Search papers, labs, and topics across Lattice.
University of British, Columbia University of Copenhagen, Carleton University University of Copenhagen Vector Institute Vector Institute, Taylor Geospatial University of Copenhagen University of British Columbia Vector Institute
6
0
8
Earth observation streams and foundation model embeddings can be compressed 1,800脳 into continuous neural fields while retaining 90%+ downstream performance and cutting data retrieval latency tenfold.
DeluluNet can adapt to new satellite sensors without extensive retraining, maintaining predictive accuracy in a rapidly evolving landscape.
A new benchmark reveals how well machine learning models can map hedgerows across diverse climates and distances, highlighting significant generalization challenges.
Current open-set TTA methods fail to effectively balance in-distribution accuracy with out-of-distribution detection, revealing a critical gap in their robustness.
Ditch unstable final-layer self-distillation: Bootleg predicts latent representations from multiple hidden layers, boosting ImageNet classification by 10% over I-JEPA.
