# Exploring approaches to Graphs + Machine Learning **URL:** \ **Category:** Projects & Collaboration\ **Created:** [October 1, 2018, 8:18pm UTC](https://community.neo4j.com/t/exploring-approaches-to-graphs-machine-learning/2196 "2018-10-01T20:18:10Z")\ **Posts on this page:** 1\ **Page:** 1
**Author:** ![david1](https://sea1.discourse-cdn.com/flex021/user_avatar/community.neo4j.com/david1/32/1532_2.png) [@david1](https://community.neo4j.com/u/david1)\ **Post date:** [October 1, 2018, 8:18pm UTC](https://community.neo4j.com/t/exploring-approaches-to-graphs-machine-learning/2196/1 "2018-10-01T20:18:10Z")
Hi, Thanks Karin for welcoming me to the community! My name's David, I'm the co-founder of [https://octavian.ai](https://octavian.ai), a research project into new approaches to machine learning on graphs. My background is from running a tech startup ( SketchDeck) and studying Computer Science and Mathematics. We use Neo4j in a bunch of our work! We've put together tutorials on how to connect Neo4j with Keras and TensorFlow, and also how to translate English into Cypher: > **[Answering English questions using knowledge graphs and sequence translation](https://medium.com/octavian-ai/answering-english-questions-using-knowledge-graphs-and-sequence-translation-2acbaa35a21d)** > > In sci-fi films like Star Trek and 2001: A Space Odyssey, crew-members walk around asking their ship questions about things in its… > **[Review prediction with Neo4j and TensorFlow](https://medium.com/octavian-ai/review-prediction-with-neo4j-and-tensorflow-1cd33996632a)** > > We show how to create an embedding to predict product reviews, using the TensorFlow machine learning framework and the Neo4j graph… > **[An introduction to using Keras with Neo4j](https://medium.com/octavian-ai/an-introduction-to-machine-learning-on-graph-databases-24ee502fd12e)** > > We demonstrate connecting a Neo4j graph database to Keras. We build a neural network achieving 100% test accuracy on a simple review… My current work focuses on performing Question-Answering on Knowledge Graphs. A lot of our approaches involve having the neural network directly read/manipulate the graph itself (E.g. without a DB engine) although I'm also really interested in combining Reinforcement Learning with Neo4j + Cypher to do graph QA. More than happy to chat with y'all about graphs and ML :)