PointGoal Navigation is an embodied task that requires agents to navigate to a specified point in an unseen environment. Wijmans et al. showed that this task is solvable but their method is computationally prohibitive, requiring 2.5 billion frames and 180 GPU-days. In this work, we develop a method to significantly increase sample and time efficiency in learning PointNav using self-supervised auxiliary tasks (e.g. predicting the action taken between two egocentric observations, predicting the distance between two observations from a trajectory,etc.).We find that naively combining multiple auxiliary tasks improves sample efficiency,but only provides marginal gains beyond a point. To overcome this, we use attention to combine representations learnt from individual auxiliary tasks. Our best agent is 5.5x faster to reach the performance of the previous state-of-the-art, DD-PPO, at 40M frames, and improves on DD-PPO's performance at 40M frames by 0.16 SPL. Our code is publicly available at https://github.com/joel99/habitat-pointnav-aux.
University of Illinois, Urbana-Champaign（伊利诺伊大学厄巴纳-尚佩恩分校）；Georgia Institute of Technology（佐治亚理工学院）；Facebook AI Research（Facebook AI研究院）
It is fundamental for personal robots to reliably navigate to a specified goal. To study this task, PointGoal navigation has been introduced in simulated Embodied AI environments. Recent advances solve this PointGoal navigation task with near-perfect accuracy (99.6% success) in photo-realistically simulated environments, assuming noiseless egocentric vision, noiseless actuation, and most importantly, perfect localization. However, under realistic noise models for visual sensors and actuation, and without access to a "GPS and Compass sensor," the 99.6%-success agents for PointGoal navigation on…
Georgia Institute of Technology（佐治亚理工学院）；Facebook AI Research（Facebook人工智能研究院）；Google Research（谷歌研究院）
PointGoal navigation has seen significant recent interest and progress, spurred on by the Habitat platform and associated challenge. In this paper, we study PointGoal navigation under both a sample budget (75 million frames) and a compute budget (1 GPU for 1 day). We conduct an extensive set of experiments, cumulatively totaling over 50,000 GPU-hours, that let us identify and discuss a number of ostensibly minor but significant design choices -- the advantage estimation procedure (a key component in training), visual encoder architecture, and a seemingly minor hyper-parameter change. Overall, …
Georgia Institute of Technology（佐治亚理工学院）；Facebook AI Research（Facebook人工智能研究院）
We propose a novel architecture and training paradigm for training realistic PointGoal Navigation -- navigating to a target coordinate in an unseen environment under actuation and sensor noise without access to ground-truth localization. Specifically, we find that the primary challenge under this setting is learning localization -- when stripped of idealized localization, agents fail to stop precisely at the goal despite reliably making progress towards it. To address this we introduce a set of auxiliary losses to help the agent learn localization. Further, we explore the idea of treating the …
Ukrainian Catholic University（乌克兰天主教大学）；Georgia Institute of Technology（佐治亚理工学院）；Meta AI
Can an autonomous agent navigate in a new environment without building an explicit map? For the task of PointGoal navigation ('Go to $Δx$, $Δy$') under idealized settings (no RGB-D and actuation noise, perfect GPS+Compass), the answer is a clear 'yes' - map-less neural models composed of task-agnostic components (CNNs and RNNs) trained with large-scale reinforcement learning achieve 100% Success on a standard dataset (Gibson). However, for PointNav in a realistic setting (RGB-D and actuation noise, no GPS+Compass), this is an open question; one we tackle in this paper. The strongest published …
Georgia Institute of Technology（佐治亚理工学院）；Facebook AI Research（Facebook人工智能研究院）；Oregon State University（俄勒冈州立大学）；Simon Fraser University（西蒙弗雷泽大学）
We present Decentralized Distributed Proximal Policy Optimization (DD-PPO), a method for distributed reinforcement learning in resource-intensive simulated environments. DD-PPO is distributed (uses multiple machines), decentralized (lacks a centralized server), and synchronous (no computation is ever stale), making it conceptually simple and easy to implement. In our experiments on training virtual robots to navigate in Habitat-Sim, DD-PPO exhibits near-linear scaling -- achieving a speedup of 107x on 128 GPUs over a serial implementation. We leverage this scaling to train an agent for 2.5 Bil…
