List of videos

Towards a sustainable solution to open source sustainability [EuroPython 2021 - Talk - 2021-07-28 - Optiver] [Online] By Tobie Langel A few years ago, Heartbleed epitomized a massive open source sustainability problem for critical parts of the internet infrastructure. The bug, which affected the popular OpenSSL cryptographic software library, notably compromised the confidentiality of 4.5 million US patient records and cost the industry an estimated $500M. It was soon revealed that the root cause of the issue was that OpenSSL was precariously understaffed. Open source sustainability became a major theme overnight. Stories of maintainer burn-out made the headlines. And tentative solutions started to emerge, most of them donation-based. In this talk, we’ll explore a number of existing strategies to fund open source and make it more sustainable, from patronage to dedicated ad networks. And we’ll defend the idea that the best path to open source sustainability is to help companies understand the tangible business value they can get from contributing to open source. License: This video is licensed under the CC BY-NC-SA 4.0 license: https://creativecommons.org/licenses/by-nc-sa/4.0/ Please see our speaker release agreement for details: https://ep2021.europython.eu/events/speaker-release-agreement/

Hypertag: an Indented Language for easy HTML Generation & Web Templating [EuroPython 2021 - Talk - 2021-07-28 - Brian] [Online] By Marcin Wojnarski HYPERTAG (http://hypertag.io) is a full-featured standalone language for HTML templating and document generation. Inspired by indentation-based template languages (Slim, Haml, Shpaml, Pug), it provides clean, readable syntax and multiple original features that enhance code reuse and modularity: native custom tags, DOM manipulation, Python-like granular imports, control blocks, compound expressions, and more. Hypertag can be called directly from Python code, or be plugged as a template backend into Django. Hypertag scripts can utilize all of Django's filter functions out of the box. Hypertag is not limited to HTML alone, but can be adapted to produce different markups. Docs: http://hypertag.io GitHub: https://github.com/mwojnars/hypertag PyPI: https://pypi.org/project/hypertag-lang/ During the talk, Hypertag's syntax will be presented with particular focus on original features not found elsewhere (custom native tags, DOM manipulation, granular imports). The talk will be illustrated with examples, and a comparison with existing template languages will be shown. License: This video is licensed under the CC BY-NC-SA 4.0 license: https://creativecommons.org/licenses/by-nc-sa/4.0/ Please see our speaker release agreement for details: https://ep2021.europython.eu/events/speaker-release-agreement/

Virtual Tourism In Covid Times [EuroPython 2021 - Talk - 2021-07-28 - Ni] [Online] By Nithish Raghunandanan The introduction to the talk would cover the motivation for the solution. As someone who loves to travel and has been travelling quite extensively in Europe over the past years, the lockdown & travel restrictions preventing travels are quire hard to adjust to. This gave me the idea to build a dashboard around my past travels to improve my spirits. In the next part, I cover how I get the data. In this case, I use the location history from Google Maps. A brief description of the data is included along with what data I select. I also cover some of the problems faced like the non standardization of the raw data (timestamps, latitude / longitude pairs) & how to resolve it. In the next part, I explain how I can do some basic analysis with the raw data like the distribution & visualization of my travels over time interactively. This will go through some code examples on how you can build these components using matplotlib & how to make them interactive using Streamlit without writing any Javascript code. In the next section, I explain how I can augment this data with data from other sources like public images from Flickr to bring out my memories. This would also explain how to integrate these components using Streamlit. This would also cover the deployment aspects of the solution. Also covered here would be some future additions that I am considering. In the final part, I conclude the presentation with my learnings from building this app to do some virtual tourism in COVID times. I cover some of the scenarios where Streamlit could be useful & other places where it might not be suitable. License: This video is licensed under the CC BY-NC-SA 4.0 license: https://creativecommons.org/licenses/by-nc-sa/4.0/ Please see our speaker release agreement for details: https://ep2021.europython.eu/events/speaker-release-agreement/

