PyData is back in Bristol and running frequently again!
We'd like to thank our hosts Cookpad for providing the venue, pizza 🍕 and refreshments 🍺🧃🥤, and Adlib for additional sponsorship.

Expect two 25-minute talks, two 5-minute lightning talks, community announcements and some relaxed networking over beers and soft drinks.

🕡 Note: doors open from 6, and the talks start at 6:30 pm sharp!

Programme:

1. Building a Public-Facing and Low-Maintenance Wordle Solver with Unsupervised Learning by Jason Chao
2. [Lightning] Software to Data: Lessons from a Crash Landing by Tom Foster
3. [Lightning] Introducing PPML: Machine Learning on data you cannot see by Valerio Maggio
4. Introduction to Federated Learning by Akiko Ogawa

If you would like to speak at this or a future event - please fill out this form: https://goo.gl/forms/8lsz1WA1986Ahbbs1

TALKS

1. Building a Public-Facing and Low-Maintenance Wordle Solver with Unsupervised Learning

Wordle is a popular online game that challenges players to guess a five-letter word in six tries. While most Wordle solvers use supervised learning techniques to target specific word lists, this talk will explore a different approach. Join Jason Chao on his journey of building a public-facing Wordle Solver using unsupervised learning. He will showcase the possibility of rapidly developing a low-maintenance and future-proof side project.
Despite time constraints, Jason built a Wordle Solver that works with multiple Wordle versions and supports variable word lengths. He will discuss how unsupervised learning helped him create a more robust and versatile solver. Additionally, he will share his experience of running the solver with almost zero maintenance and at a low operating cost.

# 2. Introduction to Federated Learning

Federated learning is an emerging machine learning methodology that involves deploying and training models on edge devices without data needing to be moved to the cloud.
The talk will explain in 25 minutes about how AI in distributed systems and IoT could solve new problems where sensor data required for training could be compromised or insufficient. It will give some insight into the opportunities for automating the coordination between machine-to-machine and machine-to-human to bring about collective intelligence and a ‘hive-mind’.
If there is time, I might give a quick showcase of an open-source federated learning Python package called ‘flwr’ which enables a gRPC network of multiple edge clients to connect with each other and train/evaluate any custom-developed machine learning model.

LIGHTNING TALKS

3. Software to Data: Lessons from a Crash Landing by Tom Foster

Moving from software development to data engineering can be a challenging transition, especially when it comes to landing your first data project. In this lightning talk, Tom Foster shares his personal journey from software to data, the key concepts he found useful to know before diving into data engineering, and the lessons he learned along the way. Tom also discusses his experience working on a PySpark and Databricks project, providing insights and tips that may be helpful to others making a similar transition.

# 4. Introducing PPML: Machine Learning on data you cannot see

What if I would tell you that any Machine learning model could be exploited to gather information about the data originally used for training? And if that data contains sensitive information, well: privacy guarantees are crucial and unavoidable. Ideally, we would need machine learning models to work on data without really looking at the data. In this talk, I will give a short introduction to model attacks, and methods to prevent them using privacy-preserving methods.

🕖 LOGISTICS
Talks kick off at 18:30 sharp; then networking/beers in The Knights Templar from 20:40.
If you realise you can't make it, please un-RSVP in good time to free up your place for your fellow community members.
Follow @pydatabristol (https://twitter.com/pydatabristol) for updates on this and future events, as well as news from the global PyData community.

📜 CODE OF CONDUCT
The PyData Code of Conduct governs this meetup (https://pydata.org/code-of-conduct/). To discuss any issues or concerns relating to the code of conduct or behaviour of anyone at the PyData meetup, please contact the PyData Bristol organisers, or you can submit a report of any potential Code of Conduct violation directly to NumFOCUS (https://numfocus.typeform.com/to/ynjGdT).