Room 200, Rockefeller Hall, Vassar College, Poughkeepsie, NY
LT1: Eric Myers - OpenSpace: A new OpenSource interactive data visualisation softrware suite designed to visualise the entire known Universe and portray out ongoing task to investigate the Cosmos! Eric is involved with the development of this marvelous bit of software and opened the meeting by announcing an event at the Hayden Planetarium to generate code for the project. Check out the OpenSpace project!
LT2: Sean Dague - Grafana - Visualising time series data. Sean presented his use of Grafana, using environmental sensors in his home to illustrate the potential of visualising data.
Main Talk: Nitya Narasimhan - Tensorflow
TensorFlow is an OpenSource library for Machine Intelligence. "Machine Learning" is the ability of systems to adapt and learn without specific programming toward a particular goal. This normally requires data "retraining" from your specific data.
TensorFlow is (from the web page) " an open source software library for numerical computation using data flow graphs. Nodes in the graph represent mathematical operations, while the graph edges represent the multidimensional data arrays (tensors) communicated between them."
Nitya presented an overview of Machine Lewarning and some of the steps taken in Tensorflow to develop basic tasks in this arena. She illustrated how a system was able to identify hand-written numbers of various quality through application of TF modules.
Nitya also discussed neural networks and their application to machine learning
She introduced methods of learning TF (TFLearn, workshops and others) and took many questions from those attending this very interesting talk. The images attached have some of the links you can use if interested in knownig more about TensorFlow.