Hi The MUFC Academy 🥇🔴⚪️⚫️🐝🍷, any idea or leads on when was the first ever direct free kick recorded ?

That is some deep learning from life

Wishing a healthy and long life for the DD news commentator who could not hold back his "tch tch tch... Disaster..." remark upon seeing this performance by BJP's national president:

Subtitles aren't just for deaf people. Lots of my hearing friends use them, too. If you're hearing and find yourself using subtitles on Netflix and TV and would quite like them at the cinema, please retweet to help normalise their presence! Big thanks #DeafAwarenessWeek

Excited to share our paper on OpenML-Python, a Python API for OpenML.org. Check it the API at openml.github.io/openml-python. With @janvanrijn, Arlind Kadra, Pieter Gijsbers, Neeratyoy Mallik, Sahithya Ravi, Andreas Mueller, Joaquin Vanschoren and Frank Hutter CC OpenML #openml

The frequency of random seeds between 0 and 1000 on github (data from grep.app)

Surprising turn in #HPO and #NAS: we now have DEHB, an evolutionary successor of BOHB that's up to 1000x faster than random search for optimizing 6 NN hypers and up to 32x faster than BOHB on NAS-Bench-B201! #IJCAI21 paper: arxiv.org/abs/2105.09821 Code: github.com/automl/DEHB

5. HPOBench: openreview.net/forum?id=1k4rJ…, with @KEggensperger, Neeratyoy Mallik, Philipp Müller, Matthias Feurer, Noor Awad, René Sass, Aaron Klein, @LindauerMarius -- 100+ blackbox and multi-fidelity benchmarks, containerized code and tabular versions. No more excuse for only Branin!

When the robot revolution comes, it will be stopped only by pip install errors

This may revolutionize data science: we introduce TabPFN, a new tabular data classification method that takes 1 second & yields SOTA performance (better than hyperparameter-optimized gradient boosting in 1h). Current limits: up to 1k data points, 100 features, 10 classes. 🧵1/6

In our #NeurIPS2023 paper "PriorBand: Practical Hyperparameter Optimization in the Age of Deep Learning", we propose PriorBand a new efficient and practical #HPO method that combines expert priors and multi-fidelity optimization. With Neeratyoy Mallik, Carl Hvarfner, Eddie Bergman,

You can now use autopilot in India with an Android with Snapdragon!! FlowPilot is an open-source fork of Comma AI's OpenPilot developed by mansin and co. More than that, it exemplifies the tremendous talent young (non-IIT) Indian devs have! 1/6

We the reporters' collective are making public all documents related to electoral bonds, including those now removed by the State Bank of India from its website. And those that it never disclosed in the first place. You can find them here: drive.google.com/drive/folders/…

This discovery excites me in confusing ways :')

This might be the first time after 10 years that boosted trees are not the best default choice when working with data in tables. Instead a pre-trained neural network is, the new TabPFN, as we just published in Nature 🎉