AWS Certified AI Practitioner Flashcards¶
74 cards from this guide's fact sheet and notes. Click a question to reveal the answer.
To study with spaced repetition, download flashcards.csv and import it into Anki or Quizlet (comma separated, first row is the header). Back to the study guide.
Exam logistics¶
How long is the AWS Certified AI Practitioner exam?
90 minutes
How many questions are on the AWS Certified AI Practitioner exam?
65 questions
What is the passing score for AWS Certified AI Practitioner?
700/1000 (scaled scoring, approximately 70%)
What does the AWS Certified AI Practitioner exam cost?
$150 USD (50% discount for previous AWS certification holders)
How long is AWS Certified AI Practitioner valid for?
3 years
How is the AWS Certified AI Practitioner exam delivered?
Pearson VUE (online proctored or testing center)
Ai ml fundamentals¶
Accuracy (ai ml fundamentals)
(TP + TN) / Total
Precision (ai ml fundamentals)
TP / (TP + FP) - Of predicted positives, how many are correct?
Recall (Sensitivity) (ai ml fundamentals)
TP / (TP + FN) - Of actual positives, how many did we find?
F1 Score (ai ml fundamentals)
2 × (Precision × Recall) / (Precision + Recall)
MAE (Mean Absolute Error) (ai ml fundamentals)
Average absolute difference
MSE (Mean Squared Error) (ai ml fundamentals)
Average squared difference
RMSE (Root Mean Squared Error) (ai ml fundamentals)
Square root of MSE
R² (R-squared) (ai ml fundamentals)
Proportion of variance explained
Sigmoid (ai ml fundamentals)
1/(1+e^-x) - Binary classification
Tanh (ai ml fundamentals)
Hyperbolic tangent - Range [-1, 1]
Softmax (ai ml fundamentals)
Multi-class probabilities
Gradient Descent (ai ml fundamentals)
Move toward minimum loss
Learning Rate (ai ml fundamentals)
Step size for updates
Batch Size (ai ml fundamentals)
Samples per update
Epochs (ai ml fundamentals)
Complete passes through data
Text (ai ml fundamentals)
GPT-4, Claude, PaLM, LLaMA
Vision (ai ml fundamentals)
CLIP, DALL-E, Stable Diffusion
Multimodal (ai ml fundamentals)
GPT-4V, Gemini, Claude 3
Code (ai ml fundamentals)
GitHub Copilot, CodeWhisperer, Codex
Amazon Rekognition (ai ml fundamentals)
Computer vision
Amazon Textract (ai ml fundamentals)
Document analysis
Amazon Translate (ai ml fundamentals)
Language translation
Amazon Lex (ai ml fundamentals)
Conversational AI
Amazon Kendra (ai ml fundamentals)
Intelligent search
SageMaker JumpStart (ai ml fundamentals)
Pre-built models
Amazon Bedrock (ai ml fundamentals)
Foundation models
Amazon CodeWhisperer (ai ml fundamentals)
Code generation
Bedrock¶
Choice of foundation models (bedrock)
Access models from multiple providers
Serverless experience (bedrock)
No infrastructure to manage
Integrated security (bedrock)
Built-in privacy and security features
Fine-tuning capabilities (bedrock)
Customize models with your data
Responsible AI (bedrock)
Built-in safeguards and monitoring
Text generation (bedrock)
High-quality content creation
Code generation (bedrock)
Programming assistance and debugging
Analysis (bedrock)
Document and data analysis
Conversation (bedrock)
Natural dialogue and Q&A
Reasoning (bedrock)
Complex problem-solving and logic
Text completion (bedrock)
Continuing and completing text
Summarization (bedrock)
Document and content summarization
Text classification (bedrock)
Content categorization
Enterprise focus (bedrock)
Business-oriented language tasks
Multilingual (bedrock)
Support for multiple languages
Embeddings (bedrock)
High-quality text representations
Classification (bedrock)
Text categorization and routing
Open source foundation (bedrock)
Transparent and customizable
Instruction following (bedrock)
Task-specific completions
Text-to-image (bedrock)
Create images from descriptions
Style control (bedrock)
Artistic style customization
High resolution (bedrock)
Detailed image generation
Inpainting (bedrock)
Image editing and completion
Safety (bedrock)
Built-in content filtering
Customization (bedrock)
Fine-tuning capabilities
Cost-effective (bedrock)
Optimized for price-performance
Domain-specific language (bedrock)
Industry or technical terminology
Specific writing style (bedrock)
Brand voice or tone requirements
Task specialization (bedrock)
Optimize for specific use cases
Performance improvement (bedrock)
Better accuracy on your data
Predictable workloads (bedrock)
Consistent traffic patterns
Performance requirements (bedrock)
Guaranteed response times
Cost optimization (bedrock)
Lower per-token costs at scale
Production applications (bedrock)
Critical business applications
Data sources (bedrock)
S3 buckets with documents
Vector database (bedrock)
Amazon OpenSearch Serverless
Retrieval (bedrock)
Semantic search and ranking
No model training (bedrock)
Your data is not used to train foundation models
Encryption (bedrock)
Data encrypted in transit and at rest
VPC endpoints (bedrock)
Private connectivity option
Access control (bedrock)
IAM-based permissions