PyData London - 48th Meetup
NOTE: A valid photo ID is required by building security. Please use your full real names when signing up, otherwise you may be refused entry!
As always, there'll be free food & drinks, generously provided by our host, AHL.
We are issuing tickets via a lottery - if you want to be in with a chance of a place - sign up for the waitlist! The lottery will be run approx 1 week before the meetup, and we will re-run the lottery to fill any spaces that free up or use the waitlist towards the time of the event.
Jan Freyberg on "Humans can support algorithms - and vice versa - with the interactive python ecosystem"
Abstract: Often, data scientists are faced with unlabelled datasets. While unsupervised machine learning methods exist, they often don’t achieve performance close to what supervised learning methods achieve. One way of meeting halfway is semi-supervised learning. A particularly intuitive subdomain of semi-supervised learning is called active learning, where the algorithm requests new data points to be labelled. Active learning employs the same classification models as supervised learning, which are iteratively improved by labelling data points which are the most useful for the algorithm. This means that a human sitting down to label datapoints only needs to label the points that are actually useful - not the ones the algorithm is already sure about. In this talk, I want to go through the rationale behind active learning, and demonstrate how it can be done in the interactive python ecosystem. I’ll also outline how to use the superintendent (github.com/janfreyberg/superintendent) package, specifically designed for active learning.
Peadar Coyle on "A modern introduction to Hamiltonian Monte Carlo and Bayesian workflows"
Abstract: Probabilistic Programming is a new paradigm enabling a better
understanding of uncertainty. I introduce a modern Bayesian workflow and explanation of Hamiltonian Monte Carlo - how it fails and how you debug it.
Robert Stojnic on: LazyData : Scaleable Data Dependencies
Abstract: lazydata is a minimalist library for including data dependencies into Python projects.
Keeping data files in git (e.g. via git-lfs) results in a bloated repository that takes ages to pull. So
lazydata only stores references to data files in git, and syncs data files on-demand when they are needed. In this brief talk, I'll show an example of how lazydata can work for a machine learning or data science project.
And if you want to include a link: https://github.com/rstojnic/lazydata
Ross Taylor on Mantra: A Deep Learning Development Kit
Abstract: Mantra is an open-source, deep learning development kit that manages the various components in a deep learning project. I'll show a brief example of how it makes routine tasks like training in the cloud, model monitoring, benchmarking and more much easier.
Doors open at 6.30pm (get there early as you have to sign-in via AHL's security), talks start at 7 pm, drinks from 9 pm in the bar. We normally have >200 folks in the room so there's plenty of people to discuss data science questions with!
Please unRSVP in good time if you realize you can't make it. We're limited by building security on the number of attendees, so please free up your place for your fellow community members!
Follow @pydatalondon (https://twitter.com/pydatalondon) for updates and early announcements. See you on the 4th!