Predictive video representations do not by themselves specify how a multimodal system should handle missing sensors. We transfer frozen V-JEPA2 features to a set of heterogeneous sensors comprised by RGB video, depth video, and radar time series. A shared fusion model uses modality queries, fixed sensor slots, and expl...
Yun-Xiang Luo, C. A. Casado, Xiao-Ting Wu et al.· 0 citations
Camera-based physiological monitoring can produce multiple estimates from several facial regions, extraction methods, and processing settings. Signal quality indices aim to select reliable estimates without a physiological reference, and their potential is often assessed with an oracle that selects the estimate closest...
M. Rahman, Nhi Nguyen, Sasan Sharifipour et al.· 0 citations
Long-tailed 3D object detection is treated as a class-frequency problem, but LiDAR supervision quality depends on object observability: similar frequencies can hide different geometric evidence. We introduce Geometry-Augmented Exponentially Weighted Instance-Aware Repeat Factor Sampling (GA-EIRFS), a detector-agnostic...
Taufiq Ahmed, C. A. Casado, Daniel Herrera Castro et al.· 0 citations
The study separates the measured benefit of predictive representations from the physiological prior used to organize their training, and presents Phenomenon-Graph JEPA, a joint-embedding predictive architecture that learns from four processed one-dimensional streams without negative pairs or synthetic augmentation in i...
C. A. Casado, Nhi Nguyen, M. Rahman et al.· 0 citations
It is shown that the structure of service-dependency graphs, modelled as DAGs of compute stages, is a primary determinant of whether decentralised, price-based resource allocation works reliably at scale.
Lauri Lovén, Alaa Saleh, Reza Farahani et al.· arXiv.org· 2 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
