The Curious Engineer

All blogs / AI / ML

FastML

fastml.com

This is not related to machine learning, but on the 25th anniversary of 9/11 we will make an exception. If you think about anything related to US …

AI / MLSociety / politics / economyEssays / personal

’15’20’25
Posts
147 from 9 Aug 2012 to 10 Sep 2026
Posting
now and then (4 in the last year)
Kind
personal blog
Feed
https://fastml.com/atom.xml

2026 3

  1. The why of 9/11
  2. Contents
  3. Detecting AI slop on Reddit in three easy steps

2025 6

  1. Paper review: SSRL - Self-Search Reinforcement Learning
  2. Stop the genocide
  3. Llama 4 smells bad
  4. SEIZED BY DOGE
  5. Beware of Musk's chatbot, Grok
  6. They are selling dollars for 69 cents now

2024 6

  1. Paper review: X-Sample Contrastive Loss
  2. They are selling a dollar coin flip for 36 cents
  3. Large language models in 2024
  4. How many letters R are there in the word STRAWBERRY?
  5. LLM snake oil drama
  6. The one thing you need to know about linear algebra for data science and machine learning

2023 6

  1. Paper review: FrugalGPT
  2. How to train your own ChatGPT Alpaca style, part two
  3. How to train your own ChatGPT Alpaca style, part one
  4. Instructions for large language models and AI assistants
  5. This is how you code now
  6. SpamGPT: watermarking large language models

2022 1

  1. Help with making a new music recommendation dataset - FastML

2021 1

  1. How to solve the cheaters problem in online shooter games - FastML

2019 1

  1. How to solve the cheaters problem in Counter Strike, with or without machine learning - FastML

2018 5

  1. One weird regularity of the stock market - FastML
  2. Classifying time series using feature extraction - FastML
  3. Google's principles on AI weapons, mass surveillence, and signing out - FastML
  4. How to use the Python debugger - FastML
  5. Preparing continuous features for neural networks with GaussRank - FastML

2017 9

  1. Two faces of overfitting - FastML
  2. Goodbooks-10k: a new dataset for book recommendations - FastML
  3. Revisiting Numerai - FastML
  4. It's embarassing, really - FastML
  5. Introduction to pointer networks - FastML
  6. Project RHUBARB: predicting mortality in England using air quality data - FastML
  7. Tuning hyperparams fast with Hyperband - FastML
  8. How to use pd.get_dummies() with the test set - FastML
  9. Data in, predictions out - FastML

2016 12

  1. On chatbots - FastML
  2. Piping in R and in Pandas - FastML
  3. Deep learning architecture diagrams - FastML
  4. Factorized convolutional neural networks, AKA separable convolutions - FastML
  5. How to make those 3D data visualizations - FastML
  6. Adversarial validation, part two - FastML
  7. ^one weird trick for training char-^r^n^ns - FastML
  8. Adversarial validation, part one - FastML
  9. Coming out - FastML
  10. Bayesian machine learning - FastML
  11. What next? - FastML
  12. What is better: gradient-boosted trees, or a random forest? - FastML

2015 14

  1. Numerai - like Kaggle, but with a clean dataset, top ten in the money, and recurring payouts - FastML
  2. What you wanted to know about TensorFlow - FastML
  3. Predicting sales: Pandas vs SQL - FastML
  4. An excerpt from The Master Algorithm - FastML
  5. Evaluating recommender systems - FastML
  6. Deep nets generating stuff - FastML
  7. Classifying text with bag-of-words: a tutorial - FastML
  8. Real-time interactive movie recommendation - FastML
  9. The emperor's new clothes: distributed machine learning - FastML
  10. What you wanted to know about AI, part II - FastML
  11. What you wanted to know about AI - FastML
  12. Juergen Schmidhuber's answers from the Reddit AMA - FastML
  13. Torch vs Theano - FastML
  14. Loading data in Torch (is a mess) - FastML

