GenAI Bedrock Flashcards
(154 cards)
1
Q
Knowledge Bases for Amazon Bedrock
A
Gives FMs and agents contextual information from your company’s private data sources for RAG to deliver more relevant; accurate; and customized responses
2
Q
Amazon Titan Text Express
A
A Foundation Model (FM) offered by Amazon on Amazon Bedrock
3
Q
Transformer models
A
Use a self-attention mechanism and implement contextual embeddings
4
Q
Amazon Bedrock
A
Fully managed service that makes foundation models from Amazon and leading AI startups available through an API
5
Q
Amazon SageMaker JumpStart
A
Machine learning hub with foundation models; built-in algorithms; and prebuilt ML solutions that you can deploy with just a few clicks
6
Q
Amazon Titan
A
Foundation models developed by AWS that are pre-trained on extensive datasets; suitable for a wide range of applications including text and image generation
7
Q
Foundation Models
A
Use self-supervised learning to create labels from input data
8
Q
Fine-tuning
A
A customization method for FMs that involves further training and does change the weights of your model
9
Q
Amazon SageMaker Ground Truth
A
Helps build high-quality training datasets for machine learning models
10
Q
Amazon SageMaker Model Dashboard
A
Centralized repository of all models created in your account
11
Q
Amazon Comprehend Medical
A
Detects and returns useful information in unstructured clinical text such as physician’s notes; discharge summaries; test results; and case notes
12
Q
Amazon SageMaker Data Wrangler
A
Reduces the time it takes to aggregate and prepare tabular and image data for ML from weeks to minutes
13
Q
Amazon SageMaker Canvas
A
Gives the ability to use machine learning to generate predictions without needing to write any code
14
Q
Amazon Forecast
A
Fully managed service that uses statistical and machine learning algorithms to deliver highly accurate time-series forecasts
15
Q
Amazon Kendra
A
Highly accurate and easy-to-use enterprise search service that’s powered by machine learning
16
Q
Amazon Textract
A
Machine learning service that automatically extracts text; handwriting; layout elements; and data from scanned documents
17
Q
Amazon Rekognition
A
Cloud-based image and video analysis service that makes it easy to add advanced computer vision capabilities to your applications
18
Q
Amazon Polly
A
Uses deep learning technologies to synthesize natural-sounding human speech
19
Q
Amazon Transcribe
A
Automatic speech recognition service that uses machine learning models to convert audio to text
20
Q
Amazon Translate
A
Text translation service that uses advanced machine learning technologies to provide high-quality translation on demand
21
Q
Amazon Lex
A
Fully managed artificial intelligence service with advanced natural language models to design; build; test; and deploy conversational interfaces in applications
22
Q
Amazon Connect
A
AI-powered cloud contact center that automatically detects customer issues and provides agents with contextual customer information
23
Q
Amazon Personalize
A
Fully managed machine learning service that uses your data to generate product and content recommendations for your users
24
Q
Amazon SageMaker Feature Store
A
Fully managed; purpose-built repository to store; share; and manage features for machine learning models
25
Amazon SageMaker Clarify
Helps identify potential bias during data preparation without writing code
26
Amazon Augmented AI (A2I)
Service that makes it easy to build the workflows required for human review of ML predictions
27
AWS DeepRacer
Autonomous 1/18th scale race car designed to test RL models by racing on a physical track
28
Reinforcement Learning
Machine learning technique where an agent learns to make decisions through interactions with an environment; receiving feedback in the form of rewards or penalties
29
Supervised learning
Involves training models with labeled data to make predictions or classify data
30
Unsupervised learning
Identifies patterns and relationships in unlabeled data
31
Semi-supervised learning
Applies both supervised and unsupervised learning techniques to a common problem
32
Deep learning
Subset of machine learning that uses neural networks with many layers to learn from large amounts of data
33
Convolutional Neural Networks (CNNs)
Type of deep learning model particularly well-suited for processing grid-like data; such as images
34
Recurrent Neural Networks (RNNs)
Designed to handle sequential data; where the order of the data points matters
35
K-Means
Unsupervised learning algorithm used for clustering data points into groups
36
K-Nearest Neighbors (KNN)
Supervised learning algorithm used for classifying data points based on their proximity to labeled examples
37
Overfitting
Occurs when a model is overly complex and captures noise or random fluctuations in the training data rather than the underlying patterns
38
Underfitting
