nowfound

Life & fun · September 18, 2021

VA

Variational Autoencoder: A Simple Illustration

I want to illustrate what the variational auctoenocder (VAE) does. Biases (z) of my coins are uniform, and I have a device which changes a bias to z*t (t is between 0 and 1).

Variational Autoencoder: A Simple Illustration launched on September 18, 2021 with 6 votes, #1,463 of 2,941 launches that month and more than 50% of that year's launches. The site was still online in September 2026, 5 years on.

Visit linkedin.com ↗Alternatives

What it does

I want to illustrate what the variational auctoenocder (VAE) does. Biases (z) of my coins are uniform, and I have a device which changes a bias to z*t (t is between 0 and 1). I pick two coins, scale biases, and toss each coin two times. First coin: Tail-Tail Second coin: Head-Tail I want to guess t and biases of two coins. I need notation H = z*t T = 1 - z*t and i(expr) is an integral of expr over z from 0 to 1. E = i(T*T)*i(H*T) is a polynomial of t with the maximum at t = 0.473 and two conditional distributions are weighted guesses for the biases (quadratic functions of z, see the cover image): Q1(z) = T*H/i(T*H) Q2(z) = T*T/i(T*T) VAE approximates those answers. I make linear…from linkedin.com

Does a similar job

all alternatives →
  • Efficient-VDVAE an Open-source memory-efficient deep hierarchical VAE2022 · github.com · ▲39

    Official Pytorch and JAX implementation of "Efficient-VDVAE: Less is more" - Rayhane-mamah/Efficient-VDVAE

  • Mini-AGI – Dynamic continual learning model trained on 8GB VRAMSep 2026 · github.com · ▲277

    Continual learning model trained from scratch on 8GB VRAM laptop with batch-1 stream of data. - volotat/mini-AGI

  • Llama 3.2 Interpretability with Sparse Autoencoders2024 · github.com · ▲579

    I spent a lot of time and money on this rather big side project of mine that attempts to replicate the mechanistic interpretability research on proprietary LLMs that was quite popular this year and produced great research papers by Anthropic [1], OpenAI [2] and Deepmind [3]. I am quite proud of this project and since I consider myself the target audience for HackerNews did I think that maybe some of you would appreciate this open research replication as well. Happy to answer any questions or face any feedback. Cheers [1]…

  • DeepSeek My User Agent2025 · jasonthorsness.com · ▲360

    I wrote an SVG editor and physics simulation to determine how pieces connected by rigid rings will dangle when hung. I was surprised at how well the simulation actually predicted reality! This was done with a lot of help from Gemini over last week (so half with the upgrade). Main link is an explanatory article. Here's the live app: https://jasonthorsness.github.io/tree-dangler/ And source: https://github.com/jasonthorsness/tree-dangler

  • Vaev – A browser engine built from scratch (It renders google.com)2025 · github.com · ▲233

    We’ve been working on Vaev, a minimal web browser engine built from scratch. It supports HTML/XHTML, the CSS cascade, @page rules for pagination, and print-to-PDF rendering. It even handles calc(), var(), and percentage units—and yes, it renders Google.com (mostly). This is an experimental project focused on learning and exploration. Networking is basic (http:// and file:// only), and grid layouts aren’t supported yet, but we’re making progress fast. We’d love your thoughts and feedback.

  • Boldly go where Gradient Descent has never gone before with DiscoGrad2024 · github.com · ▲232

    Trying to do gradient descent using automatic differentiation over branchy programs? Or to combine them with neural networks for end-to-end training? Then this might be interesting to you. We develped DiscoGrad, a tool for automatic differentiation through C++ programs involving input-dependent control flow (e.g., "if (f(x) < c) { ... }", differentiating wrt. x) and randomness. Our initial motivation was to enable the use of gradient descent with simulations, which often rely heavily on such discrete branching. The latter makes plain autodiff mostly useless, since it can only account for the…

More life & fun this month

the category →

Source: Show HN post ↗

Launched alongside, September 2021

the whole month →