Search papers, labs, and topics across Lattice.
Affiliation:
7
0
5
11
No single foundation model rules 3D radiology: across 2.1 million volumetric scans and 108 benchmarks, ConvNets consistently outperform transformers at dense spatial localization, while transformers decisively win on global semantic reasoning.
Hierarchy-aware training combined with anatomy-guided learning boosts LUS video classification accuracy while enhancing model interpretability.
Existing continual learning methods falter in preserving knowledge across diverse medical tasks, revealing critical gaps in their application to real-world clinical settings.
ZEBRA bridges the critical base-to-novel generalization gap in audio-language models, boosting novel-class performance without sacrificing base accuracy.
GMM pooling not only enhances preterm birth prediction accuracy but also sets a new standard for image-based classification tasks by effectively modeling intra-patient variability.
Top algorithms in the HECKTOR 2025 challenge achieved impressive segmentation and survival prediction metrics, showcasing the power of multimodal imaging in oncology.
A novel knowledge distillation technique allows a tiny 11M parameter model to not just match, but *beat* a 300M parameter foundation model on fetal ultrasound tasks, opening the door to real-time AI on handheld devices.
