Unsupervised & Dimensionality ReductionLevel: Core MLStudy Battlecard

t-SNE & UMAP

Non-Linear Manifold Learning, Student-t Kernels & Fuzzy Simplicial Sets

#Manifold Learning#Visualization#Embeddings#t-SNE#UMAP
Choose Presentation Mode:
STAGE 1 / 7— Anti-Pattern
Section 1: The LLM Anti-Pattern vs Right-Sized Model
The Naive Generative LLM Approach:

Asking an LLM to explain why two high-dimensional text embeddings from different topics are clustered together in 2D space.

Why It Fails in Production:

LLMs have no visibility into the gradient optimization paths of non-linear projections and cannot analyze high-dimensional topological distances.

Targeted Algorithm (t-SNE & UMAP)
Latency:150ms (UMAP)
Cost / 1M Ops:$0.00
Determinism:Mathematically Grounded Topology
Generative LLM Alternative
Latency:4,000ms
Cost / 1M Ops:$15,000
Determinism:Conjectural Reasoning