Generative AI Workshop for Patent Practitioners

E-learning: Introduction to AI & LLM

• How AI, and LLMs (Large Language Model) works
• Confidentiality, efficiency and reliability aspects

Practical exercise to warm up

• Experiments with local and cloud language models
• Optional: Programming* of a neural network (*skills not required)

Basic GenAI

• Prompt Engineering: introduction and practical exercise for common patent tasks
• Hallucinations; improving the

• How to efficiently and securely combine cloud and local AI for patent work

Advanced GenAI: Agentic AI

• Agentic AI: introduction and practical exercise for patent work
• Implementation of the 'Human-in-the-Loop' principle
• Practical exercise: Use cases for agentic AI in patent work

Introduction to Embeddings and RAG

• What is an embedding model
• What is retrieval augmented generation (RAG)
• Use case: Ask your data with RAG
• Practical exercise: How to analyse extensive documents?

Advanced GenAI: Prompting and Coding for Patent Search and Analysis

• Use case: Filtering search results with generative AI
• Use case: Automation of patent searches and infringement detection searches
• Practical exercise: Automation of searches based on invention disclosures

Advanced GenAI: Prompting and Coding for Patent Drafting

• Creating customized DIY solutions, e.g. for drafting in different languages 
• Use case: Streamlining patent drafting
• Practical exercise: LLM-generated patent and description based on human-drafted patent claims using LLMs

Advanced GenAI: Prompting and Coding for Patent Prosecution

• Use case: Streamlining prosecution
• Practical exercise: Office action response generation

Advanced Generative AI: Fine-Tuning

• Introduction to the concept of Fine-Tuning
• Training data preparation

Wrap-up and Q&A Session

• Open discussion: Addressing questions and sharing experiences
• Summary: Key takeaways and how to implement generative AI in day-to-day patent work