Nota:
Noten que el tipo de algoritmos que corre auto ML son algorirmos supervisados, porque le decimos la tabla objetivo (los resultados que debe obtener).
Compute on Google Cloud Platform
Introducci贸n al curso de Google Cloud for Developer Community
Lectura: introducci贸n de instalaci贸n
Tutorial de Qwiklabs
C贸mputo en la nube de Google
Opciones de c贸mputo en la nube
M谩quinas virtuales a profundidad
Tutorial para instalar Qwiklabs
Demo: m谩quinas virtuales a profundidad
C贸mputo sin administraci贸n con plataformas como servicio
Demo: c贸mputo sin administraci贸n
Lectura: 驴qu茅 son los contenedores?
C贸mputo contenerizado con App Engine Flex
C贸mputo contenerizado con Cloud Run
Funciones serverless
Continuous Integration, Continuous Delivery
CI/CD en Google Cloud Platform
Estrategias de Despliegue
Repositorios de c贸digo
Construcci贸n y despliegue de artefactos
Infraestructura como c贸digo
Despliegue en Servicios Serverless
Google Kubernetes Engine
Kubernetes Overview
Demo Kubernetes
Planeaci贸n de tu despliegue
Anthos
Cloud Run for Anthos
Demo Cloud Run for Anthos
Anthos Service Mesh
Site Reliability Engineering con Anthos
Streaming Data Analytics
Integraci贸n de datos e ingesta de datos totalmente administrada sobre GCP
Demo: ingesta de datos
Ingesta de datos confiable en streaming sobre GCP
Demo: ingesta de datos confiable
Demo: configuraci贸n de Apache Kafka
Visualizaci贸n de mensajes de una base de datos relacional en Google Cloud
Data Warehouse: el modelo tradicional para construir un repositorio de datos empresarial
Data Lakehouse: el nuevo y moderno enfoque para construir un repositorio de datos empresarial
El portafolio de gesti贸n de datos en Google Cloud
Desglose del portafolio de gesti贸n de datos (Bases de datos) en Google Cloud
Gobierno de datos de punta a punta para garantizar la seguridad en tu Data Lake
Gobierno de datos: calidad y monitoreo
Machine Learning
驴Qu茅 es ML y AI?
Plataforma de AI en GCP
Auto ML con datos estructurados
Demo Auto ML con datos estructurados
Predicci贸n de tarifas usando AI notebooks
Demo predicci贸n de tarifas usando AI notebooks
TensorFlow Extended
Sesiones en vivo
Sesi贸n en vivo con Pablo P茅rez Villanueva
You don't have access to this class
Keep learning! Join and start boosting your career
AutoML, an advanced machine learning technology, allows companies to implement models without the need to be Machine Learning experts. It lies between pre-built solutions and fully customized development. AutoML makes it easy for downstream users to tackle artificial intelligence and machine learning problems quickly, saving them time and technical effort.
The process simplifies complex tasks by defining data schemas, labeling information and using preconfigured models. AutoML solutions, such as AutoML Vision and AutoML Tables, are ideal for common structured data in business scenarios, generating efficient and personalized predictions.
Google Cloud offers AutoML as a solution to extend technical capabilities with pre-built models that can be customized. This is where Cloud AutoML comes in, providing the flexibility to parameterize models according to business specificities and input data without building a model from scratch.
Define data and attribute schemas: This includes classifying input data (floats, dates, strings) and setting labels for supervised learning.
Attribute analysis: Identifying which attributes have the greatest impact on predictions is crucial. Domain experts must determine which features in the dataset are the most relevant.
Feature engineering and selection: AutoML automatically selects models and parameters based on the data provided. Users define the initial parameters and AutoML adjusts the hyperparameters automatically.
Model training and evaluation: It is possible to evaluate the behavior of the algorithm by testing and dynamically adjust the model according to specific business needs.
Deployment: Production-ready, the model can be deployed via APIs, scaled on demand and used in a variety of contexts, from mobile applications to cloud infrastructures.
AutoML Tables focuses on structured data, representing numerous use cases, such as sales forecasting, risk assessment or weather predictions in IoT systems. AutoML enables an optimized learning and deployment process, lending itself to scenarios that do not require intensive computational investment.
In addition to tabular data, AutoML extends its functionality to images, videos and natural language. For example, technologies behind services such as Google Photos and Google Translate, initially developed in-house, are now available for customization in AutoML. AutoML Vision allows training customized visual models, leveraging advanced computer vision algorithms.
Suppose you have Google Analytics data: using AutoML Tables, you can identify buying patterns based on parameters such as session ID, geography or domain. This capability allows you to make accurate predictions and optimize business strategies from well-categorized historical data.
AutoML uses automatic feature engineering to automatically recognize different types of variables, identifying possible imbalances or correlated labels that affect model performance. Its ability to handle missing data or outliers is crucial for generating robust models.
The underlying algorithms, such as linear regressions, simple neural networks or decision trees, are implemented and selected according to the data, resulting in a robust and scalable solution.
In short, AutoML on Google Cloud democratizes access to machine learning, turning complex challenges into accessible solutions for companies without the need for advanced technical expertise. This tool, through its structured and user-friendly approach, achieves cutting-edge results in different application areas.
Contributions 2
Questions 1
Nota:
Noten que el tipo de algoritmos que corre auto ML son algorirmos supervisados, porque le decimos la tabla objetivo (los resultados que debe obtener).
Wow esto si es ML made easy!
Want to see more contributions, questions and answers from the community?