created by some of the most innovative companies in the space. It’s a great service, giving companies without massive machine learning and data science teams models they can use to solve a wide variety of real-world business challenges, and we’re thrilled to be an inaugural partner.
Figure Eight has been hard at work creating models for the marketplace. Our machine learning team has provided eleven different models, from facial anonymizers for privacy in images to audio categorization and more, and we’re incredibly proud to see them out in the wild.
To learn a bit more about how to use Figure Eight’s models with Amazon SageMaker, watch the video below featuring our Sr. Machine Learning Engineer, Rahul Parundekar.
Here are a few of our favorite machine learning model offerings:
This model blurs faces in imagery. It can be powerful for everything from mapping software to security services or really any other space where privacy is of paramount concern.
Audio Gender Detection
This model helps identify whether a speaker in an audio snippet is a man or woman. Trained with English speakers, mostly from North America (though with myriad accents and dialects), this model can be helpful for customer service, sales forces, audio assistants, and more.
Skin Disease Classifier
This model classifies skin images which potentially contain disease. Trained on acne rosacea, actinic keratosis, basal cell carcinoma, and more, this model makes predictions about skin disease from images.
Background Noise Classifier
This model classifies audio as “noisy” or “not noisy” to allow your model to identify speakers in audio with static, ambient noise, barking dogs, street sounds, or any other background noise.
This model will generate lyrics for your next Billboard-topping single. Admittedly, this one won’t solve a ton of your core business problems, but we had a lot of fun with it and wanted to share it with everybody.
Want to take any of these for a spin? Head to the AWS Machine Learning Marketplace! They’re available now.