Arjun Gupta

Arjun Gupta, Max E. Shepherd, Saurabh Gupta

Mobile manipulation tasks such as opening a door, pulling open a drawer, or lifting a toilet lid require constrained motion of the end-effector under environmental and task constraints. This, coupled with partial information in novel environments, makes it challenging to employ classical motion planning approaches at test time. Our key insight is to cast it as a learning problem to leverage past experience of solving similar planning problems to directly predict motion plans for mobile manipulation tasks in novel situations at test time. To enable this, we develop a simulator, ArtObjSim, that simulates articulated objects placed in real scenes. We then introduce SeqIK+$θ_0$, a fast and flexible representation for motion plans. Finally, we learn models that use SeqIK+$θ_0$ to quickly predict motion plans for articulating novel objects at test time. Experimental evaluation shows improved speed and accuracy at generating motion plans than pure search-based methods and pure learning methods.

Matthew Chang, Arjun Gupta, Saurabh Gupta

This paper tackles the problem of learning value functions from undirected state-only experience (state transitions without action labels i.e. (s,s',r) tuples). We first theoretically characterize the applicability of Q-learning in this setting. We show that tabular Q-learning in discrete Markov decision processes (MDPs) learns the same value function under any arbitrary refinement of the action space. This theoretical result motivates the design of Latent Action Q-learning or LAQ, an offline RL method that can learn effective value functions from state-only experience. Latent Action Q-learning (LAQ) learns value functions using Q-learning on discrete latent actions obtained through a latent-variable future prediction model. We show that LAQ can recover value functions that have high correlation with value functions learned using ground truth actions. Value functions learned using LAQ lead to sample efficient acquisition of goal-directed behavior, can be used with domain-specific low-level controllers, and facilitate transfer across embodiments. Our experiments in 5 environments ranging from 2D grid world to 3D visual navigation in realistic environments demonstrate the benefits of LAQ over simpler alternatives, imitation learning oracles, and competing methods.

Antoni Rosinol, Andrew Violette, Marcus Abate et al.

Humans are able to form a complex mental model of the environment they move in. This mental model captures geometric and semantic aspects of the scene, describes the environment at multiple levels of abstractions (e.g., objects, rooms, buildings), includes static and dynamic entities and their relations (e.g., a person is in a room at a given time). In contrast, current robots' internal representations still provide a partial and fragmented understanding of the environment, either in the form of a sparse or dense set of geometric primitives (e.g., points, lines, planes, voxels) or as a collection of objects. This paper attempts to reduce the gap between robot and human perception by introducing a novel representation, a 3D Dynamic Scene Graph(DSG), that seamlessly captures metric and semantic aspects of a dynamic environment. A DSG is a layered graph where nodes represent spatial concepts at different levels of abstraction, and edges represent spatio-temporal relations among nodes. Our second contribution is Kimera, the first fully automatic method to build a DSG from visual-inertial data. Kimera includes state-of-the-art techniques for visual-inertial SLAM, metric-semantic 3D reconstruction, object localization, human pose and shape estimation, and scene parsing. Our third contribution is a comprehensive evaluation of Kimera in real-life datasets and photo-realistic simulations, including a newly released dataset, uHumans2, which simulates a collection of crowded indoor and outdoor scenes. Our evaluation shows that Kimera achieves state-of-the-art performance in visual-inertial SLAM, estimates an accurate 3D metric-semantic mesh model in real-time, and builds a DSG of a complex indoor environment with tens of objects and humans in minutes. Our final contribution shows how to use a DSG for real-time hierarchical semantic path-planning. The core modules in Kimera are open-source.

Arjun Gupta, Dashiell Kosaka, Edwin Pan et al.

Millimeter-Wave (mmWave) radar sensors are gaining popularity for their robust sensing and increasing imaging capabilities. However, current radar signal processing is hardware specific, which makes it impossible to build sensor agnostic solutions. OpenRadar serves as an interface to prototype, research, and benchmark solutions in a modular manner. This enables creating software processing stacks in a way that has not yet been extensively explored. In the wake of increased AI adoption, OpenRadar can accelerate the growth of the combined fields of radar and AI. The OpenRadar API was released on Oct 2, 2019 as an open-source package under the Apache 2.0 license. The codebase exists at https://github.com/presenseradar/openradar.

Aditya Prakash, Arjun Gupta, Saurabh Gupta

Objects undergo varying amounts of perspective distortion as they move across a camera's field of view. Models for predicting 3D from a single image often work with crops around the object of interest and ignore the location of the object in the camera's field of view. We note that ignoring this location information further exaggerates the inherent ambiguity in making 3D inferences from 2D images and can prevent models from even fitting to the training data. To mitigate this ambiguity, we propose Intrinsics-Aware Positional Encoding (KPE), which incorporates information about the location of crops in the image and camera intrinsics. Experiments on three popular 3D-from-a-single-image benchmarks: depth prediction on NYU, 3D object detection on KITTI & nuScenes, and predicting 3D shapes of articulated objects on ARCTIC, show the benefits of KPE.

Avi Schwarzschild, Arjun Gupta, Amin Ghiasi et al.

