λFPAI

Research Papers

A place to collect paper breakdowns by theme. Each entry will eventually link to a full blog post that explains the paper from first principles.

Optimization

  • SGD and variants

    How simple stochastic gradient descent and its momentum-based cousins actually move on a loss surface.

    (Later this will link to a /blog/[slug] breakdown.)

Architectures (CNNs / Transformers)

  • Attention Is All You Need

    The original transformer paper — sequence modeling with attention instead of recurrence.

    (Later this will link to a /blog/[slug] breakdown.)

Generative Models

  • Diffusion Models

    Turning noise into structure with a forward noising process and a learned denoiser.

    (Later this will link to a /blog/[slug] breakdown.)

RAG & LLM Systems

  • Retrieval-Augmented Generation

    Combining parametric knowledge in an LLM with non-parametric retrieval over external data.

    (Later this will link to a /blog/[slug] breakdown.)