JWTH: AI that looks at cells and the big picture to find biomarkers

Kari Jaaskelainen

10 Nov 2025 — 1 min read

Most pathology AIs read slides patch by patch, often missing the subtle shapes of individual cells. JWTH (Joint-Weighted Token Hierarchy) changes that by fusing cell-level clues with whole-tissue context in one foundation model.

What’s new: a cell-centric post-tuning step and attention pooling that combine local and global tokens.
Why it matters: infer molecular biomarkers directly from standard H&E slides, making results more interpretable and robust.
How it performed: across four biomarkers and eight patient cohorts, JWTH reached up to 8.3% higher balanced accuracy and a 1.2% average gain over prior pathology foundation models.
Under the hood: large-scale self-supervised pretraining, then joint weighting to merge cell and tissue signals.

This brings AI-assisted biomarker detection closer to real-world pathology workflows.

By Jingsong Liu, Han Li, Nassir Navab, and Peter J. Schüffler. Read more: http://arxiv.org/abs/2511.05150v1

Paper: http://arxiv.org/abs/2511.05150v1

Register: https://www.AiFeta.com

#AI #DigitalPathology #Biomarkers #Cancer #MachineLearning #Healthcare #ExplainableAI

Automating GDPR Compliance: A Roadmap for Companies and Law Firms

GDPR compliance is more than checkboxes. A new roadmap from the Privatech project shows how automation and machine learning can help companies and law firms assess—and even generate—privacy compliance. * Shift the focus to data processors’ real workflows: drafting policies, mapping data uses, documenting decisions. * Break compliance into machine-ready

FPGAs for Faster, Leaner Deep Learning: A Review of CNN Accelerators

Deep learning drives image search, robots, and medical scans. Most systems lean on CPUs and GPUs. This review asks: what if we run convolutional neural networks (CNNs) on FPGAs—reconfigurable chips you can tailor to the model? * Why FPGAs: custom dataflows, low latency, and strong energy efficiency—great for cameras,

Dynamic-K: Recommendations That Know When to Stop

Most apps show a fixed number of “top” items—say 10 movies or 20 products—assuming there are always enough good options. But that’s not always true: sometimes there are few relevant items, or some users are extra picky. The result? Filler recommendations. Dynamic-K flips the script. Instead of

Teaching chatbots to stop contradicting themselves (DECODE)

Teaching chatbots to stop contradicting themselves Ever had a bot say one thing, then the opposite a few turns later? This study introduces DECODE—a new task and dataset for spotting contradictions in everyday conversations, drawn from both human-human and human-bot chats. * New data beats existing natural language inference (NLI)