Good machine learning practices: Learnings from the modern pharmaceutical discovery enterprise.

Area of Science:

• Pharmaceutical industry
• Drug discovery and development
• Artificial Intelligence (AI)
• Machine Learning (ML)

Background:

• AI and ML are crucial in modern drug discovery and development.
• Pharmaceutical teams face challenges in implementing AI/ML solutions effectively.
• There's a need to understand and address these implementation hurdles.

Purpose of the Study:

• To systematically analyze business practices related to AI and ML in the pharmaceutical sector.
• To identify common challenges faced by personnel working with AI/ML technologies.
• To propose best practices for overcoming these identified issues.

Main Methods:

• Conducted an industry-wide assessment of AI and Machine Learning practices.
• Performed a systematic business analysis of personas within pharmaceutical discovery.
• Examined the interaction of these personas with AI and ML technologies.

Main Results:

• Identified 23 common business problems encountered by pharmaceutical professionals using AI/ML.
• Detailed the specific challenges related to the delivery of AI/ML solutions.
• Established a framework of Good Machine Learning Practices.

Conclusions:

• Addressing identified business problems is key to successful AI/ML integration in pharma.
• Good Machine Learning Practices offer a pathway to overcome implementation challenges.
• Optimizing AI/ML adoption can enhance the efficiency of drug discovery and development.