by Aisling Love
29 July, 2016 - 2 minute read

On Thursday 28th July, Winton hosted the latest in the series of the TensorFlow London meetups.

TensorFlowâ„¢ is 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. The flexible architecture allows you to deploy computation to one or more CPUs or GPUs in a desktop, server, or mobile device with a single API. TensorFlow was originally developed by researchers and engineers working on the Google Brain Team within Google’s Machine Intelligence research organization for the purposes of conducting machine learning and deep neural networks research, but the system is general enough to be applicable in a wide variety of other domains as well.

The first speaker at the event was Steven Hutt of CME group. Steven mercifully gave a thorough introduction to the structure and mechanism of the models he had used in the work that he was presenting. These included Recurrent Neural Networks, Autoencoders and LSTMs. Steven reported promising results in recognising the limit order book structure of currency futures contracts and outlined how he had implemented the training of these models with TensorFlow.

The discussion during the Q&A largely focussed on whether it was practical for High Frequency Traders to use these techniques. The verdict was maybe not but pretty soon it will be!

The second speaker was Yaz Santissi. Yaz currently works freelance for Google and is the lead developer of the Google Developer Group for The Cloud. Yaz gave an accelerated workshop on image recognition with TensorFlow. He trained the Inception v3 image recognition model (packaged with TensorFlow) on images of flowers and performed an impressive live test of its ability to identify a tulip from a random google image search.

The volume of questions is testament to the fact that attendees enjoyed the talks as much as the free pizza and beer. We look forward to hosting similar events in the future and to further integration with the machine learning community.

To attend future TensorFlow meet-ups, you can sign up here: