Theory Meets Practice #4: Explainable Scientific Discovery with AI Memory
We are excited to host a BLISS x dida workshop featuring Lazar Obradović (Lead Data Scientist @ cognee.ai), who will guide us through an interactive session on Embedded AI.
Title: Explainable Scientific Discovery with AI Memory: Hypothesis Generation and Validation with Cognee and Bayer
📅 Date: 27.05.2026
🕕 Time: 18:00
📍 Location: TU Berlin Marchstrasse 23 [Room 0.011]
The session will last around 2 hours, followed by a networking session with dida and fellow AI enthusiasts (and free pizza!🍕). Bring your laptop!
Abstract: Building "AI scientist" systems that can reliably generate and evaluate research ideas remains effectively an open problem. While LLMs make hypothesis generation cheap, current systems still struggle to connect evidence across papers, domains, and mechanisms in a structured and reusable way. Evaluation is equally difficult, often relying on opaque “LLM-as-a-judge” methods that are hard to interpret or reproduce. We use Cognee to ingest scientific literature and construct a graph-based representation of papers, claims, and entities with associated vector embeddings. We then use this structure to select and organize the context used for hypothesis generation. Candidate hypotheses are evaluated using explicit graph- and vector-based metrics grounded in the underlying corpus. Initial evaluations by Bayer across targeted biomedical domains demonstrated that graph-based AI memory can reliably support scientific hypothesis generation and validation.
Who is this event for?
This session is targeted at anyone with basic ML knowledge. Experience with Python is very helpful. We will cover the most important aspects as part of the workshop.
We are BLISS e.V., Berlin’s AI community connecting like-minded individuals passionate about machine learning and data science. Our BLISS Workshops connect students and young professionals with industry partners, offering an inside look into how machine learning is applied in real-world settings - from research and development to deployment.
We are dida, scientific engineers who believe that reliable AI should not be a "black box". That is why we prioritize transparency, mathematics, and code over hype. By bridging the gap between theoretical research and production, we develop custom white-box AI solutions that are fully explainable, rigorously engineered, and free of "magic".
dida Website: https://dida.do
dida Youtube: https://www.youtube.com/@dida-do
BLISS Website: https://bliss.berlin
BLISS Youtube: https://www.youtube.com/@bliss.ev.berlin