Skip to main content

AI and ML

Updated Sep 21, 2020 ·

Overview​

Artificial Intelligence is defined as the ability of machines to mimic human thinking and decision-making. Machine Learning, a branch of AI, enables machines to learn from data and make predictions or decisions.

  • AI simulates human problem-solving
  • ML improves performance automatically over time
  • Machines adapt without needing explicit programming for every task

AI and ML allow machines to handle complex tasks while improving accuracy over time.

Data Quality in ML​

ML models rely on the data they are trained with. High-quality data ensures better predictions.

  • Data should be clean, relevant, and up-to-date
  • Poor data leads to inaccurate predictions
  • Good data improves reliability and trust in ML models

Quality data is key because ML models reflect the data they learn from.

GCP AI and ML Tools​

GCP offers AI and ML solutions built on a fast and reliable global network. These tools are usable by beginners and experts alike.

  • Pre-built models for easy integration
  • Scalable infrastructure for AI workloads
  • Tools that work with all GCP storage and compute services

Pre-Trained Models​

GCP provides models that are ready to use for images, videos, and text analysis.

  • Vision AI detects objects and reads handwriting
  • Video AI extracts metadata and identifies objects
  • Natural Language AI analyzes text and sentiment

These pre-built models save time and cost, letting businesses add AI capabilities quickly.

ML in BigQuery​

BigQuery allows users to train basic ML models directly in the data warehouse.

  • Use familiar analytics tools to build models
  • Train models on structured data
  • Seamless transition from data analysis to ML

BigQuery makes it easy to turn data insights into predictive models.

BigQuery and Vertex AI​

BigQuery and Vertex AI solve different parts of the ML workflow, and they work well together.

  • BigQuery is best for storing, querying, and preparing analytical data
  • BigQuery ML lets you train simpler models close to the data
  • Vertex AI is better for managing the full ML lifecycle, including training, deployment, and prediction services

Use BigQuery when the goal is analysis or lightweight model building, and use Vertex AI when you need a broader platform for production ML workflows.

Vertex AI​

Vertex AI is a platform for managing, training, and deploying ML models efficiently.

  • Supports both AutoML and custom training
  • Centralizes GCP’s ML services on one platform
  • Streamlines building and deploying models

You can analyze images, translate text, or predict future outcomes. With no-code tools or advanced options, Vertex AI adapts to business needs.

AutoML vs Custom Training​

AutoML automates model building for beginners, while Custom Training offers flexibility for experts.

  • AutoML works on tabular, image, and text data
  • Custom training allows fine-tuning and advanced workflows
  • Users can choose based on expertise and project needs

GCP provides options for both beginners and advanced users to leverage AI.

AI in Business​

GCP AI and ML solutions are applied across industries to create value.

  • Retail: Predict customer churn with BigQuery ML and deploy on Vertex AI
  • Healthcare: Label medical images with AutoML for provisional diagnoses
  • Finance: Build personalized recommendation models using custom training on Vertex AI

Businesses can use AI to improve decisions, automate tasks, and deliver personalized services.