AI Implementation Best Practices and Methodologies - CPMAI

Flashcards covering best practices, methodologies, common mistakes, and lifecycle considerations for successful AI implementation, including CPMAI.

What is a fundamental difference in focus between AI projects and traditional software development?

Traditional software development emphasizes functionality, while AI projects are fundamentally about leveraging data to create systems that learn and and evolve.

How do AI systems, particularly those based on machine learning, improve their performance?

AI systems improve their performance through exposure to data, enhancing their capabilities as they learn from new data inputs.

What is a common mistake made when approaching AI projects with a traditional mindset?

A common mistake is a misplaced focus on functionality, prioritizing feature delivery rather than understanding underlying data requirements.

What are the three key practices for effective data management in AI projects?

Data Collection, Data Cleansing, and Data Preparation.

Name one common reason for failure in AI projects.

Data quantity and quality issues.

Why do AI project lifecycles require continuous iteration?

AI projects require continuous iteration because data and user needs change, necessitating updates and retraining of AI models to maintain effectiveness.

What are the three quantifiable metrics commonly used to assess the Return on Investment (ROI) of AI projects?

Cost Savings, Time Savings, and Resource Efficiency.

What is the first critical question to ask before initiating an AI project?

What problem are we trying to solve?

What caused Amazon's AI Recruiting Tool to display bias against women?

The tool was trained on historical hiring data that lacked sufficient representation of female candidates.

What is the main difference between a Proof of Concept (PoC) and a Pilot Project?

A PoC demonstrates feasibility in a controlled environment, often with idealized data, while a Pilot Project tests a solution under actual operating conditions with real data and users.

What principle emphasizes that poor-quality data will lead to ineffective AI models?

The principle of 'garbage in, garbage out'.

What three common issues are associated with falling for vendor hype in AI projects?

Product mismatch, overhype, and oversell.

What is the recommended approach to avoid overpromising and underdelivering in AI projects?

Think Big, Start Small, Iterate Often.

What is Uncanny Valley awareness and mitigation in the context of user experience for AI?

It is preventing user rejection due to overly intrusive or creepy AI implementations, especially when systems know too much personal information, causing user sentiment to plummet.

How do Agile methodologies enhance AI projects?

Agile methodologies enhance responsiveness and efficiency by promoting iterative development and continuous feedback.

What percentage of an AI project is often attributed to data problems versus application functionality?

80% data problem, 20% application functionality.

What specialized role must a Data Product Owner understand in an Agile AI team?

The complete data lifecycle from collection and ingestion through preparation, transformation, and consumption.

What are two key Agile practices mentioned for AI projects?

Sprints and Daily Standups.

What is a significant limitation of CRISP-DM in the current AI landscape?

It has not been updated since its first release in 1999 and does not account for agile methodologies or more advanced forms of data analytics like AI/ML.

What is CPMAI, and what does it emphasize for AI projects?

CPMAI (Cognitive Project Management for AI) is an updated methodology specifically for AI projects that incorporates Agile principles and emphasizes continuous iteration and AI-specific processes, focusing on operational success driven by reliable, well-understood data.

What is the objective of Phase I (Business Understanding) in CPMAI?

To define the business problem and project goals to ensure alignment with organizational objectives.

What is the objective of Phase III (Data Preparation) in CPMAI?

To prepare the data for modeling by cleaning, transforming, and structuring it appropriately.

What is a key activity in Phase IV (Modeling) of CPMAI?

Hyperparameter Tuning.

What is the objective of Phase VI (Operationalization) in CPMAI?

To deploy the model into a production environment where it can deliver real-world value.

What is a key characteristic of Data Projects that distinguishes them from application development projects?

They focus on what insights or actions need to be derived from the data rather than starting with functionality.

Why can fragmented outputs be a challenge in Agile Data Projects?

Individual data projects may lack cohesion, leading to redundancy and inefficiency if efforts are not aligned with broader organizational goals.