Vision-language navigation is a task that requires an agent to follow instructions to navigate in environments. It becomes increasingly crucial in the field of embodied AI, with potential applications in autonomous navigation, search and rescue, and human-robot interaction. In this paper, we propose to address a more practical yet challenging counterpart setting - vision-language navigation in continuous environments (VLN-CE). To develop a robust VLN-CE agent, we propose a new navigation framework, ETPNav, which focuses on two critical skills: 1) the capability to abstract environments and generate long-range navigation plans, and 2) the ability of obstacle-avoiding control in continuous environments. ETPNav performs online topological mapping of environments by self-organizing predicted waypoints along a traversed path, without prior environmental experience. It privileges the agent to break down the navigation procedure into high-level planning and low-level control. Concurrently, ETPNav utilizes a transformer-based cross-modal planner to generate navigation plans based on topological maps and instructions. The plan is then performed through an obstacle-avoiding controller that leverages a trial-and-error heuristic to prevent navigation from getting stuck in obstacles. Experimental results demonstrate the effectiveness of the proposed method. ETPNav yields more than 10% and 20% improvements over prior state-of-the-art on R2R-CE and RxR-CE datasets, respectively.

Scene Rearrangement Planning (SRP) has recently emerged as a crucial interior scene task; however, current approaches still face two primary issues. First, prior works define the action space of SRP using handcrafted coarse-grained actions, which are inflexible for scene arrangement transition and impractical for real-world deployment. Secondly, the scarcity of realistic indoor scene rearrangement data hinders popular data-hungry learning approaches and quantitative evaluation. To tackle these issues, we propose a fine-grained action space definition and curate a large-scale scene rearrangement dataset to facilitate the training of learning approaches and comprehensive benchmarking. Building upon this dataset, we introduce a novel framework, PLATO, designed for efficient agent training and inference. Our approach features an exPert-assisted curriculum Learning (PL) paradigm that possesses a Behavior Cloning (BC) and an offline Reinforcement Learning (RL) curriculum for agent training, along with an advanced tree-search-based planner enhanced by an Adaptive Trade-Off (ATO) strategy to improve expert agent performance further. We demonstrate the superior performance of our method over baseline agents through extensive experiments and provide a detailed analysis to elucidate its rationale. Our project website can be accessed at plato.github.io.

Video Question Answering (VideoQA) is challenging as it requires capturing accurate correlations between modalities from redundant information. Recent methods focus on the explicit challenges of the task, e.g. multimodal feature extraction, video-text alignment and fusion. Their frameworks reason the answer relying on statistical evidence causes, which ignores potential bias in the multimodal data. In our work, we investigate relational structure from a causal representation perspective on multimodal data and propose a novel inference framework. For visual data, question-irrelevant objects may establish simple matching associations with the answer. For textual data, the model prefers the local phrase semantics which may deviate from the global semantics in long sentences. Therefore, to enhance the generalization of the model, we discover the real association by explicitly capturing visual features that are causally related to the question semantics and weakening the impact of local language semantics on question answering. The experimental results on two large causal VideoQA datasets verify that our proposed framework 1) improves the accuracy of the existing VideoQA backbone, 2) demonstrates robustness on complex scenes and questions. The code will be released at https://github.com/Chuanqi-Zang/Discovering-the-Real-Association.

VLN-CE is a recently released embodied task, where AI agents need to navigate a freely traversable environment to reach a distant target location, given language instructions. It poses great challenges due to the huge space of possible strategies. Driven by the belief that the ability to anticipate the consequences of future actions is crucial for the emergence of intelligent and interpretable planning behavior, we propose Dreamwalker — a world model based VLN-CE agent. The world model is built to summarize the visual, topological, and dynamic properties of the complicated continuous environment into a discrete, structured, and compact representation. Dreamwalker can simulate and evaluate possible plans entirely in such internal abstract world, before executing costly actions. As opposed to existing model-free VLN-CE agents simply making greedy decisions in the real world, which easily results in shortsighted behaviors, Dreamwalker is able to make strategic planning through large amounts of "mental experiments." Moreover, the imagined future scenarios reflect our agent’s intention, making its decision-making process more transparent. Extensive experiments and ablation studies on VLN-CE dataset confirm the effectiveness of the proposed approach and outline fruitful directions for future work.

Since the rise of vision-language navigation (VLN), great progress has been made in instruction following - building a follower to navigate environments under the guidance of instructions. However, far less attention has been paid to the inverse task: instruction generation - learning a speaker to generate grounded descriptions for navigation routes. Existing VLN methods train a speaker independently and often treat it as a data augmentation tool to strengthen the follower, while ignoring rich cross-task relations. Here we describe an approach that learns the two tasks simultaneously and exploits their intrinsic correlations to boost the training of each: the follower judges whether the speaker-created instruction explains the original navigation route correctly, and vice versa. Without the need of aligned instruction-path pairs, such cycle-consistent learning scheme is complementary to task-specific training targets defined on labeled data, and can also be applied over unlabeled paths (sampled without paired instructions). Another agent, called creator is added to generate counterfactual environments. It greatly changes current scenes yet leaves novel items - which are vital for the execution of original instructions - unchanged. Thus more informative training scenes are synthesized and the three agents compose a powerful VLN learning system. Extensive experiments on a standard benchmark show that our approach improves the performance of various follower models and produces accurate navigation instructions.