2014 28

  1. Interactive in-browser 3D visualization of datasets - FastML
  2. How to run external programs from Python and capture their output - FastML
  3. Geoff Hinton's Dark Knowledge - FastML
  4. Vowpal Wabbit, Liblinear/SBM and StreamSVM compared - FastML
  5. ICLR 2014 tidbits - FastML
  6. Comparing large-scale linear learners - FastML
  7. Michael Jordan on deep learning - FastML
  8. Kaggle vs industry, as seen through lens of the Avito competition - FastML
  9. Math for machine learning - FastML
  10. Classifier calibration with Platt's scaling and isotonic regression - FastML
  11. Vowpal Wabbit eats big data from the Criteo competition for breakfast - FastML
  12. Optimizing hyperparams with hyperopt - FastML
  13. Extreme Learning Machines - FastML
  14. How a Russian mathematician constructed a decision tree - by hand - to solve a medical problem - FastML
  15. Yann LeCun's answers from the Reddit AMA - FastML
  16. Impute missing values with Amelia - FastML
  17. Converting categorical data into numbers with Pandas and Scikit-learn - FastML
  18. Predicting happiness from demographics and poll answers - FastML
  19. Deep learning these days - FastML
  20. Exclusive Geoff Hinton interview - FastML
  21. If you use R, you may want RStudio - FastML
  22. Good representations, distance, metric learning and supervised dimensionality reduction - FastML
  23. PyBrain - a simple neural networks library in Python - FastML
  24. Are stocks predictable? - FastML
  25. Yesterday a kaggler, today a Kaggle master: a wrap-up of the cats and dogs competition - FastML
  26. Why IPy: reasons for using IPython interactively - FastML
  27. How to get predictions from Pylearn2 - FastML
  28. Classifying images with a pre-trained deep network - FastML

2013 35

  1. Regularizing neural networks with dropout and with DropConnect - FastML
  2. A/B testing with bayesian bandits in Google Analytics - FastML
  3. 13 NIPS papers that caught our eye - FastML
  4. Object recognition in images with cuda-convnet - FastML
  5. CUDA on a Linux laptop - FastML
  6. Maxing out the digits - FastML
  7. How much data is enough? - FastML
  8. Big data made easy - FastML
  9. Pylearn2 in practice - FastML
  10. What you wanted to know about AUC - FastML
  11. Predicting solar energy from weather forecasts plus a NetCDF4 tutorial - FastML
  12. Our followers and who else they follow - FastML
  13. A bag of words and a nice little network - FastML
  14. Accelerometer Biometric Competition - FastML
  15. Running things on a GPU - FastML
  16. Introducing phraug - FastML
  17. Processing large files, line by line - FastML
  18. Go non-linear with Vowpal Wabbit - FastML
  19. Amazon aspires to automate access control - FastML
  20. More on sparse filtering and the Black Box competition - FastML
  21. And deliver us from Weka - FastML
  22. Deep learning made easy - FastML
  23. Regression as classification - FastML
  24. Gender discrimination - FastML
  25. Dimensionality reduction for sparse binary data - an overview - FastML
  26. Large scale L1 feature selection with Vowpal Wabbit - FastML
  27. Choosing a machine learning algorithm - FastML
  28. Dimensionality reduction for sparse binary data - FastML
  29. Predicting advertised salaries - FastML
  30. The secret of the big guys - FastML
  31. A very fast denoising autoencoder - FastML
  32. Feature selection in practice - FastML
  33. Intro to random forests - FastML
  34. Machine learning courses online - FastML
  35. Madelon: Spearmint's revenge - FastML

2012 13

  1. Spearmint with a random forest - FastML
  2. Tuning hyperparams automatically with Spearmint - FastML
  3. Predicting wine quality - FastML
  4. The Facebook challenge HOWTO - FastML
  5. So you want to work for Facebook - FastML
  6. Merck challenge - FastML
  7. Predicting closed questions on Stack Overflow - FastML
  8. Best Buy mobile contest - full disclosure - FastML
  9. Best Buy mobile contest - big data - FastML
  10. Best Buy mobile contest - FastML
  11. Running Unix apps on Windows - FastML
  12. Kaggle job recommendation challenge - FastML
  13. What you wanted to know about Mean Average Precision - FastML

Undated 7

  1. Background images - FastML
  2. GitHub recommendations - FastML
  3. Ideas - FastML
  4. Popular: top ten most viewed pages, as reported by Google Analytics - FastML
  5. Solutions - FastML
  6. Thank you - FastML
  7. Toy store - FastML