Learning to Navigate Efficiently with Only 0.58M Trainable Parameters

• URL: http://arxiv.org/abs/2607.11029v1
• Date: Mon, 13 Jul 2026 02:49:55 GMT
• Title: Learning to Navigate Efficiently with Only 0.58M Trainable Parameters
• Abstract summary: We propose a navigation model in which operations with known closed-form structure are computed analytically and serve as interfaces between three small learned modules.<n>The system trains only 0.58M out of a total of 22.7M parameters, on 44k frames in under one GPU-hour, yet approaches the performance of state-of-the-art models on navigation tasks.
• Score: 12.68304107799412
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Recent progress in visual navigation has largely been driven by scale: end-to-end policies with hundreds of millions of parameters trained on billions of frames or large-scale simulated data. We ask how much of this scale a single task family actually requires, and what structure can substitute for it. We propose a decomposed navigation model in which operations with known closed-form structure, such as projective geometry, occupancy, and coordinate transforms, are computed analytically and serve as interfaces between three small learned modules: an egress predictor that grounds the episode goal as a local subgoal in the current view, a navigation predictor that estimates a goal-conditioned posterior over where trajectories travel, and an endpoint-pinned residual diffusion generator that samples trajectory shapes from this posterior. The system trains only 0.58M out of a total of 22.7M parameters, on 44k frames in under one GPU-hour, yet approaches the performance of state-of-the-art models on navigation tasks across 6060 point-goal episodes and 60 environments, while having 233x fewer trainable parameters, the lowest collision rate among all evaluated methods, and 50 Hz inference speed. The decomposition further transfers to no-goal exploration by retraining only the 123k-parameter egress head, and its failure modes under sensor corruption are transparent and analytically correctable.

