Cognitive Scaffold

Preparing your thinking workspace

arrow_back_ios_new
MENTAL MODEL · M5919

Saddle Point

Saddle Point
TechnicalHigh supportMathematics
Included
account_tree

Version 1.0.0 · Updated 2026-07-30

CORE DEFINITION

In high-dimensional space, a point that is a maximum in one direction (peak) and a minimum in another direction (valley). The shape resembles a saddle. In optimization, algorithms may mistake a saddle point for a terminal point (because the gradient is zero), thus halting progress.

SCAFFOLDING EFFECT

psychology

Reduce cognitive load

A false comfort zone. You might feel you have achieved the extreme in one dimension (e.g., strong technical skills), but in another dimension you are actually at the bottom (e.g., low emotional intelligence). This illusion of a 'local optimum' can lock in your growth; you must introduce perturbations in new dimensions to escape.

anchor

Anchor fast decisions

A saddle point is a stationary point where the gradient is zero but it is neither a local minimum nor a maximum: along some directions it is a peak, along others a valley; in high-dimensional optimization, it is more common than local minima and can easily cause algorithms to stop prematurely.

MINIMUM ACTION

In progress 0/3

Practice this model in one real situation:

Check to track your progress (stored locally)
Learning progress0%
account_treeGenealogyexpand_more
menu_bookReferencesexpand_more

Source support: Explicit

  • link
    en.wikipedia.orghttps://en.wikipedia.org/wiki/Saddle_PointZH · Explicit
    verified

RELATED MODELS

Related model

Vanishing Gradient

Very small backpropagated gradients can slow or stall learning in earlier layers or across long time intervals.

arrow_forwardRelated model

Graham's Number

An enormous finite integer associated with an upper bound for a Ramsey-theory problem on colored connections between hypercube vertices, precisely defined by a 64-term recursive sequence.

arrow_forwardRelated model

Curse of Dimensionality

The curse of dimensionality refers to various phenomena that arise when analyzing and organizing data in high-dimensional spaces that do not occur in low-dimensional settings such as the three-dimensional physical space of everyday experience. The expression was coined by Richard E. Bellman when considering problems in dynamic optimization.

arrow_forward