My lesson

 0    30 fiche    heynonynony
Télécharger mP3 Imprimer jouer consultez
 
question réponse
Supervised Learning
commencer à apprendre
Training a model on labeled data to make predictions
Unsupervised Learning
commencer à apprendre
Training a model on unlabeled data to discover patterns
Reinforcement Learning
commencer à apprendre
Training a model to take actions for maximum reward
Supervised vs Unsupervised Learning
unsupervised uses unlabeled data
commencer à apprendre
Supervised uses labeled data
Supervised vs Reinforcement Learning
reinforcement uses trial and error
commencer à apprendre
Supervised uses labeled data
Unsupervised vs Reinforcement Learning
reinforcement uses trial and error
commencer à apprendre
Unsupervised uses unlabeled data
Types of Machine Learning
unsupervised
commencer à apprendre
Supervised
reinforcement
"Ground Truth"
commencer à apprendre
Correct output for a given input in machine learning
Model Performance Evaluation
commencer à apprendre
Comparing predictions to ground truth
Supervised Learning Applications
natural language processing
commencer à apprendre
Image classification
Unsupervised Learning Applications
anomaly detection
commencer à apprendre
Clustering
Reinforcement Learning Applications
robot control
commencer à apprendre
Game playing
Input Data Structure
unsupervised has unlabeled data
commencer à apprendre
Supervised has labeled data
reinforcement has no specific guidance
Labeled Data Requirements
reinforcement does not
commencer à apprendre
Supervised requires labeled data
Unsupervised Learning Limitations
may require human guidance
commencer à apprendre
May not discover all patterns
Reinforcement Learning Limitations
may not be practical for all tasks
commencer à apprendre
May require a lot of trial and error
Machine Learning Suitability
commencer à apprendre
Depends on problem and data available
Output Type Impact
commencer à apprendre
Categorical or numerical output may affect choice of approach
Input Data Impact
commencer à apprendre
Structured or unstructured data may affect choice of approach
Labeled Data Requirements
unsupervised does not
commencer à apprendre
Supervised requires labeled data
Clear Objective Impact
commencer à apprendre
Having a clear objective may affect choice of approach
Human Supervision Impact
commencer à apprendre
Amount of required supervision may affect choice of approach
Model Provided with Correct Output
commencer à apprendre
Supervised learning
Model Not Provided with Specific Instructions
commencer à apprendre
Unsupervised learning
Model Learns through Trial and Error
commencer à apprendre
Reinforcement learning
Supervised Learning Goal
commencer à apprendre
Make predictions based on labeled data
Unsupervised Learning Goal
commencer à apprendre
Discover patterns in unlabeled data
Reinforcement Learning Goal
commencer à apprendre
Maximize reward through trial and error
Machine Learning to Group Similar Data
commencer à apprendre
Unsupervised learning
Machine Learning to Optimize Performance Over Time
commencer à apprendre
Reinforcement learning

Vous devez vous connecter pour poster un commentaire.