What we're about
This meetup is primarily for those interested in Deep Learning, using any toolkit. This project is an open source effort to expose cutting edge machine learning and AI to users. The Autonomous Intelligent Agent is a high level user interface and autonomous agent which utilizes trained machine learning models to perform a variety of tasks. The agent will be written in Java and Python. In the initial version it will perform the following.
Creative: Train on large collections of poetry, prose, music lyrics, images and art, and produce music, poetry, and art. For example, train RNNs (Recursive Neural Networks) with the Shakespeare sonnets and create new ones. Use Adversarial networks to generate high quality art. The agent will retrain models using input data selected and ranked for the user.
Economic: Stock market prediction and portfolio recommendations using Deep Learning and AI. Sports gambling recommendations.
Biology: Classification of plant, animal, and bacteria images. Ability to automate workflows for DIY Biology, e.g. qPCR.
The agent will utilize a natural language interface. It will have an AI based planning system and utilize a distributed workflow system for retraining. It will have a built in system for semantic reasoning. Future versions will have causality inference.
Platforms for this project will be iOS, Android, PC, and Mac. The main Deep Learning library will be Tensorflow, and the client apps will be written in Java and Python.
Please attend if you're interested in developing on the Intelligent Agent framework!