SQLite, an (un) known super ant [EuroPython 2021 - Talk - 2021-07-28 - Parrot [Data Science]] [Online] By María Andrea Vignau Outline - Introduction: advantages, best and worst use cases. - Why this little database is amazing and where it can be used at its best. - SQLite & Python: best practices - Use of context managers, row factories, executing many operations at once and the concept of PRAGMAs - Transactions, isolation and concurrency - Transactions management in SQLite, use of WAL, Rollback and no journaling modes. Isolation levels: exclusive, immediate and deferred - Indexing - Why and why not indexing. Partial index, expression index, FTS index. - Special data types - Date & Time data, personalized data types License: This video is licensed under the CC BY-NC-SA 4.0 license: https://creativecommons.org/licenses/by-nc-sa/4.0/ Please see our speaker release agreement for details: https://ep2021.europython.eu/events/speaker-release-agreement/

Introduction to Quantum Deep Learning [EuroPython 2021 - Talk - 2021-07-28 - Optiver] [Online] By Abhilash Majumder Introduction to Quantum Deep Learning Abstract The aim of the lightning talk is to shed light into the field of Quantum computation in the field of Deep Learning. Qubits , which form the fundamental units of quantum computing can be used to create quantum variational circuits which can be placed over traditional deep learning networks to create hybrid quantum-deep learning models. These models not only rely on the gradient convergence properties of general backpropagation technique, but also on the final probabilistic states of the Qubits. Essentially there has been quite a development to optimize the gradient convergence of these hybrid models with the help of Fischer approximation and Natural Gradient Descent.The talk would focus on the importance of Quantum Variational Deep Learning Circuits and how they provide an advantage over traditional Autograd based Circuits. The application of Quantum Variational circuits in the field of Reinforcement Learning as well as NLP would be one of the main points of the talk. There has been sufficient development in the field of quantum computing and this talk aims to throw light on how to exploit the probabilistic states of Qubits to enhance deep learning models. Topics: Introduction to Quantum Computing and Qubit system Quantum Variational Circuits Creating Hybrid Circuits (Classical-Quantum-Classical etc.) Realizing Performance of Hybrid Circuits Applications in the field of Quantum RL and Quantum NLP (research) Democratizing adoption of Quantum Circuits over traditional deep learning circuits Resources Resources (slides, repositories) would be added in course of time. Resources: https://drive.google.com/file/d/1gtCc8JViacFhlV3-XfKlGruh_OqI0ZIa/view?usp=sharing License: This video is licensed under the CC BY-NC-SA 4.0 license: https://creativecommons.org/licenses/by-nc-sa/4.0/ Please see our speaker release agreement for details: https://ep2021.europython.eu/events/speaker-release-agreement/

Pydon'ts [EuroPython 2021 - Talk - 2021-07-28 - Ni] [Online] By Rodrigo Girão Serrão EDIT: Talk slides available at https://github.com/mathspp/talks; you can reach out to me with comments/questions on Twitter @mathsppblog or through email at rodrigo at mathspp dot com. In this talk I show what idiomatic Python code looks like and illustrate how the subjective notion of “beautiful”, or “elegant” code, can make your code objectively better, regardless of your Python skill level. We do that by visiting several vanilla Python features that people sometimes forget to learn, as they rush to importing all the shiny modules and frameworks they really want to use. In order to achieve this, we go through the learning journey of a hypothetical Pythonista, taking a function written by them and refactoring it incrementally through the use of Pythonic idioms and patterns. The whole talk will build around a weekly series of articles I have been publishing (at https://mathspp.com/blog/pydonts), where I explore all the vanilla Python features that make Python one of the most interesting programming languages to learn. There are no prerequisites for this talk, and both advanced users and beginners can benefit from it. License: This video is licensed under the CC BY-NC-SA 4.0 license: https://creativecommons.org/licenses/by-nc-sa/4.0/ Please see our speaker release agreement for details: https://ep2021.europython.eu/events/speaker-release-agreement/