Occurs when a model is too simple to capture the underlying patterns in the data
39
Bias
Error introduced by approximating a real-world problem with a simpler model
40
Variance
Error introduced by the model's sensitivity to small fluctuations in the training data
41
Feature extraction
Reduces the number of features by transforming data into a new space
42
Feature selection
Reduces the number of features by selecting the most relevant ones from the existing features
43
Precision
Measures the accuracy of the positive predictions
44
Recall
Measures the ability of the classifier to identify all positive instances
45
F1-Score
Harmonic mean of Precision and Recall
46
Amazon Q Developer
Generative AI-powered assistant that can help you understand; build; extend; and operate AWS applications
47
Amazon Q Business
Fully managed; generative-AI powered assistant that you can configure to answer questions; provide summaries; generate content; and complete tasks based on your enterprise data
48
Amazon Q in QuickSight
Generative BI assistant that allows business analysts to use natural language to build BI dashboards in minutes and easily create visualizations and complex calculations
49
Amazon Q in Connect
Uses real-time conversation with the customer along with relevant company content to automatically recommend what to say or what actions an agent should take to better assist customers
50
Large Language Models (LLMs)
Used for generating human-like text; translating languages; summarizing text; and answering questions based on large datasets
51
Generative AI
Type of AI that can create new content and ideas; including conversations; stories; images; videos; and music
52
Diffusion models
Create new data by iteratively making controlled random changes to an initial data sample
53
Generative Adversarial Networks (GANs)
Work by training two neural networks in a competitive manner
54
Variational autoencoders (VAEs)
Learn a compact representation of data called latent space
55
Prompt engineering
Practice of carefully designing prompts to efficiently tap into the capabilities of FMs
56
Zero-shot Prompting
Technique used in generative AI where the model is asked to perform a task or generate content without having seen any examples of that specific task during training
57
Few-shot Prompting
Technique where you provide a few examples of a task to the model to guide its output
58
Chain-of-thought prompting
Technique that breaks down a complex question into smaller; logical parts that mimic a train of thought
59
Negative prompting
Technique used to guide a generative AI model to avoid certain outputs or behaviors when generating content
60
Retrieval Augmented Generation (RAG)
Process of optimizing the output of a large language model by referencing an authoritative knowledge base outside of its training data sources before generating a response
61
Agents for Amazon Bedrock
Fully managed capabilities that make it easier for developers to create generative AI-based applications that can complete complex tasks for a wide range of use cases
62
Guardrails for Amazon Bedrock
Help you implement safeguards for your generative AI applications based on your use cases and responsible AI policies
63
Watermark detection for Amazon Bedrock
Allows you to identify images generated by Amazon Titan Image Generator
64
Continued pre-training
Process where you provide unlabeled data to pre-train a foundation model by familiarizing it with certain types of inputs
65
Tokenization
Process of converting raw text into a sequence of tokens
66
Embeddings
Process of condensing information by transforming input into a vector of numerical values
67
Context window
Number of tokens that an LLM can consider when generating text
68
Hallucination
When AI models make something up that may sound plausible and factual but which may not be correct
69
Toxicity
AI model-generated content that can be deemed as offensive; disturbing; or inappropriate
70
Exposure
Risk of exposing sensitive or confidential information to a model during training or inference
71
Prompt injection
Influencing the outputs by embedding specific instructions within the prompts themselves
72
Hijacking
Manipulating an AI system to serve malicious purposes or to misbehave in unintended ways
73
Jailbreaking
Bypassing the built-in restrictions and safety measures of AI systems to unlock restricted functionalities or generate prohibited content
74
Reinforcement Learning from Human Feedback (RLHF)
Machine learning technique that uses human feedback to optimize ML models to self-learn more efficiently
75
Model transparency
Understanding the internal mechanisms of a machine learning model
76
Model interpretability
Providing understandable reasons for the model's predictions and behaviors to stakeholders
77
Model customization
Process of using training data to adjust the model parameter values in a base model to create a custom
