Applied AICollège Unica
4.2SpecializationSession 4 · Applied AI Systems & MLOps

Unsupervised learning algorithms

20classes60htotalTH2 · PR2 · PNW2weighting per class

Class-by-class breakdown

20 classes

TH = theory · PR = practical · PNW = personal work. These are ministry weighting codes (not hours) used to split each class's minutes. Class length = course hours ÷ class count; the three time-boxes below always sum to that length.

0%

Quiz progress

0 of 0 classes attempted

First 2 classes free with a referral 19 more with enrollment180 min per classTheory 60mPractical 60mPersonal Work 60m
1

Beyond Labels — Unsupervised Learning for GenAI

FreeTH2PR2PNW2

A Data/Business Analyst who clusters support tickets without labels uncovers themes no one mapped — unsupervised learning is how you make sense of unstructured GenAI data.

Theory60m
  • Supervised vs unsupervised in a GenAI context
  • Why unlabelled text and embeddings are the norm in agent work
  • Three goals: cluster, reduce dimensions, detect anomalies
Practical60m
  • Classify 8 sample GenAI tasks as supervised or unsupervised
  • Pick one unsupervised task and describe the hoped-for insight
Personal Work60m
  • Write a one-page note on a business problem unsupervised learning could solve for an agent
  • Micro-task: submit your note
2

Similarity and Distance — How Agents Measure 'Alike'

LockedTH2PR2PNW2

Locked class preview

An AI Integration Manager who understands distance metrics can explain why an agent retrieved one document over another — distance is the foundation of retrieval

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
3

Feature Scaling for Embeddings and Agent Data

LockedTH2PR2PNW2

Locked class preview

A Data/Business Analyst who skips scaling when combining embeddings with other features gets meaningless clusters — scaling is the unglamorous step that fixes it

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
4

Clustering Embeddings — Grouping Similar Documents

LockedTH2PR2PNW2

Locked class preview

A Digital Marketing Strategist who clusters customer feedback embeddings finds the five real themes across thousands of messages — k-means on embeddings turns noise into segments

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
5

Choosing K for Business Clusters

LockedTH2PR2PNW2

Locked class preview

A Sales and Marketing Manager who picks k by business logic, not just a score, gets segments that are actually actionable for campaigns

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
6

Turning Clusters into Personas — Interpreting GenAI Segments

LockedTH2PR2PNW2

Locked class preview

An E-commerce Director who turns raw embedding clusters into named personas gives the marketing team something they can actually target

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
7

Limits of K-Means on Embeddings

LockedTH2PR2PNW2

Locked class preview

A Data/Business Analyst who knows k-means' limits avoids forcing round clusters onto irregular GenAI data — wrong assumptions mean wrong insights

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
8

Hierarchical Clustering — Topic Trees for Agent Knowledge

LockedTH2PR2PNW2

Locked class preview

An AI Integration Manager who builds a hierarchy of topics gives an agent a structured knowledge map it can drill into — hierarchy beats flat groups for navigation

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
9

DBSCAN — Density Clustering for Noisy Text

LockedTH2PR2PNW2

Locked class preview

A Data/Business Analyst dealing with messy feedback uses DBSCAN to find clusters of any shape and flag the noise as outliers

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
10

Evaluating Clusters Without Labels

LockedTH2PR2PNW2

Locked class preview

A Project Manager who demands a quality metric trusts clustering results — even without ground truth, you can measure cluster cohesion

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
11

The Curse of Dimensionality for Embeddings

LockedTH2PR2PNW2

Locked class preview

A Data/Business Analyst who dumps every embedding dimension into k-means often gets worse clusters — high dimensions make distance meaningless

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
12

PCA on Embeddings — Reducing Dimensions for Retrieval

LockedTH2PR2PNW2

Locked class preview

An AI Integration Manager who reduces 768-dim embeddings to 50 for retrieval cuts cost and speeds search without losing the signal

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
13

Choosing How Many Components to Keep

LockedTH2PR2PNW2

Locked class preview

A Data/Business Analyst who keeps enough variance (say 95%) reduces noise without losing the signal — that's the art of dimensionality reduction

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
14

Visualizing Embedding Spaces — PCA and t-SNE

LockedTH2PR2PNW2

Locked class preview

A Digital Marketing Strategist who visualizes customer-feedback embeddings in 2D spots themes a spreadsheet could never show

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
15

Hallucination and Anomaly Detection — Catching Bad GenAI Output

LockedTH2PR2PNW2

Locked class preview

An E-commerce Director who flags an agent's off-brand or hallucinated response automatically protects the brand — anomaly detection on embeddings spots the unusual

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
16

Statistical Anomaly Detection for Agent Outputs

LockedTH2PR2PNW2

Locked class preview

A Data/Business Analyst who flags agent responses whose embedding is far from the norm catches odd outputs in seconds — simple stats, big payoff

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
17

Model-Based Anomaly Detection — Isolation Forests

LockedTH2PR2PNW2

Locked class preview

An AI Integration Manager who needs scalable hallucination detection uses Isolation Forest on embeddings — it handles many features and large logs

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
18

Mini-Project — Cluster and Monitor Agent Data

LockedTH2PR2PNW2

Locked class preview

A Sales and Marketing Manager who clusters and monitors real agent outputs delivers trustworthy insights, not just theory

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
19

Customer Segmentation With Embeddings — A Real Use Case

LockedTH2PR2PNW2

Locked class preview

A Digital Marketing Strategist who segments real customer feedback with embeddings tailors offers to each theme — this is unsupervised learning's most common GenAI win

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions
20

Present and Reflect — Insights Without Labels for GenAI

LockedTH2PR2PNW2

Locked class preview

A Data/Business Analyst who presents unsupervised insights about an agent earns trust by showing patterns, not just the math — this class closes the loop

Enroll to unlock every class in this course — the full agenda, the AI tutor, and the quiz.

Talk to admissions

Unlock all 20 classes

Get the full agenda, AI tutor, and quiz for every class in 4.2 · Unsupervised learning algorithms.

Apply via admissions

Online checkout not yet live — contact admissions to enroll.

Previous course

4.1 · Script programming and task automation

Next course

4.3 · Cloud computing

Back to Session 4