University of Illinois at Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）；Allen Institute for AI（艾伦人工智能研究所）
While deep reinforcement learning (RL) promises freedom from hand-labeled data, great successes, especially for Embodied AI, require significant work to create supervision via carefully shaped rewards. Indeed, without shaped rewards, i.e., with only terminal rewards, present-day Embodied AI results degrade significantly across Embodied AI problems from single-agent Habitat-based PointGoal Navigation (SPL drops from 55 to 0) and two-agent AI2-THOR-based Furniture Moving (success drops from 58% to 1%) to three-agent Google Football-based 3 vs. 1 with Keeper (game score drops from 0.6 to 0.1). As…
Georgia Institute of Technology（佐治亚理工学院）；FAIR, Meta AI（元人工智能基础人工智能研究院）；Simon Fraser University（西蒙弗雷泽大学）；Google Research Atlanta（谷歌亚特兰大研究院）；Oregon State University（俄勒冈州立大学）
Animal navigation research posits that organisms build and maintain internal spatial representations, or maps, of their environment. We ask if machines -- specifically, artificial intelligence (AI) navigation agents -- also build implicit (or 'mental') maps. A positive answer to this question would (a) explain the surprising phenomenon in recent literature of ostensibly map-free neural-networks achieving strong performance, and (b) strengthen the evidence of mapping as a fundamental mechanism for navigation by intelligent embodied agents, whether they be biological or artificial. Unlike animal…
In recent years several learning approaches to point goal navigation in previously unseen environments have been proposed. They vary in the representations of the environments, problem decomposition, and experimental evaluation. In this work, we compare the state-of-the-art Deep Reinforcement Learning based approaches with Partially Observable Markov Decision Process (POMDP) formulation of the point goal navigation problem. We adapt the (POMDP) sub-goal framework proposed by [1] and modify the component that estimates frontier properties by using partial semantic maps of indoor scenes built fr…
Georgia Institute of Technology（佐治亚理工学院）；Meta AI；Google Atlanta（谷歌亚特兰大）
We present Variable Experience Rollout (VER), a technique for efficiently scaling batched on-policy reinforcement learning in heterogenous environments (where different environments take vastly different times to generate rollouts) to many GPUs residing on, potentially, many machines. VER combines the strengths of and blurs the line between synchronous and asynchronous on-policy RL methods (SyncOnRL and AsyncOnRL, respectively). VER learns from on-policy experience (like SyncOnRL) and has no synchronization points (like AsyncOnRL). VER leads to significant and consistent speed-ups across a bro…
University of Modena and Reggio Emilia（摩德纳和雷焦艾米利亚大学）
The research field of Embodied AI has witnessed substantial progress in visual navigation and exploration thanks to powerful simulating platforms and the availability of 3D data of indoor and photorealistic environments. These two factors have opened the doors to a new generation of intelligent agents capable of achieving nearly perfect PointGoal Navigation. However, such architectures are commonly trained with millions, if not billions, of frames and tested in simulation. Together with great enthusiasm, these results yield a question: how many researchers will effectively benefit from these a…
IndustryNav: Exploring Spatial Reasoning of Embodied Agents in Dynamic Industrial Navigation
Michigan State University（密歇根州立大学）；Independent Researcher（独立研究者）；University of California, Irvine（加州大学尔湾分校）；Ohio State University（俄亥俄州立大学）；University of North Carolina at Chapel Hill（北卡罗来纳大学教堂山分校）；University of California, Santa Barbara（加州大学圣巴巴拉分校）；University of Pittsburgh（匹兹堡大学）；University of Virginia（弗吉尼亚大学）；Arizona State University（亚利桑那州立大学）；Purdue University Northwest（普渡大学西北分校）
Consistent Attack: Universal Adversarial Perturbation on Embodied Vision Navigation
Department of Computer Science and Technology, Beijing National Research Center for Information Science and Technology, Tsinghua-Bosch Joint Center for Machine Learning, Institute for Artificial Intelligence, Tsinghua University, Beijing 100084, China（计算机科学与技术系，北京信息科学国家研究中心，清华-博世联合机器学习中心，人工智能研究院，清华大学，北京）
Monash University（莫纳什大学）；University of Chinese Academy of Sciences（中国科学院大学）；Sun Yat-sen University（中山大学）；Dark Matter AI Inc.（暗物质人工智能公司）