78
Knowledge Bases for Amazon Bedrock
Gives FMs and agents contextual information from your company's private data sources for RAG to deliver more relevant; accurate; and customized responses
79
Amazon Titan Text Express
A Foundation Model (FM) offered by Amazon on Amazon Bedrock
80
Transformer models
Use a self-attention mechanism and implement contextual embeddings
81
Amazon Bedrock
Fully managed service that makes foundation models from Amazon and leading AI startups available through an API
82
Amazon SageMaker JumpStart
Machine learning hub with foundation models; built-in algorithms; and prebuilt ML solutions that you can deploy with just a few clicks
83
Amazon Titan
Foundation models developed by AWS that are pre-trained on extensive datasets; suitable for a wide range of applications including text and image generation
84
Foundation Models
Use self-supervised learning to create labels from input data
85
Fine-tuning
A customization method for FMs that involves further training and does change the weights of your model
86
Amazon SageMaker Ground Truth
Helps build high-quality training datasets for machine learning models
87
Amazon SageMaker Model Dashboard
Centralized repository of all models created in your account
88
Amazon Comprehend Medical
Detects and returns useful information in unstructured clinical text such as physician's notes; discharge summaries; test results; and case notes
89
Amazon SageMaker Data Wrangler
Reduces the time it takes to aggregate and prepare tabular and image data for ML from weeks to minutes
90
Amazon SageMaker Canvas
Gives the ability to use machine learning to generate predictions without needing to write any code
91
Amazon Forecast
Fully managed service that uses statistical and machine learning algorithms to deliver highly accurate time-series forecasts
92
Amazon Kendra
Highly accurate and easy-to-use enterprise search service that's powered by machine learning
93
Amazon Textract
Machine learning service that automatically extracts text; handwriting; layout elements; and data from scanned documents
94
Amazon Rekognition
Cloud-based image and video analysis service that makes it easy to add advanced computer vision capabilities to your applications
95
Amazon Polly
Uses deep learning technologies to synthesize natural-sounding human speech
96
Amazon Transcribe
Automatic speech recognition service that uses machine learning models to convert audio to text
97
Amazon Translate
Text translation service that uses advanced machine learning technologies to provide high-quality translation on demand
98
Amazon Lex
Fully managed artificial intelligence service with advanced natural language models to design; build; test; and deploy conversational interfaces in applications
99
Amazon Connect
AI-powered cloud contact center that automatically detects customer issues and provides agents with contextual customer information
100
Amazon Personalize
Fully managed machine learning service that uses your data to generate product and content recommendations for your users
101
Amazon SageMaker Feature Store
Fully managed; purpose-built repository to store; share; and manage features for machine learning models
102
Amazon SageMaker Clarify
Helps identify potential bias during data preparation without writing code
103
Amazon Augmented AI (A2I)
Service that makes it easy to build the workflows required for human review of ML predictions
104
AWS DeepRacer
Autonomous 1/18th scale race car designed to test RL models by racing on a physical track
105
Reinforcement Learning
Machine learning technique where an agent learns to make decisions through interactions with an environment; receiving feedback in the form of rewards or penalties
106
Supervised learning
Involves training models with labeled data to make predictions or classify data
107
Unsupervised learning
Identifies patterns and relationships in unlabeled data
108
Semi-supervised learning
Applies both supervised and unsupervised learning techniques to a common problem
109
Deep learning
Subset of machine learning that uses neural networks with many layers to learn from large amounts of data
110
Convolutional Neural Networks (CNNs)
Type of deep learning model particularly well-suited for processing grid-like data; such as images
111
Recurrent Neural Networks (RNNs)
Designed to handle sequential data; where the order of the data points matters
112
K-Means
Unsupervised learning algorithm used for clustering data points into groups
113
K-Nearest Neighbors (KNN)
Supervised learning algorithm used for classifying data points based on their proximity to labeled examples
114
Overfitting
Occurs when a model is overly complex and captures noise or random fluctuations in the training data rather than the underlying patterns
115
Underfitting
Occurs when a model is too simple to capture the underlying patterns in the data