Auto-Sklearn: Automated Machine Learning in Python [EuroPython 2021 - Talk - 2021-07-28 - Parrot [Data Science]] [Online] By Matthias Feurer, Katharina Eggensperger AutoML makes machine learning accessible for everyone! Machine Learning is hard since developing well-performing machine learning pipelines requires a lot of expertise, time and manual tuning. AutoML automates this development process by using latest optimization methods to efficiently search for well performing solutions. In this talk, we will cover how to move from manually constructing and tuning machine learning pipelines to using efficient hyperparameter optimization algorithms and full AutoML using the popular open-source Auto-sklearn library. Auto-sklearn is a drop-in replacement for any scikit-learn estimator and is developed by the ML Lab of the University of Freiburg. More specifically, you’ll learn the following: What is AutoML and for what can you use it? How does Auto-sklearn work? How can you use it? This talk assumes basic understanding of machine learning and statistics. The main target audience are data scientists and domain experts using machine learning. The talk will be designed such that anyone with a basic understanding of machine learning pipelines in scikit-learn and the Python language would be able to understand the concepts and to use our tool. License: This video is licensed under the CC BY-NC-SA 4.0 license: https://creativecommons.org/licenses/by-nc-sa/4.0/ Please see our speaker release agreement for details: https://ep2021.europython.eu/events/speaker-release-agreement/

The Pattern: Machine Learning Natural Language Processing meets VR/AR [EuroPython 2021 - Interactive session - 2021-07-28 - Argument-Clinic [Interactive]] [Online] By Alexander Mikhalev, Dvir Dukhan Open-source project "The Pattern" is the platform to inspire collaboration for building modern natural language processing tools and techniques and making them practically useful for end-users. In this session, I will walk through creating 3 Machine Learning NLP pipelines and deploying them using Redis and Redis Modules: RedisGears, RedisGraph and RedisAI. The first pipeline demonstrates how to use traditional NLP techniques, such as the Aho-Corasic algorithm to build a knowledge graph from medical literature using a medical thesaurus (UMLS). The second pipeline demonstrates how to build and deploy BERT Question/Answering model and create API for text to speech interface. Leveraging Redis Cluster sharding and capturing Redis Gears "keymiss" event to trigger processing of BERT QA in RedisAI. The third pipeline demonstrates how to deploy Google's T5 (text to text transfer transformers) to build summary of each article. License: This video is licensed under the CC BY-NC-SA 4.0 license: https://creativecommons.org/licenses/by-nc-sa/4.0/ Please see our speaker release agreement for details: https://ep2021.europython.eu/events/speaker-release-agreement/

DIY Home Automation with Microcontrollers and CircuitPython [EuroPython 2021 - Talk - 2021-07-28 - Optiver] [Online] By Maria Jose Molina-Contreras, PhD For many years, people have been filling their home with devices that are based on Microcontrollers, like virtual assistants, smart fridges, washing machines, security systems, and many others that can be either programmable, or even replicated with a couple of boards and our Python knowledge. During this talk, we will start exploring CircuitPython without the need of a physical device, using a microcontroller emulator. If you are new to the microcontroller world, don’t worry, we will have an introduction to the main concepts, so we have the base knowledge to start doing little experiments by ourselves. After the introduction, you will be able to start programming microcontrollers using the emulator or in real life, moreover, you will see a couple of real projects (with real devices!) that could easily be replicated by yourself: a low-cost robot floor cleaner, and a plant watering and monitoring system, as examples. Outline Talk motivation Introduction to Microcontrollers and CircuitPython. How to use an emulator for hardware programming. DIY Home Automation (Real projects) : Low-cost robot cleaner. Watering and monitoring plant system. License: This video is licensed under the CC BY-NC-SA 4.0 license: https://creativecommons.org/licenses/by-nc-sa/4.0/ Please see our speaker release agreement for details: https://ep2021.europython.eu/events/speaker-release-agreement/