Private custom model (a list of policy numbers, claim numbers, or SSN)
PDFs, Plain Text, MS Word docs
Custom Classification
build custom text classification models using your business-specific labels without learning M
Sentiment Analysis
Subtopic 6
Comprehend Medical
Amazon Lex
(AI) service with advanced natural language models to design, build, test, and deploy conversational interfaces
Uses cases
Self-service voice assistants and chatbots – build a call center bot
Informational bot – build an automated customer support agent or bot that answers questions
Application/Transactional bot – build a stand-alone pizza ordering agent or a travel bot
Enterprise Productivity bot – build custom bots to connect to enterprise data resources
Device Control bot– use Amazon Lex to issue control commands to connected devices
Components
Intent
identify a set of actions
BookTickets
utterance
spoken or typed phrase to invoke an intent
slots
needs information from the user
prompts
action fulfilled
Amazon Polly
Converts Text to Speech
Using DL
Natural Human Speech
Lifelike Speech: talkign Apps
Speech Synthesis Markup Language (SSML) tags like prosody so you can adjust the speech rate, pitch, or volume
W3C standard, XML-based markup language for speech synthesis applications
pronunciation of particular words, such as company names, acronyms, foreign words and neologisms, e.g. “P!nk”, “ROTFL”, “C’est la vie” using custom lexicons.
Integrate Speech into apps
return Audio
send audio stream
Features
Speech Synchronization
Send Audio Stream, Metadata Stream,
Contains info: Sentences, Words, Sounds
Custom Lexicons
Modify pronunciation
Company names, Acronyms, ROTFL Neologisms
Lexical Entries
Speech Synthesis
API, Console, SDK
AWS SDK (Java, Node.js, .NET, PHP, Python, Ruby, Go, and C++) and AWS Mobile SDK (iOS/Android).
Pricing
Standart
You are billed monthly for the number of characters of text that you processed
Free tier
5 million characters per month for speech or Speech Marks requests, for the first 12 months
Amazon Translate
neural machine translation service customizable language translation
supports translation between the following 75 languages:
deep learning techniques to produce more accurate and fluent translation than traditional statistical and rule-based translation models
Features
Named Entity Translation Customization
Language Identification
Batch and Real-Time Translations
Secure Machine Translation
SSL encryption
File Formats
large number of Word documents (docx), PowerPoint presentations (pptx), Excel spreadsheets (xlsx), text, and HTML documents
pricing
Standart
You are billed monthly for the total number of characters
Free tier
500,000 characters per month for 2 months
Amazon Kendra
Amazon Kendra is a highly scalable, intelligent enterprise search service.
It uses machine learning for improved accuracy in search results and the ability to search unstructured data.
Features
Natural Language Processing (NLP): Amazon Kendra uses NLP to get highly accurate answers without the need for machine learning (ML) expertise.
Fine-tuning Search Results: You can fine-tune your search results based on content attributes, freshness, user behavior, and more.
ML-powered Instant Answers: Amazon Kendra delivers ML-powered instant answers, FAQs, and document ranking as a fully managed service.
Use Cases
Enhance Internal Search Experiences for Employees: Improve employee productivity and unlock the insights employees need to make data-driven business decisions through a single search interface.
Improve Customer Interactions: Reduce contact center costs with intuitive self-service bots, agent-assist solutions, and frictionless document access.
Integrate Search into SaaS Applications: Helps you find information faster with ML-powered in-app searches.
File Formats
Amazon Kendra can handle a variety of document types and formats, including PDF, HTML, Word, PowerPoint, and others.
It also supports additional formats like RTF, JSON, Markdown, CSV, MS Excel, XML, and XSLT.
Pricing
Kendra Enterprise Edition (KEE): Provides a high-availability service for production workloads. It costs $1,008 per month.
Kendra Developer Edition (KDE): Provides developers with a lower-cost option ($810 per month) to build a proof-of-concept. This edition is not recommended for production workloads.
Free Tier: You can get started for free with the Amazon Kendra Developer Edition, that provides free usage of up to 750 hours for the first 30 days.
Amazon Forecast
Amazon Forecast is a fully managed service that uses machine learning to deliver highly accurate forecasts. It is designed to harness the same technology that Amazon.com uses for its own forecasting needs.
Features
Data Integration: Amazon Forecast can automatically load historical time-series data from Amazon S3, or you can use the AWS Data Pipeline to import data from other sources.
Automatic Data Preprocessing: The service handles data preprocessing steps like missing value imputation, outlier detection, and variable transformation.
Machine Learning Algorithms: Amazon Forecast employs a range of sophisticated machine learning algorithms, including traditional statistical methods like ARIMA and more advanced deep learning algorithms like DeepAR+. The service automatically selects the best algorithm for your data, or you can specify an algorithm if you have a preference.
Customizable Models: While Amazon Forecast can automatically choose an algorithm, users can also customize models based on their specific needs. This includes the ability to fine-tune parameters and incorporate additional datasets to improve forecast accuracy.
Forecast Dimensions: The service supports complex forecasting scenarios, including item-level forecasting, location-based forecasting, and demand planning for multiple products or services simultaneously.
Evaluation Metrics: Amazon Forecast provides metrics like Weighted Quantile Loss (wQL) to help evaluate the performance of the forecasts.
Scalability and Performance: Given its cloud-based nature, Amazon Forecast can scale according to the size of the data and the complexity of the forecasting task.