Vision-Language Navigation (VLN) is a task where agents learn to navigate following natural language instructions. The key to this task is to perceive both the visual scene and natural language sequentially. Conventional approaches exploit the vision and language features in cross-modal grounding. However, the VLN task remains challenging, since previous works have neglected the rich semantic information contained in the environment (such as implicit navigation graphs or sub-trajectory semantics). In this paper, we introduce Auxiliary Reasoning Navigation (AuxRN), a framework with four self-su…
Georgia Institute of Technology（佐治亚理工学院）；Salesforce Research（Salesforce 研究院）；University of Maryland, College Park（马里兰大学帕克分校）
The Vision-and-Language Navigation (VLN) task entails an agent following navigational instruction in photo-realistic unknown environments. This challenging task demands that the agent be aware of which instruction was completed, which instruction is needed next, which way to go, and its navigation progress towards the goal. In this paper, we introduce a self-monitoring agent with two complementary components: (1) visual-textual co-grounding module to locate the instruction completed in the past, the instruction required for the next action, and the next moving direction from surrounding images…
In Vision-and-Language Navigation (VLN), researchers typically take an image encoder pre-trained on ImageNet without fine-tuning on the environments that the agent will be trained or tested on. However, the distribution shift between the training images from ImageNet and the views in the navigation environments may render the ImageNet pre-trained image encoder suboptimal. Therefore, in this paper, we design a set of structure-encoding auxiliary tasks (SEA) that leverage the data in the navigation environments to pre-train and improve the image encoder. Specifically, we design and customize (1)…
The Robotics Institute, Carnegie Mellon University（卡内基梅隆大学机器人研究所）；General Motors Global R&D（通用汽车全球研发中心）
A safe and robust on-road navigation system is a crucial component of achieving fully automated vehicles. NVIDIA recently proposed an End-to-End algorithm that can directly learn steering commands from raw pixels of a front camera by using one convolutional neural network. In this paper, we leverage auxiliary information aside from raw images and design a novel network structure, called Auxiliary Task Network (ATN), to help boost the driving performance while maintaining the advantage of minimal training data and an End-to-End training method. In this network, we introduce human prior knowledg…
Current datasets to train social behaviors are usually borrowed from surveillance applications that capture visual data from a bird's-eye perspective. This leaves aside precious relationships and visual cues that could be captured through a first-person view of a scene. In this work, we propose a strategy to exploit the power of current game engines, such as Unity, to transform pre-existing bird's-eye view datasets into a first-person view, in particular, a depth view. Using this strategy, we are able to generate large volumes of synthetic data that can be used to pre-train a social navigation…
Most common navigation tasks in human environments require auxiliary arm interactions, e.g. opening doors, pressing buttons and pushing obstacles away. This type of navigation tasks, which we call Interactive Navigation, requires the use of mobile manipulators: mobile bases with manipulation capabilities. Interactive Navigation tasks are usually long-horizon and composed of heterogeneous phases of pure navigation, pure manipulation, and their combination. Using the wrong part of the embodiment is inefficient and hinders progress. We propose HRL4IN, a novel Hierarchical RL architecture for Inte…
University of Washington（华盛顿大学）；Facebook AI Research（脸书人工智能研究院）；Georgia Institute of Technology（佐治亚理工学院）
We propose SplitNet, a method for decoupling visual perception and policy learning. By incorporating auxiliary tasks and selective learning of portions of the model, we explicitly decompose the learning objectives for visual navigation into perceiving the world and acting on that perception. We show dramatic improvements over baseline models on transferring between simulators, an encouraging step towards Sim2Real. Additionally, SplitNet generalizes better to unseen environments from the same simulator and transfers faster and more effectively to novel embodied navigation tasks. Further, given …
Ecole Polytechnique Fédérale de Lausanne (EPFL)（洛桑联邦理工学院）；National Institute of Informatics（国立信息学研究所）
