paper-with-me

Papers

On Pre-Training for Visuo-Motor Control: Revisiting a Learning-from-Scratch Baseline

2022-12-12 · Nicklas Hansen, Zhecheng Yuan, Yanjie Ze, Tongzhou Mu, Aravind Rajeswaran, Hao Su, Huazhe Xu, Xiaolong Wang

In this paper, we examine the effectiveness of pre-training for visuo-motor control tasks. We revisit a simple Learning-from-Scratch (LfS) baseline that incorporates data augmentation and a shallow ConvNet, and find that this baseline is surprisingly competitive with recent approaches (PVR, MVP, R3M) that leverage frozen visual representations trained on large-scale vision datasets -- across a variety of algorithms, task domains, and metrics in simulation and on a real robot. Our results demonstrate that these methods are hindered by a significant domain gap between the pre-training datasets and current benchmarks for visuo-motor control, which is alleviated by finetuning. Based on our findings, we provide recommendations for future research in pre-training for control and hope that our simple yet strong baseline will aid in accurately benchmarking progress in this area.

📄 PDF Abstract BibTeX arXiv:2212.05749

Code (1)

gemcollector/learning-from-scratch 공식 구현 pytorch

Tasks

BenchmarkingData Augmentation

Similar Papers 제목 키워드 기반

Introspective Visuomotor Control: Exploiting Uncertainty in Deep Visuomotor Control for Failure Recovery

2021-03-22 · Chia-Man Hung, Li Sun, Yizhe Wu, Ioannis Havoutis 외

End-to-end visuomotor control is emerging as a compelling solution for robot manipulation tasks. However, imitation learning-based visuomotor control approaches tend to suffer from a common limitation, lacking the abilit…

Imitation LearningRobot Manipulation

Adversarial Feature Training for Generalizable Robotic Visuomotor Control

2019-09-17 · Xi Chen, Ali Ghadirzadeh, Mårten Björkman, Patric Jensfelt

Deep reinforcement learning (RL) has enabled training action-selection policies, end-to-end, by learning a function which maps image pixels to action outputs. However, it's application to visuomotor robotic policy traini…

Deep Reinforcement LearningReinforcement LearningReinforcement Learning (RL)Transfer Learning

Adapting Pretrained ViTs with Convolution Injector for Visuo-Motor Control

2024-06-10 · Dongyoon Hwang, Byungkun Lee, Hojoon Lee, Hyunseung Kim 외

Vision Transformers (ViT), when paired with large-scale pretraining, have shown remarkable performance across various computer vision tasks, primarily due to their weak inductive bias. However, while such weak inductive …

Inductive Bias

Efficient Training of Generalizable Visuomotor Policies via Control-Aware Augmentation

2024-01-17 · Yinuo Zhao, Kun Wu, Tianjiao Yi, Zhiyuan Xu 외

Improving generalization is one key challenge in embodied AI, where obtaining large-scale datasets across diverse scenarios is costly. Traditional weak augmentations, such as cropping and flipping, are insufficient for i…

Data AugmentationReinforcement Learning (RL)Robot Manipulation

Unified Visuomotor Targets: Supervising VLAs Beyond Physical Actions

2026-08-04 · Zhenyang Feng, Unnat Jain arxiv

VLA models are trained to predict robot actions from visual and language observations. This is a natural choice, but it creates a mismatch: VLMs encode rich, high-level representations of scenes and goals, while robot ac…