Linear Memory SE(2) Invariant Attention

• URL: http://arxiv.org/abs/2507.18597v1
• Date: Thu, 24 Jul 2025 17:28:57 GMT
• Title: Linear Memory SE(2) Invariant Attention
• Abstract summary: We propose a mechanism for SE(2) invariant scaled dot-product attention that requires linear memory relative to the number of objects in the scene.<n>Our approach is practical to implement and improves performance compared to comparable non-invariant architectures.
• Score: 8.734687343991364
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Abstract: Processing spatial data is a key component in many learning tasks for autonomous driving such as motion forecasting, multi-agent simulation, and planning. Prior works have demonstrated the value in using SE(2) invariant network architectures that consider only the relative poses between objects (e.g. other agents, scene features such as traffic lanes). However, these methods compute the relative poses for all pairs of objects explicitly, requiring quadratic memory. In this work, we propose a mechanism for SE(2) invariant scaled dot-product attention that requires linear memory relative to the number of objects in the scene. Our SE(2) invariant transformer architecture enjoys the same scaling properties that have benefited large language models in recent years. We demonstrate experimentally that our approach is practical to implement and improves performance compared to comparable non-invariant architectures.

Related papers

• Efficient Equivariant Transformer for Self-Driving Agent Modeling [23.858855004235128]
  DriveGATr is a transformer-based architecture for agent modeling.<n>It achieves SE(2)-equivariance without the computational cost of existing methods.<n>DriveGATr is comparable to the state-of-the-art in traffic simulation.
  arXiv  Detail & Related papers  (2026-04-01T23:20:06Z)
• AsCAN: Asymmetric Convolution-Attention Networks for Efficient Recognition and Generation [48.82264764771652]
  We introduce AsCAN -- a hybrid architecture, combining both convolutional and transformer blocks. AsCAN supports a variety of tasks: recognition, segmentation, class-conditional image generation. We then scale the same architecture to solve a large-scale text-to-image task and show state-of-the-art performance.
  arXiv  Detail & Related papers  (2024-11-07T18:43:17Z)
• Transformer Network for Multi-Person Tracking and Re-Identification in Unconstrained Environment [0.6798775532273751]
  Multi-object tracking (MOT) has profound applications in a variety of fields, including surveillance, sports analytics, self-driving, and cooperative robotics. We put forward an integrated MOT method that marries object detection and identity linkage within a singular, end-to-end trainable framework. Our system leverages a robust memory-temporal memory module that retains extensive historical observations and effectively encodes them using an attention-based aggregator.
  arXiv  Detail & Related papers  (2023-12-19T08:15:22Z)
• Bilateral Network with Residual U-blocks and Dual-Guided Attention for Real-time Semantic Segmentation [18.393208069320362]
  We design a new fusion mechanism for two-branch architecture which is guided by attention computation. To be precise, we use the Dual-Guided Attention (DGA) module we proposed to replace some multi-scale transformations. Experiments on Cityscapes and CamVid dataset show the effectiveness of our method.
  arXiv  Detail & Related papers  (2023-10-31T09:20:59Z)
• Multi-body SE(3) Equivariance for Unsupervised Rigid Segmentation and Motion Estimation [49.56131393810713]
  We present an SE(3) equivariant architecture and a training strategy to tackle this task in an unsupervised manner. Our method excels in both model performance and computational efficiency, with only 0.25M parameters and 0.92G FLOPs.
  arXiv  Detail & Related papers  (2023-06-08T22:55:32Z)
• CASAPose: Class-Adaptive and Semantic-Aware Multi-Object Pose Estimation [2.861848675707602]
  We present a new single-stage architecture called CASAPose. It determines 2D-3D correspondences for pose estimation of multiple different objects in RGB images in one pass. It is fast and memory efficient, and achieves high accuracy for multiple objects.
  arXiv  Detail & Related papers  (2022-10-11T10:20:01Z)
• Segmenting Moving Objects via an Object-Centric Layered Representation [100.26138772664811]
  We introduce an object-centric segmentation model with a depth-ordered layer representation. We introduce a scalable pipeline for generating synthetic training data with multiple objects. We evaluate the model on standard video segmentation benchmarks.
  arXiv  Detail & Related papers  (2022-07-05T17:59:43Z)
• Joint Spatial-Temporal and Appearance Modeling with Transformer for Multiple Object Tracking [59.79252390626194]
  We propose a novel solution named TransSTAM, which leverages Transformer to model both the appearance features of each object and the spatial-temporal relationships among objects. The proposed method is evaluated on multiple public benchmarks including MOT16, MOT17, and MOT20, and it achieves a clear performance improvement in both IDF1 and HOTA.
  arXiv  Detail & Related papers  (2022-05-31T01:19:18Z)
• Scalable Video Object Segmentation with Identification Mechanism [125.4229430216776]
  This paper explores the challenges of achieving scalable and effective multi-object modeling for semi-supervised Video Object (VOS) We present two innovative approaches, Associating Objects with Transformers (AOT) and Associating Objects with Scalable Transformers (AOST) Our approaches surpass the state-of-the-art competitors and display exceptional efficiency and scalability consistently across all six benchmarks.
  arXiv  Detail & Related papers  (2022-03-22T03:33:27Z)
• Pairwise Similarity Knowledge Transfer for Weakly Supervised Object Localization [53.99850033746663]
  We study the problem of learning localization model on target classes with weakly supervised image labels. In this work, we argue that learning only an objectness function is a weak form of knowledge transfer. Experiments on the COCO and ILSVRC 2013 detection datasets show that the performance of the localization model improves significantly with the inclusion of pairwise similarity function.
  arXiv  Detail & Related papers  (2020-03-18T17:53:33Z)
• Improving Few-shot Learning by Spatially-aware Matching and CrossTransformer [116.46533207849619]
  We study the impact of scale and location mismatch in the few-shot learning scenario. We propose a novel Spatially-aware Matching scheme to effectively perform matching across multiple scales and locations.
  arXiv  Detail & Related papers  (2020-01-06T14:10:20Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.