Heor

In a recent paper, Tinglong Dai, Risa Wolf, and Haiyang Yang wrote about unlearning in Medical AI. With more and more CPU and storage thrown at Large Language Models (LLMs) and Generative Artificial Intelligence (GenAI) in general, the capacity to “memorise” information grows larger and larger with each generation of LLM. This is further accelerated by the capacity to add specific details to generalist LLMs using Retrieval Augmented Generation (RAG) and agents (e.g., with the ability to query real-world systems at the interface with the physical world). LLMs are learning more, but what about unlearning? Dai and colleagues didn’t evoke the analogy with human memory: our capacity to learn more relies, in part, on our capacity to forget, to reorganise, to summarise, and to prioritise the learnt information. Sleep and stress play a role in this reorganisation of information; this was the overarching topic of my Ph.D. thesis [link]. I will de-prioritise the visual cues along the path leading to a bakery if I no longer go to this bakery (“unlearning”). However, practising navigation to the bakery improved this skill, and this improvement will serve me later when I need to go to another place (something I could call “secondary learning”). It may seem we diverge from AI, but Dai and colleagues actually start their paper with the EU GDPR possibility for a patient to remove their data from a database, wondering how this is technically possible with LLMs (where data is not structured like in a traditional relational database and where the way data is retrieved is often unknown). The “unlearning” process in LLMs can be considered from three encapsulated levels: algorithm, legal, and ethical levels.

After Modelling and Regulations & Pricing, and just a few days before ISPOR25, here is my take on the potentially interesting sessions on Artificial Intelligence (AI, which generally means: the use of Generative AI, or GenAI, in HEOR). First, Sven Klijn, William Rawlinson, and Tim Reason are again offering their introductory course on Applied Generative AI for HEOR. Last year, I followed it in Barcelona, and it was nice. In my opinion, “nice” means that although I didn’t learn much more than previous presentations by the authors and my own experience, it was a great course for beginners because it struck the right balance between theory (which too many sessions end up only covering) and practical examples. Don’t expect hands-on exercises (that would be too long, and the course synopsis doesn’t mention that either). But “nice” to me means that the presenters dared to show actual working code, with all the humility that it implies. This year, they mention they’ll cover Retrieval-Augmented Generation (RAG) and agents. Hopefully, their coverage of these aspects will be as good as last year’s on the other topics. ...

ISPOR25, the annual North American conference for the International Society for Pharmacoeconomics and Outcomes Research, is in three weeks. As usual, I’m planning for it by browsing its program. This time, I decided to share a few of my interests on my blog. ISPOR usually covers many topics, from “hardcore” statistical methods to top-level overviews of some issues, so I will focus on only a few topics. Feel free to connect with me if you want to discuss anything at or around the conference (or virtually). (And before we start, full disclaimer: I’m currently working for Parexel, but opinions shared here are only mine; otherwise, I would have written them on the company blog.) ...