Prepare and test your skills

Prepare and test your skills

Free sample question with answers and detailed explanation below. The Cloud Digital Leader practice bank has an additional 3,070 practice questions.Try 10 free realistic questions: instant score, no credit card, and no account required.
A home services company stores thousands of recorded customer phone calls and PDF invoices in Cloud Storage. The business intelligence team wants to extract actionable insights from these unstructured files to meet the following goals:
How should the company use Google Cloud machine learning capabilities to transform this unstructured data into structured information?
A tabular regression model is a supervised machine learning model trained on structured data organized in rows and columns. It analyzes input attributes (called features) from a table to predict a single continuous numerical value (the target column), such as predicting a customer's monthly spending or a home's sale price.
245.50), meaning it cannot transcribe spoken conversations into text or extract text fields like customer names, addresses, and service requests.200 rows multiplied by the number of input features to learn reliable statistical patterns.This solution would be the right choice if the company already had a structured table of historical customer transactions and wanted to predict a continuous numerical outcome, such as how much money a customer will spend on home services next month.
While tabular regression is a core machine learning capability, it operates at a later stage in the analytics lifecycle. Organizations first use task-specific AI services—like Speech-to-Text and Document AI—to convert unstructured files into structured tables, and can then use those extracted tabular features to train predictive regression or classification models.
The Cloud Translation API is a pre-trained machine learning service that translates written text from one language into another (for example, converting Spanish customer comments into English). A search index in BigQuery optimizes the SEARCH function to perform fast, tokenized keyword lookups across unstructured text or semi-structured JSON columns.
ML.TRANSLATE) operates on text data stored in standard tables and cannot convert spoken audio recordings into written transcripts.SEARCH INDEX allows users to find specific words within already-extracted text, but it does not parse raw PDF files or organize invoice entities into relational database columns.ML.TRANSLATE is specifically built for cross-language text translation rather than speech recognition.SEARCH function locates rows containing specific text tokens (such as 'Google') rather than performing entity extraction.This approach would be correct if the company already had written customer reviews stored in a BigQuery table in multiple foreign languages, wanted to translate them into a single familiar language using ML.TRANSLATE, and needed to run fast keyword searches over the translated text using a search index.
In a complete unstructured data pipeline, search indexes and machine translation often complement extraction tools. For example, after Document AI (ML.PROCESS_DOCUMENT) extracts text from PDF files into a BigQuery table, an analyst can create a search index on the resulting table to quickly locate specific terms across thousands of processed documents.
Cloud AI APIs and BigQuery ML provide pre-trained machine learning and generative AI capabilities that unlock business value from unstructured data such as audio recordings, images, and PDF documents. By referencing these services over object tables in BigQuery, organizations can analyze unstructured files stored in Cloud Storage using standard SQL queries and convert raw media into structured relational tables.
ML.TRANSCRIBE function in BigQuery ML) automatically transcribes spoken audio files into written text.AI.GENERATE_TABLE or AI.GENERATE, analyze unstructured call transcripts and format extracted details—like customer names, requests, and pricing quotes—directly into a custom SQL schema.ML.PROCESS_DOCUMENT) uses specialized document processors to parse unstructured invoice files and extract structured entities so they can be joined with existing customer tables.Unlike traditional rule-based programming or tabular-only models, combining Speech-to-Text, Document AI, and BigQuery ML generative AI directly addresses multimodal, unstructured formats. It provides an automated, scalable pipeline that turns raw audio and PDFs into structured database records for immediate analytical querying.
Classical programming relies on human developers writing explicit, step-by-step conditional instructions and manual filters for a computer to follow. In this traditional paradigm, the software only recognizes patterns that a programmer has explicitly anticipated and coded into a sequence of rigid rules.
This approach would be appropriate if the company only needed to apply deterministic validation logic—such as filtering structured database rows where a numeric invoice_total column is greater than $1,000 or checking if a standardized status code equals PAID.
A primary business value of machine learning is its ability to overcome the limitations of classical programming on unstructured data. When a problem involves natural language, audio, or images where rules would expand endlessly, organizations need ML systems that learn patterns from examples rather than following manual step-by-step instructions.
Answer them yourself and get your score the moment you finish. No account needed.
Full access includes
Keep the momentum going with these hand-picked practice scenarios
Last updated