Insights & Ideas
Thoughts on AI, intellectual property, and the future of patent drafting — from the team building it.
How India's generic drug revolution shattered Big Pharma pricing—and the critical IP lessons every innovator must learn.
Privacy is non-negotiable when it comes to patent data. Your inventions are your most valuable intellectual property, and we treat them that way. Here's exactly how PatentAssist.ai secures your data — both through compliance and through engineering.
Unstructured and messy text data often poses challenges while fine-tuning text generation pipeline. Traditional rule-based text cleaning methods often fail as they do not understand the nuances of a language. Therefore, at PatentAssist.AI we are working on using Meta's Llama 3.1 8B model for pre-processing text dataset and preparing it for fine-tuning a text generation pipeline. Using this technique, we have processed and cleaned over 3000 patent drafts to create a training dataset. Here is an article explaining our work.
We read New Guidelines for Examination of Computer-Related Inventions (CRIs), so you don't have to!
The Impact of AI on Patent Law Practice: What Patent Lawyers Need to Prepare For The integration of Artificial Intelligence (AI) into patent law is not just a trend; it represents a fundamental shift in how the profession operates. As AI continues to evolve, patent lawyers need to be aware of how these changes will affect their practices, the skills required, and their interactions with clients. This article explores the various ways AI is reshaping patent law and what lawyers need to prepare for in this new era.
"I want to use AI in doing Prior Art search and Patentability Search." "The use case of semantic based search better than keyword based search for Prior art search?" Intelectual Property stuff is at fore front of AI revolution. However, recent research reveals: current LLM-based legal tools are prone to significant inaccuracies, including “hallucinations” – generating false or misleading information. In a study, researchers assessed leading AI legal research tools, including Lexis+ AI and Westlaw AI based on RAG on performance for Legal use. While Retrieval-Augmented Generation (RAG) shows promise for prior art search, it doesn’t entirely eliminate hallucinations. Read this article which delves into "How one can think of a robust RAG architecture to address the issues of misinterpreting patent claims and confusing inventor descriptions with prior art." Robust Retrieval System for Patent Search: - Finetuning with Domain-Specific Knowledge. - Fact-checking: Combine keyword and semantic search, and prompt LLMs to think through citations. - Chain-of-Verification Prompting: Verify answers using a loop of questions. - Human-in-the-Loop Approach: Involve Experts in critical stages. hashtagAI hashtagIntellectualProperty hashtagLegalTech hashtagRAG hashtagLLM hashtagGenAI hashtagIP hashtagPatent hashtagSearch
Enhancing Patent Draft Quality: Insights from Syngenta Participations vs Controller Of Patents And Designs Gain critical insights from the recent Syngenta Participations AG vs. Controller of Patents & Designs case. Our latest article breaks down essential takeaways for patent agents and attorneys: 🔹 Precision in claim drafting 🔹 Differentiating from prior art 🔹 Importance of comprehensive data 🔹 Strategies for amendments and hearings Learn how specificity and clarity in your patent applications can safeguard your innovations and navigate objections effectively. Read the full article to enhance your patent drafting skills and fortify your intellectual property protections! 📜🔬 Patents IntellectualProperty Innovation PatentLaw LegalInsights ResearchAndDevelopment TechLaw