It is widely believed that deep neural networks contain layer specialization, wherein neural networks extract hierarchical features representing edges and patterns in shallow layers and complete objects in deeper layers. Unlike common feed-forward models that have distinct filters at each layer, recurrent networks reuse the same parameters at various depths. In this work, we observe that recurrent models exhibit the same hierarchical behaviors and the same performance benefits with depth as feed-forward networks despite reusing the same filters at every recurrence. By training models of various feed-forward and recurrent architectures on several datasets for image classification as well as maze solving, we show that recurrent networks have the ability to closely emulate the behavior of non-recurrent deep models, often doing so with far fewer parameters.

Liam Fowl, Micah Goldblum, Arjun Gupta et al.

Generative models are increasingly able to produce remarkably high quality images and text. The community has developed numerous evaluation metrics for comparing generative models. However, these metrics do not effectively quantify data diversity. We develop a new diversity metric that can readily be applied to data, both synthetic and natural, of any type. Our method employs random network distillation, a technique introduced in reinforcement learning. We validate and deploy this metric on both images and text. We further explore diversity in few-shot image generation, a setting which was previously difficult to evaluate.

Matthew Chang, Arjun Gupta, Saurabh Gupta

Semantic cues and statistical regularities in real-world environment layouts can improve efficiency for navigation in novel environments. This paper learns and leverages such semantic cues for navigating to objects of interest in novel environments, by simply watching YouTube videos. This is challenging because YouTube videos don't come with labels for actions or goals, and may not even showcase optimal behavior. Our method tackles these challenges through the use of Q-learning on pseudo-labeled transition quadruples (image, action, next image, reward). We show that such off-policy Q-learning from passive data is able to learn meaningful semantic cues for navigation. These cues, when used in a hierarchical navigation policy, lead to improved efficiency at the ObjectGoal task in visually realistic simulations. We observe a relative improvement of 15-83% over end-to-end RL, behavior cloning, and classical methods, while using minimal direct interaction.

Arjun Gupta, Luca Carlone

Several autonomy pipelines now have core components that rely on deep learning approaches. While these approaches work well in nominal conditions, they tend to have unexpected and severe failure modes that create concerns when used in safety-critical applications, including self-driving cars. There are several works that aim to characterize the robustness of networks offline, but currently there is a lack of tools to monitor the correctness of network outputs online during operation. We investigate the problem of online output monitoring for neural networks that estimate 3D human shapes and poses from images. Our first contribution is to present and evaluate model-based and learning-based monitors for a human-pose-and-shape reconstruction network, and assess their ability to predict the output loss for a given test input. As a second contribution, we introduce an Adversarially-Trained Online Monitor ( ATOM ) that learns how to effectively predict losses from data. ATOM dominates model-based baselines and can detect bad outputs, leading to substantial improvements in human pose output quality. Our final contribution is an extensive experimental evaluation that shows that discarding outputs flagged as incorrect by ATOM improves the average error by 12.5%, and the worst-case error by 126.5%.

Antoni Rosinol, Arjun Gupta, Marcus Abate et al.

We present a unified representation for actionable spatial perception: 3D Dynamic Scene Graphs. Scene graphs are directed graphs where nodes represent entities in the scene (e.g. objects, walls, rooms), and edges represent relations (e.g. inclusion, adjacency) among nodes. Dynamic scene graphs (DSGs) extend this notion to represent dynamic scenes with moving agents (e.g. humans, robots), and to include actionable information that supports planning and decision-making (e.g. spatio-temporal relations, topology at different levels of abstraction). Our second contribution is to provide the first fully automatic Spatial PerceptIon eNgine(SPIN) to build a DSG from visual-inertial data. We integrate state-of-the-art techniques for object and human detection and pose estimation, and we describe how to robustly infer object, robot, and human nodes in crowded scenes. To the best of our knowledge, this is the first paper that reconciles visual-inertial SLAM and dense human mesh tracking. Moreover, we provide algorithms to obtain hierarchical representations of indoor environments (e.g. places, structures, rooms) and their relations. Our third contribution is to demonstrate the proposed spatial perception engine in a photo-realistic Unity-based simulator, where we assess its robustness and expressiveness. Finally, we discuss the implications of our proposal on modern robotics applications. 3D Dynamic Scene Graphs can have a profound impact on planning and decision-making, human-robot interaction, long-term autonomy, and scene prediction. A video abstract is available at https://youtu.be/SWbofjhyPzI

Igor Gilitschenski, Guy Rosman, Arjun Gupta et al.

In this paper, we propose a novel approach for agent motion prediction in cluttered environments. One of the main challenges in predicting agent motion is accounting for location and context-specific information. Our main contribution is the concept of learning context maps to improve the prediction task. Context maps are a set of location-specific latent maps that are trained alongside the predictor. Thus, the proposed maps are capable of capturing location context beyond visual context cues (e.g. usual average speeds and typical trajectories) or predefined map primitives (such as lanes and stop lines). We pose context map learning as a multi-task training problem and describe our map model and its incorporation into a state-of-the-art trajectory predictor. In extensive experiments, it is shown that use of learned maps can significantly improve predictor accuracy. Furthermore, the performance can be additionally boosted by providing partial knowledge of map semantics.