Enabling users to interactively navigate through different viewpoints of a static scene is a new interesting functionality in 3D streaming systems. While it opens exciting perspectives towards rich multimedia applications, it requires the design of novel representations and coding techniques in order to solve the new challenges imposed by interactive navigation. Interactivity clearly brings new design constraints: the encoder is unaware of the exact decoding process, while the decoder has to reconstruct information from incomplete subsets of data since the server can generally not transmit ima…
This paper presents a novel approach for aerial drone autonomous navigation along predetermined paths using only visual input form an onboard camera and without reliance on a Global Positioning System (GPS). It is based on using a deep Convolutional Neural Network (CNN) combined with a regressor to output the drone steering commands. Furthermore, multiple auxiliary navigation paths that form a navigation envelope are used for data augmentation to make the system adaptable to real-life deployment scenarios. The approach is suitable for automating drone navigation in applications that exhibit re…
Agents navigating in 3D environments require some form of memory, which should hold a compact and actionable representation of the history of observations useful for decision taking and planning. In most end-to-end learning approaches the representation is latent and usually does not have a clearly defined interpretation, whereas classical robotics addresses this with scene reconstruction resulting in some form of map, usually estimated with geometry and sensor models and/or learning. In this work we propose to learn an actionable representation of the scene independently of the targeted downs…
Contrastive language image pretraining (CLIP) encoders have been shown to be beneficial for a range of visual tasks from classification and detection to captioning and image manipulation. We investigate the effectiveness of CLIP visual backbones for Embodied AI tasks. We build incredibly simple baselines, named EmbCLIP, with no task specific architectures, inductive biases (such as the use of semantic maps), auxiliary tasks during training, or depth maps -- yet we find that our improved baselines perform very well across a range of tasks and simulators. EmbCLIP tops the RoboTHOR ObjectNav lead…
Meta AI Research（Meta人工智能研究院）；Georgia Institute of Technology（佐治亚理工学院）
How should we learn visual representations for embodied agents that must see and move? The status quo is tabula rasa in vivo, i.e. learning visual representations from scratch while also learning to move, potentially augmented with auxiliary tasks (e.g. predicting the action taken between two successive observations). In this paper, we show that an alternative 2-stage strategy is far more effective: (1) offline pretraining of visual representations with self-supervised learning (SSL) using large-scale pre-rendered images of indoor environments (Omnidata), and (2) online finetuning of visuomoto…
UniWAM Technical Report: Unified Mobile Manipulation via Mixed-Stream World-Action Modeling and Manipulation Anchor Pose Supervision
Physical Superintelligence Lab, Fysics AI（Fysics AI 物理超级智能实验室）；College of Intelligent Robotics and Advanced Manufacturing, Fudan University（复旦大学智能机器人与先进制造学院）
South China University of Technology（华南理工大学）；Pazhou Laboratory（琶洲实验室）；Shenzhen University（深圳大学）；Information Technology R&D Innovation Center of Peking University（北京大学信息技术研发创新中心）；MIT-IBM Watson AI Lab（MIT-IBM沃森人工智能实验室）；UMass Amherst（马萨诸塞大学阿默斯特分校）
We address a practical yet challenging problem of training robot agents to navigate in an environment following a path described by some language instructions. The instructions often contain descriptions of objects in the environment. To achieve accurate and efficient navigation, it is critical to build a map that accurately represents both spatial location and the semantic information of the environment objects. However, enabling a robot to build a map that well represents the environment is extremely challenging as the environment often involves diverse objects with various attributes. In th…
Facebook AI Research（Facebook人工智能研究）；UT Austin（德克萨斯大学奥斯汀分校）；Georgia Tech（佐治亚理工学院）；Simon Fraser University（西蒙菲莎大学）；Cornell University（康奈尔大学）
We present the Habitat-Matterport 3D (HM3D) dataset. HM3D is a large-scale dataset of 1,000 building-scale 3D reconstructions from a diverse set of real-world locations. Each scene in the dataset consists of a textured 3D mesh reconstruction of interiors such as multi-floor residences, stores, and other private indoor spaces. HM3D surpasses existing datasets available for academic research in terms of physical scale, completeness of the reconstruction, and visual fidelity. HM3D contains 112.5k m^2 of navigable space, which is 1.4 - 3.7x larger than other building-scale datasets such as MP3D an…
