Get Started
Home
Topics
Search
Library
Research questionHow can learned models approximate trajectory similarity across multiple granularities without losing relative similarity rankings?Trajectory similarity supports spatio-temporal data mining, but learned surrogates can miss patterns that emerge at different trajectory granularities. They must also preserve which trajectories are more similar, not only fit pairwise similarity values.
AI
Machine Learning
Research Paper
Statistical Machine Learning
Technology
Latest papersRecent research connected to this question, newest first.TSMini: A Simple Yet Highly Effective Trajectory Similarity Learning ModelApplies to TSMini, a learned trajectory-similarity model that approximates conventional measures using sub-view modeling and a k-nearest-neighbor-based training loss. The supplied evidence reports experiments on widely used trajectory similarity measures and an average 15% improvement over prior state-of-the-art models; it does not specify datasets or deployment constraints.research paper · Sep 7, 2026
Related questions
How can representation-learning methods for dynamic heterogeneous graphs be compared across temporal granularities?How can learned graph simulators preserve contact history for accurate long-horizon granular-dynamics rollouts?How can model-free robust Q-learning learn from one trajectory with second-moment targets and a noncontractive Bellman operator?How can human motion be retargeted across different humanoid morphologies without hand-crafted mappings or losing pose detail?