Easy Integration: The generated forecasts can be easily integrated into other applications or business processes through AWS SDKs and APIs.
Security and Compliance: Amazon Forecast adheres to AWS’s robust security and compliance standards, ensuring that your data is protected.
Amazon Fraud Detector
Amazon Fraud Detector is a powerful tool for businesses looking to leverage machine learning for fraud detection without needing deep expertise in the field. Its key capabilities include easy model creation, pre-built templates, real-time detection, customizable logic, risk scoring, and seamless integration with AWS services, all underpinned by Amazon's extensive experience in fraud detectio
Features
Machine Learning Models: Amazon Fraud Detector uses machine learning to identify potentially fraudulent activities. It leverages algorithms that have been trained on historical fraud patterns to detect anomalies and risky transactions.
Easy Model Creation: Users can create a fraud detection model with just a few clicks in the AWS Management Console. The service uses your historical data to train and deploy custom fraud detection models.
Pre-built Templates: Amazon Fraud Detector offers pre-built model templates based on common fraud scenarios, like online payment fraud, fake account creation, etc., which can be used to quickly deploy fraud detection models.
Data Integration: The service allows for easy integration of your historical event data for model training. It can handle various types of data, including numerical, categorical, and textual.
Real-time Detection: It provides real-time fraud detection, which is crucial for businesses where immediate action is required, such as e-commerce transactions.
Customizable Detection Logic: Users can create custom rules to fine-tune their fraud detection strategies. This includes setting up specific conditions and thresholds to trigger alerts.
Risk Scoring: Amazon Fraud Detector assigns risk scores to transactions or activities, which helps in determining the level of risk and the appropriate action to take.
Batch Import and Export: In addition to real-time analysis, it supports batch import and export of data for offline analysis and model retraining.
Integration with Other AWS Services: It can be easily integrated with other AWS services like Amazon S3, AWS Lambda, and Amazon SageMaker for extended functionality.
Security and Compliance: As part of AWS, it adheres to strict security standards, ensuring the confidentiality and integrity of your data.
Global Reach: Leveraging AWS infrastructure, it can be deployed globally, enabling businesses to implement consistent fraud detection strategies across different regions.
API Access: Amazon Fraud Detector provides APIs for easy integration with your applications, allowing for automated fraud checks within existing systems.
Amazon SageMaker
A fully managed service that allows data scientists and developers to easily build, train, and deploy machine learning models at scale.
Features
SageMaker AutoPilot – automates the process of building, tuning, and deploying machine learning models based on a tabular dataset (CSV or Parquet). SageMaker Autopilot automatically explores different solutions to find the best model.
SageMaker GroundTruth – a data labeling service that lets you use workforce (human annotators) through your own private annotators, Amazon Mechanical Turk, or third-party services.
SageMaker Data Wrangler – a visual data preparation and cleaning tool that allows data scientists and engineers to easily clean and prepare data for machine learning.
SageMaker Neo – allows you to optimize machine learning models for deployment on edge devices to run faster with no loss in accuracy.
SageMaker Automatic Model Tuning – automates the process of hyperparameter tuning based on the algorithm and hyperparameter ranges you specify. This can result in saving a significant amount of time for data scientists and engineers.
Amazon SageMaker Debugger – provides real-time insights into the training process of machine learning models, enabling rapid iteration. It allows you to monitor and debug training issues, optimize model performance, and improve overall accuracy by analyzing various model-related metrics, such as weights, gradients, and biases.
Managed Spot Training – allows data scientists and engineers to save up to 90% on the cost of training machine learning models by using spare compute capacity.
Distributed Training – allows for splitting the data and distributing the workload across multiple instances, improving speed and performance. It supports various distributed training frameworks such as TensorFlow, PyTorch, and MXNet.
Amazon Bedrock
Amazon Bedrock is a fully managed service that allows you to build and scale generative AI applications. These applications can generate text, images, audio, and synthetic data in response to prompts
Features
Model Choice: Amazon Bedrock provides access to a variety of high-performing foundation models from leading AI companies such as AI21 Labs, Anthropic, Cohere, Meta, Stability AI, and Amazon. You can easily experiment with and evaluate these models for your use case.
Customization: You can privately customize the models with your data using techniques such as fine-tuning and Retrieval Augmented Generation (RAG).
Agents: You can build agents that execute tasks using your enterprise systems and data sources.
Serverless: Since Amazon Bedrock is serverless, you don’t have to manage any infrastructure.
Integration: You can securely integrate and deploy generative AI capabilities into your applications using the AWS services you are already familiar with.
Knowledge base for Amazon Bedrock: The knowledge base for Amazon Bedrock provides the capability of amassing data sources into a repository of information. With knowledge bases, you can easily build an application that takes advantage of retrieval augmented generation (RAG), a technique in which the retrieval of information from data sources augments the generation of model responses.
Text, Image, and Chat playgrounds: Amazon Bedrock provides playgrounds for text, chat, and image models. In these playgrounds, you can experiment with models before deciding to use them in an application
Embeddings: Amazon Bedrock provides text and image embeddings that represent meaningful vector representations of unstructured text, such as documents, paragraphs, and sentences.
Fine-tuning and Continued Pre-training: Amazon Bedrock provides a new capability that allows you to train Amazon Titan Text Express and Amazon Titan Text Lite foundation models and customize them using your own unlabeled data in a secure and managed environment.