We are a consulting R&D partner with world class expertise in machine and deep learning providing solutions in healthcare and biosciences
@forecomai.bsky.social
@UniOfSurrey@PeopleCentredAI We introduce TRex — a two-stage framework that adapts static SSL features to biologically meaningful tasks like MoA classification. It doubles mAP (0.15 → 0.32), improves compound recognition and generalisation — without retraining large models.
Why do pre-o3 LLMs struggle with generalization tasks like @arcprize? It's not what you might think.
OpenAI o3 shattered the ARC-AGI benchmark. But the hardest puzzles didn’t stump it because of reasoning, and this has implications for the benchmark as a whole.
Analysis below🧵
.@PeopleCentredAI unveils a cutting-edge #AI system that can accurately and quickly identify protein patterns in cells. The new system has the potential to speed progress in #cancerresearch https://t.co/A419lqCXnp #ProudtobeSurrey
Join the State of Precision Medicine webinar, where physicians, scientists, directors, and patient advocates will discuss the current state and future of the field. Speakers include Professor Mathias Uhlén who will discuss related advances of multiomics.
https://t.co/YtVnK2EkJF
Really proud to be published in a Nature Portfolio journal for the first time!
We set a new SOTA for single-cell protein localisation on the @ProteinAtlas, building on our work in the 2nd HPA Kaggle comp.
https://t.co/ETeQoRtYv4 @ForecomAI@cvssp_research@d_minskiy
The Hybrid subCellular Protein Localiser (HCPL) improves single-cell classification for protein localization, allowing for large-scale data annotation through deep-learning architecture. https://t.co/dvTy59ecGb
This technique can rapidly identify the subcellular localisation of proteins of interest in images, streamlining high throughput screening for various applications including the study of tumour heterogeneity, identification of disease and toxicity biomarkers, and drug discovery.
My paper introducing the concept of backdoored NN architectures and their security implications was accepted to @CVPR!
Big thanks to my BA dissertation supervisors @iliaishacked, Aaron, @RobDMullins, and also @NicolasPapernot.
Pre-print: https://t.co/Y6CUTk42Td