Sign inSign in
ConceptAI & ML

Inductive bias

The set of assumptions that a learning algorithm uses to predict outputs for inputs it has not encountered.
as Irina Rish defines itUniversité de Montréal, MILAThe wording closest to all 9 definitions
Watch it explained

Know first

Named as a prerequisite by the lectures that define this. The number is how many say so. A grey one has no page yet.

Opens up

Concepts that name this one as a prerequisite.

Lecturers

9

Each in their own terms, from the easiest level.

  • Hyung Won Chung

    Stanford University · Undergraduate

    Assumptions and structures built into a model to help it learn faster with less data, which can later act as a ceiling to scaling.
    • 11:36Shaping the Future of AI from the History of Transformer Architectures
  • Jason Wei & Hyung Won Chung

    Stanford University · Undergraduate

    Additional structure or modeling assumptions imposed on a machine learning method to help it perform well under low-compute or low-data regimes.
    • 43:17Intuitions on Language Models and the History of Transformer Architectures
  • Sara Beery

    MIT · Undergraduate

    Assumptions and structural priors built into a learning model about the kinds of functions it should expect to learn.
    • 3:30Architectures: Grids
  • Phillip Isola

    MIT · Undergraduate

    Assumptions built into a learning algorithm that cause it to prefer certain solutions over others, enabling generalization.
    • 59:56Generalization Theory
  • Stefano Ermon

    Stanford University · Graduate

    The set of assumptions a model uses to predict outputs for inputs it has not encountered.
    • 1:11:21Generative Modeling by Estimating Gradients of the Data Distribution
  • Drew Jaegle

    Stanford University · Graduate

    Assumptions built into a learning algorithm that help it generalize to unseen data.
    • 1:17DeepMind's Perceiver and Perceiver IO: New data family architecture
  • Tom Mitchell

    UC Berkeley · Graduate

    The set of assumptions that a machine learning algorithm uses to predict outputs for unseen inputs.
    • 8:57Questions for Theory in the New Age of Machine Learning
  • Lucas Beyer

    Stanford University · Graduate

    Built-in assumptions like locality or translation equivariance that may limit models when scaling up data.
    • 53:49Transformers in Vision: Tackling problems in Computer Vision

Lectures that assume it

2

Connections

18

The lectures that define it or assume it, and the concepts next to it.

9 lecturers. Definitions are written by AI from each lecture’s transcript. A time with a play mark starts the lecture at the moment the lecturer says it.