116
Bias
Error introduced by approximating a real-world problem with a simpler model
117
Variance
Error introduced by the model's sensitivity to small fluctuations in the training data
118
Feature extraction
Reduces the number of features by transforming data into a new space
119
Feature selection
Reduces the number of features by selecting the most relevant ones from the existing features
120
Precision
Measures the accuracy of the positive predictions
121
Recall
Measures the ability of the classifier to identify all positive instances
122
F1-Score
Harmonic mean of Precision and Recall
123
Amazon Q Developer
Generative AI-powered assistant that can help you understand; build; extend; and operate AWS applications
124
Amazon Q Business
Fully managed; generative-AI powered assistant that you can configure to answer questions; provide summaries; generate content; and complete tasks based on your enterprise data
125
Amazon Q in QuickSight
Generative BI assistant that allows business analysts to use natural language to build BI dashboards in minutes and easily create visualizations and complex calculations
126
Amazon Q in Connect
Uses real-time conversation with the customer along with relevant company content to automatically recommend what to say or what actions an agent should take to better assist customers
127
Large Language Models (LLMs)
Used for generating human-like text; translating languages; summarizing text; and answering questions based on large datasets
128
Generative AI
Type of AI that can create new content and ideas; including conversations; stories; images; videos; and music
129
Diffusion models
Create new data by iteratively making controlled random changes to an initial data sample
130
Generative Adversarial Networks (GANs)
Work by training two neural networks in a competitive manner
131
Variational autoencoders (VAEs)
Learn a compact representation of data called latent space
132
Prompt engineering
Practice of carefully designing prompts to efficiently tap into the capabilities of FMs
133
Zero-shot Prompting
Technique used in generative AI where the model is asked to perform a task or generate content without having seen any examples of that specific task during training
134
Few-shot Prompting
Technique where you provide a few examples of a task to the model to guide its output
135
Chain-of-thought prompting
Technique that breaks down a complex question into smaller; logical parts that mimic a train of thought
136
Negative prompting
Technique used to guide a generative AI model to avoid certain outputs or behaviors when generating content
137
Retrieval Augmented Generation (RAG)
Process of optimizing the output of a large language model by referencing an authoritative knowledge base outside of its training data sources before generating a response
138
Agents for Amazon Bedrock
Fully managed capabilities that make it easier for developers to create generative AI-based applications that can complete complex tasks for a wide range of use cases
139
Guardrails for Amazon Bedrock
Help you implement safeguards for your generative AI applications based on your use cases and responsible AI policies
140
Watermark detection for Amazon Bedrock
Allows you to identify images generated by Amazon Titan Image Generator
141
Continued pre-training
Process where you provide unlabeled data to pre-train a foundation model by familiarizing it with certain types of inputs
142
Tokenization
Process of converting raw text into a sequence of tokens
143
Embeddings
Process of condensing information by transforming input into a vector of numerical values
144
Context window
Number of tokens that an LLM can consider when generating text
145
Hallucination
When AI models make something up that may sound plausible and factual but which may not be correct
146
Toxicity
AI model-generated content that can be deemed as offensive; disturbing; or inappropriate
147
Exposure
Risk of exposing sensitive or confidential information to a model during training or inference
148
Prompt injection
Influencing the outputs by embedding specific instructions within the prompts themselves
149
Hijacking
Manipulating an AI system to serve malicious purposes or to misbehave in unintended ways
150
Jailbreaking
Bypassing the built-in restrictions and safety measures of AI systems to unlock restricted functionalities or generate prohibited content
151
Reinforcement Learning from Human Feedback (RLHF)
Machine learning technique that uses human feedback to optimize ML models to self-learn more efficiently
152
Model transparency
Understanding the internal mechanisms of a machine learning model
153
Model interpretability
Providing understandable reasons for the model's predictions and behaviors to stakeholders
154
Model customization
Process of using training data to adjust the model parameter values in a base model to create a custom
