Search papers, labs, and topics across Lattice.
University of Bonn, Lamarr Institute for Machine Learning and Artificial Intelligence
2
0
3
Inertial activity recognition models can now learn rich, global motion structure from unlabeled data by discovering and grouping motion primitives, achieving state-of-the-art performance with orders of magnitude less data.
Generative instability in text-to-video models is actually a feature, not a bug: measuring output variance across synthetic generations provides a zero-annotation signal to prune spurious spatio-temporal dimensions from foundation embeddings.
