A company is developing an AI character for a video game. The AI character needs to learn how to navigate a complex environment and make decisions to achieve certain objectives within the game. When the AI takes actions that lead to positive outcomes, like finding a reward or overcoming an obstacle, it receives a positive score. When it takes actions that lead to negative outcomes, like hitting a wall or losing progress, it receives a negative score. Through this process of trial and error, the AI gradually improves the character??s ability to play the game effectively. What machine learning should the company use?

1. A. Reinforcement learning
2. B. Unsupervised learning
3. C. Supervised learning
4. D. Deep learning

Correct Answer: A

A company is defining their generative AI strategy. They want to follow Google- recommended practices to increase their chances of success. Which strategy should they use?

1. A. Rapid implementation strategy
2. B. Bottom-up strategy
3. C. Multi-directional strategy
4. D. Top-down strategy

Correct Answer: D

A large multinational corporation with geographically dispersed teams struggles with knowledge silos and inconsistent access to crucial internal information. What is a key business benefit of using Google Agentspace in this scenario?

1. A. Improved IT infrastructure management across offices.
2. B. Seamless knowledge sharing and collaboration across internal systems.
3. C. Enhanced data encryption and compliance for internal communications.
4. D. Automation of employee performance reviews using AI.

Correct Answer: B

A software developer needs a highly efficient, open-source large language model that can be fine-tuned on a local machine for rapid prototyping of a chatbot application. They require a model that offers strong performance in natural language understanding and generation, while being lightweight enough to run on limited hardware. Which Google-developed family of models should they use?

1. A. Veo
2. B. Gemini
3. C. Gemma
4. D. Imagen

Correct Answer: C

A company??s large learning model (LLM) is producing hallucinations that are a result of the Knowledge cutoff. How does retrieval-augmented generation (RAG) overcome this limitation?

1. A. RAG fine-tunes the LLM on specific customer query patterns to improve the speed andefficiency of response generation.
2. B. RAG enhances the creative writing capabilities of the LLM to generate more engaging and informative responses.
3. C. RAG enables the LLM to retrieve relevant and up-to-date information from knowledge sources.
4. D. RAG uses human oversight to ensure accuracy before presenting information to the customer.

Correct Answer: C