Explore core concepts and practical knowledge on cloud-based machine learning APIs and AI services. Sharpen your understanding of features, workflows, and best practices related to scalable AI solutions in a cloud environment.
Which API function would you use to detect the original language of a text before translating 'Guten Morgen' to English in an application?
Explanation: Language Detection is specifically designed to determine the original language of a given text, such as identifying that 'Guten Morgen' is German. Sentiment Analysis focuses on interpreting the emotional tone rather than the language. Entity Extraction seeks to find named objects in text, while Vision Labeling analyzes images instead of text. Thus, only Language Detection matches the required function for translation workflows.
If you need to filter inappropriate images uploaded by users to your cloud service, which approach using AI APIs would best automate this process?
Explanation: Label Detection in image analysis is capable of identifying objects or content (such as adult material) within uploaded images, supporting automated moderation. Text-to-Speech synthesizes spoken audio from text, which is not relevant to image content. Speech Recognition is for converting audio into text, and Structured data import deals with tabular or form data, not images. Therefore, only Label Detection offers automated image content moderation.
What is a primary benefit of using an automated machine learning (AutoML) service to train a custom model for classifying online product reviews?
Explanation: AutoML services are designed to simplify model development so that users do not need detailed knowledge of machine learning architecture. While they increase accessibility, they do not guarantee perfect accuracy, so the second option is inaccurate. Labeled data is still required to train supervised models, making the third option incorrect. The speed of AutoML varies and is not necessarily faster than all manual methods, so the fourth choice is not accurate.
To automatically detect spoken topics and label scenes within hours of video lectures uploaded by students, which AI service feature is most suitable?
Explanation: Video Intelligence with speech transcription and entity detection can process video files to transcribe spoken content and label elements, making it suitable for analyzing lectures. Document translation is limited to text, not video. Object detection in tabular data refers to analyzing data tables, which does not apply. Text annotation in spreadsheets is unrelated to processing video content. Therefore, the first choice is the only appropriate answer for this use case.
Which API workflow is most appropriate for analyzing customer feedback from thousands of product reviews and summarizing overall sentiment trends?
Explanation: Bulk Sentiment Analysis allows for efficient processing of large volumes of text to assess sentiment, and data aggregation can then summarize overall trends. Image labeling does not apply to text reviews, making the second option incorrect. Option three suggests converting text to audio and using OCR, which is unnecessary since the feedback is already textual. The fourth option relates to image data rather than text analysis. Thus, the most relevant workflow is the first option.