Blog

In a world of recycled content and SEO-driven noise, I’m carving out a space for something different. Here, I share some of my favorite blogs from others—those that stand out for their originality and substance. If I publish my own work, you can trust it’s something I’ve put real thought into, aiming for authenticity over clicks. This is a place for ideas, not algorithms. Let’s keep it real, and explore content that’s worth reading.

Flow Reinforce Blog

by Google DeepMind and University of Cambridge researchers

Best blog to showcase their paper ever, with interactive demos and clear explanations.

KV Frans's Blog

by Kevin Frans

This blog offers some of the highest-quality resources on deep learning and reinforcement learning.

Value Scaling

by Researchers at Google DeepMind

A blog post discussing scaling laws and their implications for large language models, providing a clear, detailed analysis.

WTF happened to BLOGS

by Michal Pándy

Now, instead of passionate creators sharing their thoughts, we have armies of underpaid writers churning out “10 Ways to Boost Your Productivity (You Won’t Believe #7!)”

Graph Based SLAM with GTSAM

by Shubodh Sai, Udit Singh Parihar, Ansh Shah, August 2023

This blog contains lecture material and coding exercise for Graph Based SLAM. I was a visiting lecturer for the Mobile Robotics course in IIIT Hyderabad. I took lectures on SLAM and PoseGraph Optimisation

Diffusion Meets Flow Matching: Two Sides of the Same Coin

by Ruiqi Gao, Emiel Hoogeboom, Jonathan Heek, Valentin De Bortoli, Kevin P. Murphy, Tim Salimans

How Flow Matching and Noise Conditioned Score Models come to the same underlying principles

Energy Based Models

by Michal Pándy

This blog provides an introduction to Energy-Based Models (EBMs), explaining how they assign energy values to data configurations for modeling probability distributions. It covers the fundamentals of EBMs, their applications in machine learning, and their increasing use in areas like generative modeling and unsupervised learning.

Diffusion Models

by Lilian Weng

Diffusion models are a class of generative models that learn the data distribution by iteratively diffusing a simple base distribution.

Spinning Up RL

by OpenAI when they were open

Holistic resource on hands-on basic RL algorithms.