Sun Yat-sen University（中山大学）；University of Chinese Academy of Sciences（中国科学院大学）
Language instruction plays an essential role in the natural language grounded navigation tasks. However, navigators trained with limited human-annotated instructions may have difficulties in accurately capturing key information from the complicated instruction at different timesteps, leading to poor navigation performance. In this paper, we exploit to train a more robust navigator which is capable of dynamically extracting crucial factors from the long instruction, by using an adversarial attacking paradigm. Specifically, we propose a Dynamic Reinforced Instruction Attacker (DR-Attacker), whic…
DreamFly: Causal Memory and Receding-Horizon Diffusion Planning for Aerial Vision-Language Navigation
School of Electronic Information Engineering, Xi’an Technological University（西安工业大学电子信息工程学院）；School of Computer Science and Engineering, Xi’an Technological University（西安工业大学计算机科学与工程学院）
Effectively utilizing the vast amounts of ego-centric navigation data that is freely available on the internet can advance generalized intelligent systems, i.e., to robustly scale across perspectives, platforms, environmental conditions, scenarios, and geographical locations. However, it is difficult to directly leverage such large amounts of unlabeled and highly diverse data for complex 3D reasoning and planning tasks. Consequently, researchers have primarily focused on its use for various auxiliary pixel- and image-level computer vision tasks that do not consider an ultimate navigational obj…
EvolveNav: Empowering LLM-Based Vision-Language Navigation via Self-Improving Embodied Reasoning
Shanghai Jiao Tong University（上海交通大学）；Shenzhen Campus of Sun Yat-sen University（中山大学深圳校区）；Peng Cheng Laboratory（鹏城实验室）；Tsinghua Shenzhen International Graduate School（清华大学深圳国际研究生院）；Mohamed Bin Zayed University of Artificial Intelligence（Mohamed Bin Zayed人工智能大学）；Yinwang Intelligent Technology Co.（亿维智能技术有限公司）
SRI International（SRI国际）；Arizona State University（亚利桑那州立大学）
Visual navigation for autonomous agents is a core task in the fields of computer vision and robotics. Learning-based methods, such as deep reinforcement learning, have the potential to outperform the classical solutions developed for this task; however, they come at a significantly increased computational load. Through this work, we design a novel approach that focuses on performing better or comparable to the existing learning-based solutions but under a clear time/computational budget. To this end, we propose a method to encode vital scene semantics such as traversable paths, unexplored area…
Université de Montréal（蒙特利尔大学）；Mila – Quebec AI Institute（米拉-魁北克人工智能研究所）；Allen Institute for AI（艾伦人工智能研究所）
Deep reinforcement learning has shown promising results on an abundance of robotic tasks in simulation, including visual navigation and manipulation. Prior work generally aims to build embodied agents that solve their assigned tasks as quickly as possible, while largely ignoring the problems caused by collision with objects during interaction. This lack of prioritization is understandable: there is no inherent cost in breaking virtual objects. As a result, "well-trained" agents frequently collide with objects before achieving their primary goals, a behavior that would be catastrophic in the re…
Learning to navigate in complex environments with dynamic elements is an important milestone in developing AI agents. In this work we formulate the navigation question as a reinforcement learning problem and show that data efficiency and task performance can be dramatically improved by relying on additional auxiliary tasks leveraging multimodal sensory inputs. In particular we consider jointly learning the goal-driven reinforcement learning problem with auxiliary depth prediction and loop closure classification tasks. This approach can learn to navigate from raw sensory input in complicated 3D…
Georgia Institute of Technology（佐治亚理工学院）；Facebook AI Research（脸书人工智能研究院）
ObjectGoal Navigation (ObjectNav) is an embodied task wherein agents are to navigate to an object instance in an unseen environment. Prior works have shown that end-to-end ObjectNav agents that use vanilla visual and recurrent modules, e.g. a CNN+RNN, perform poorly due to overfitting and sample inefficiency. This has motivated current state-of-the-art methods to mix analytic and learned components and operate on explicit spatial maps of the environment. We instead re-enable a generic learned agent by adding auxiliary learning tasks and an exploration reward. Our agents achieve 24.5% success a…