Related papers

• FlowPilot: Real-Time World-Action Modeling for Agile UAV Navigation [18.092834753587454]
  FlowPilot is a compact world-action model for real-time onboard UAV navigation from depth.<n>A dual-stream mixture-of-transformers couples video and action experts through shared attention, allowing future-scene prediction and trajectory generation to inform each other.
  arXiv  Detail & Related papers  (2026-08-01T12:43:31Z)
• DepthART: Scaling Foundation Monocular Depth to Tiny Models [51.10174763031444]
  DepthART is a compact MDE model for on-device deployment across diverse scenes.<n>We show how DepthART consistently surpasses previous tiny baselines in both zero-shot generalization and metric accuracy.<n>We also provide a scalable model family, with DepthART-S reaching 347/245 FPS (strict FP32) on an GTX A6000 at $2242/4482$, 102 FPS (TF32) on a Orin NX 8GB, and over 15 FPS (FP32) on a Jetson Nano 4GB.
  arXiv  Detail & Related papers  (2026-07-19T06:54:52Z)
• DRIFT: Drift and Aggregation for Motion Planning [20.309189828710362]
  DRIFT is a fixed-depth trajectory planner that combines one-step drifting in a compact trajectory latent space with scene-aware proposal aggregation.<n>A lightweight Aggregation Head integrates these features with scene, navigation, and ego-state information and directly predicts the final trajectory.<n>Its output is trained with expert-trajectory imitation and a map-derived boundary regularizer that penalizes waypoints outside the drivable polygon.
  arXiv  Detail & Related papers  (2026-07-16T02:43:13Z)
• ShipNet: A Geometric Deep Learning Surrogate for Real-Time Ship Hydrodynamics [0.0]
  ShipNet is a geometric deep-learning surrogate that predicts both hull-surface pressure distributions and far-field free-surface wave patterns directly from hull geometry and speed.<n>Inference requires approximately 0.15s per case, yielding over a 550x speedup relative to the potential-flow solver on conventional hardware.
  arXiv  Detail & Related papers  (2026-06-13T15:38:24Z)
• Towards Real-Time Autonomous Navigation: Transformer-Based Catheter Tip Tracking in Fluoroscopy [35.565354067805835]
  This paper aims to develop and evaluate a real-time catheter tip tracking pipeline under fluoroscopy.<n>Deep learning segmentation models, including U-Net, U-Net+Transformer, and SegFormer, were trained and benchmarked.<n>The two-class SegFormer achieved a mean absolute error of 4.44 mm, outperforming U-Net (4.60 mm), U-Net+Transformer (6.20 mm), and all three-class models (5.19-7.74 mm)
  arXiv  Detail & Related papers  (2026-05-14T01:42:56Z)
• Trajectory-Aware Adaptive Inference in Object Detection Models [0.19845233274307664]
  We build upon an existing object detection framework by incorporating GPS trajectory data into the inference process.<n>We introduce an early-exit mechanism in a YOLOv8-based detector that incorporates motion cues.<n>Results demonstrate that this strategy maintains satisfactory detection performance while significantly reducing inference time and computational cost.
  arXiv  Detail & Related papers  (2026-05-12T16:04:07Z)
• WUTDet: A 100K-Scale Ship Detection Dataset and Benchmarks with Dense Small Objects [68.44670476527155]
  WUTDet is a large-scale ship detection dataset with 100,576 images and 381,378 annotated ship instances.<n>We systematically evaluate 20 baseline models from three mainstream detection architectures, namely CNN, Transformer, and Mamba.<n>Results on Ship-GEN show that models trained on WUTDet exhibit stronger generalization under different data distributions.
  arXiv  Detail & Related papers  (2026-04-09T03:28:05Z)
• HCLSM: Hierarchical Causal Latent State Machines for Object-Centric World Modeling [0.0]
  HCLSM operates on three principles: object-centric decomposition, spatial broadcast decoding, and hierarchical temporal dynamics.<n>System spans 8,478 lines of Python across 51 modules with 171 unit tests.
  arXiv  Detail & Related papers  (2026-03-31T00:11:29Z)
• End-to-End Spatial-Temporal Transformer for Real-time 4D HOI Reconstruction [74.31251139839047]
  THO is an end-to-end Spatial-Temporal Transformer that predicts human motion and coordinated object motion in a forward fashion from the given video and 3D template.<n> experiments demonstrate that THO operates at an inference speed of 31.5 FPS on a single 4090 GPU, achieving a >600x speedup over prior optimization-based methods.
  arXiv  Detail & Related papers  (2026-03-15T15:21:36Z)
• MetricNet: Recovering Metric Scale in Generative Navigation Policies [51.90872764552077]
  MetricNet is an effective add-on for generative navigation that predicts the metric distance between waypoints.<n>We show that executing MetricNet-scaled waypoints significantly improves both navigation and exploration performance.<n>We also propose MetricNav, which integrates MetricNet into a navigation policy to guide the robot away from obstacles while still moving towards the goal.
  arXiv  Detail & Related papers  (2025-09-17T13:37:13Z)
• CabiNet: Scaling Neural Collision Detection for Object Rearrangement with Procedural Scene Generation [54.68738348071891]
  We first generate over 650K cluttered scenes - orders of magnitude more than prior work - in diverse everyday environments. We render synthetic partial point clouds from this data and use it to train our CabiNet model architecture. CabiNet is a collision model that accepts object and scene point clouds, captured from a single-view depth observation.
  arXiv  Detail & Related papers  (2023-04-18T21:09:55Z)
• Ultra-low Power Deep Learning-based Monocular Relative Localization Onboard Nano-quadrotors [64.68349896377629]
  This work presents a novel autonomous end-to-end system that addresses the monocular relative localization, through deep neural networks (DNNs), of two peer nano-drones. To cope with the ultra-constrained nano-drone platform, we propose a vertically-integrated framework, including dataset augmentation, quantization, and system optimizations. Experimental results show that our DNN can precisely localize a 10cm-size target nano-drone by employing only low-resolution monochrome images, up to 2m distance.
  arXiv  Detail & Related papers  (2023-03-03T14:14:08Z)
• Learning Spatio-Temporal Transformer for Visual Tracking [108.11680070733598]
  We present a new tracking architecture with an encoder-decoder transformer as the key component. The whole method is end-to-end, does not need any postprocessing steps such as cosine window and bounding box smoothing. The proposed tracker achieves state-of-the-art performance on five challenging short-term and long-term benchmarks, while running real-time speed, being 6x faster than Siam R-CNN.
  arXiv  Detail & Related papers  (2021-03-31T15:19:19Z)
• Object Tracking by Detection with Visual and Motion Cues [1.7818230914983044]
  Self-driving cars need to detect and track objects in camera images. We present a simple online tracking algorithm that is based on a constant velocity motion model with a Kalman filter. We evaluate our approach on the challenging BDD100 dataset.
  arXiv  Detail & Related papers  (2021-01-19T10:29:16Z)

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.