Wikiversity:Sandbox

Welcome to the Wikiversity Sandbox. You can experiment here. Click the or tab above to edit and make changes, or click on if you are in the mobile site. Save changes by clicking Save page. Changes here are temporary and occasionally removed. Most changes will still be visible in history. See the "View history" link above. For help with using Wikiversity, see editing help, help desk, or the Colloquium.

If you are a registered user, you might consider to edit on your sandbox.

Please do not add copyrighted, offensive, slanderous, or libelous content. Thanks!


Deep Learning, Computer Vision, and 3D Geometry edit

by --Pinakinathc (discusscontribs) 15:24, 17 March 2024 (UTC)[reply]

Contents edit

  1. Introduction to Deep Learning
    • Gradient Backpropagation
    • Optimisation (e.g., SGD, Adam, AdamW, RMSProp etc.)
    • Reinforcement Learning
  2. Neural Networks Architectures
    • Convolutional neural networks (CNNs).
    • Transformers.
    • Graph neural networks.
    • Graph convolutional neural networks.
  3. Training a neural network (start coding yourself).
    • A simple image classifier using CNN.
    • A simple text-based image retrieval using CNN.
    • Using a deep learning framework like PyTorch.
    • Large-scale training of Foundation Models.
  4. Object Detection
    • Supervised training methods.
    • Weakly supervised training methods.
    • Using large-scale foundation models.
  5. Probability and Information Theory in Deep Learning
    • Variational AutoEncoding (VAEs).
    • Flow-based Models.
    • Diffusion Models.
    • Generative Flow Models.
  6. Basic Concepts in 3D Geometry
    • Camera Parameters (e.g., Intrinsics and Extrinsics)
    • Polar Coordinates
    • Generate 3D objects using Signed Distance Fields (SDFs).
    • Generate 3D objects using Neural Radiance Fields (NeRFs).
    • Generate 3D objects using Gaussian Splatting.