Contrast & Compress: Learning Lightweight Embeddings For Short Trajectories

Abstract

The ability to retrieve semantically and directionally similar short-range trajectories with both accuracy and efficiency is foundational for downstream applications such as motion forecasting and autonomous navigation. However, prevailing approaches often depend on computationally intensive heuristics or latent anchor representations that lack interpretability and controllability. In this work, we propose a novel framework for learning fixed-dimensional embeddings for short trajectories by leveraging a Transformer encoder trained with a contrastive triplet loss that emphasize the importance of discriminative feature spaces for trajectory data. We analyze the influence of Cosine and FFT-based similarity metrics within the contrastive learning paradigm, with a focus on capturing the nuanced directional intent that characterizes short-term maneuvers. Our empirical evaluation on the Argoverse 2 dataset demonstrates that embeddings shaped by Cosine similarity objectives yield superior clus

Stats

• citations0
• S2 citations—
• github stars0
• HF likes0
• heat score0.00
• arxiv keyvivekanandan2025contrast

Related papers

• Towards Robust Trajectory Embedding For Similarity Computation: When Triangle Inequality Violations In Distance Metrics Matter (2025)0.00
• One Trajectory, One Token: Grounded Video Tokenization Via Panoptic Sub-object Trajectory (2025)0.00
• Contrastive Learning Of Visual-semantic Embeddings (2021)0.00
• Thinking Fast And Slow: Efficient Text-to-visual Retrieval With Transformers (2021)15.16
• Shadow Loss: Memory-linear Deep Metric Learning For Efficient Training (2023)0.00
• Contra: (con)text (tra)nsformer For Cross-modal Video Retrieval (2022)2.26
• Optimization Of Latent-space Compression Using Game-theoretic Techniques For Transformer-based Vector Search (2025)0.00
• Geopth: A Lightweight Approach To Category-based Trajectory Retrieval Via Geometric Prototype Trajectory Hashing (2025)0.00