With the emergence of varied visual navigation tasks (e.g., image-/object-/audio-goal and vision-language navigation) that specify the target in different ways, the community has made appealing advances in training specialized agents capable of handling individual navigation tasks well. Given plenty of embodied navigation tasks and task-specific solutions, we address a more fundamental question: can we learn a single powerful agent that masters not one but multiple navigation tasks concurrently? First, we propose VXN, a large-scale 3D dataset that instantiates~four classic navigation tasks in standardized, continuous, and audiovisual-rich environments. Second, we propose Vienna, a versatile embodied navigation agent that simultaneously learns to perform the four navigation tasks with one model. Building upon a full-attentive architecture, Vienna formulates various navigation tasks as a unified, parse-and-query procedure: the target description, augmented with four task embeddings, is comprehensively interpreted into a set of diversified goal vectors, which are refined as the navigation progresses, and used as queries to retrieve supportive context from episodic history for decision making. This enables the reuse of knowledge across navigation tasks with varying input domains/modalities. We empirically demonstrate that, compared with learning each visual navigation task individually, our multitask agent achieves comparable or even better performance with reduced complexity.

Recently, numerous algorithms have been developed to tackle the problem of vision-language navigation (VLN), i.e., entailing an agent to navigate 3D environments through following linguistic instructions. However, current VLN agents simply store their past experiences/observations as latent states in recurrent networks, failing to capture environment layouts and make long-term planning. To address these limitations, we propose a crucial architecture, called Structured Scene Memory (SSM). It is compartmentalized enough to accurately memorize the percepts during navigation. It also serves as a structured scene representation, which captures and disentangles visual and geometric cues in the environment. SSM has a collect-read controller that adaptively collects information for supporting current decision making and mimics iterative algorithms for long-range reasoning. As SSM provides a complete action space, i.e., all the navigable places on the map, a frontier-exploration based navigation decision making strategy is introduced to enable efficient and global planning. Experiment results on two VLN datasets (i.e., R2R and R4R) show that our method achieves state-of-the-art performance on several metrics.

We propose a novel approach for automatically generating a move plan for scene arrangement. 1 Given a scene like an apartment with many furniture objects, to transform its layout into another layout, one would need to determine a collision-free move plan. It could be challenging to design this plan manually because the furniture objects may block the way of each other if not moved properly; and there is a large complex search space of move action sequences that grow exponentially with the number of objects. To tackle this challenge, we propose a learning-based approach to generate a move plan automatically. At the core of our approach is a Monte Carlo tree that encodes possible states of the layout, based on which a search is performed to move a furniture object appropriately in the current layout. We trained a policy neural network embedded with a LSTM module for estimating the best actions to take in the expansion step and simulation step of the Monte Carlo tree search process. Leveraging the power of deep reinforcement learning, the network learned how to make such estimations through millions of trials of moving objects. We demonstrated our approach for moving objects under different scenarios and constraints. We also evaluated our approach on synthetic and real-world layouts, comparing its performance with that of humans and other baseline approaches.

3D object reconstruction is a fundamental task of many robotics and AI problems. With the aid of deep convolutional neural networks (CNNs), 3D object reconstruction has witnessed a significant progress in recent years. However, possibly due to the prohibitively high dimension of the 3D object space, the results from deep CNNs are often prone to missing some shape details. In this paper, we present an approach which aims to preserve more shape details and improve the reconstruction quality. The key idea of our method is to leverage object mask and pose estimation from CNNs to assist the 3D shape learning by constructing a probabilistic singleview visual hull inside of the network. Our method works by first predicting a coarse shape as well as the object pose and silhouette using CNNs, followed by a novel 3D refinement CNN which refines the coarse shapes using the constructed probabilistic visual hulls. Experiment on both synthetic data and real images show that embedding a single-view visual hull for shape refinement can significantly improve the reconstruction quality by recovering more shapes details and improving shape consistency with the input image.

Transferring objects from one place to another place is a common task performed by human in daily life. During this process, it is usually intuitive for humans to choose an object as a proper container and to use an efficient pose to carry objects; yet, it is non-trivial for current computer vision and machine learning algorithms. In this paper, we propose an approach to jointly infer container and human pose for transferring objects by minimizing the costs associated both object and pose candidates. Our approach predicts which object to choose as a container while reasoning about how humans interact with physical surroundings to accomplish the task of transferring objects given visual input. In the learning phase, the presented method learns how humans make rational choices of containers and poses for transferring different objects, as well as the physical quantities required by the transfer task (e.g., compatibility between container and containee, energy cost of carrying pose) via a structured learning approach. In the inference phase, given a scanned 3D scene with different object candidates and a dictionary of human poses, our approach infers the best object as a container together with human pose for transferring a given object.