In this paper, we develop a structured critique of robotic simulations for real-world manipulation, by arguing that scaling simulators is neither necessary nor sufficient for making progress in general-purpose real-world robotic manipulation agents that are compliant with human preferences. With the ubiquity of robotic simulators, and recent efforts to scale them for diverse tasks, and at the same time the interest in generally capable real-world manipulation systems, we believe it is important to address the limitations of using simulation for real-world manipulation, so that as a community, we can focus our collective resources, energy, and time on approaches that have more principled odds of success. We further demonstrate the unique challenges that real-world manipulation presents, and show through examples and arguments why scaling simulation doesn't get us closer to solving these challenges required for diverse real-world deployment.

We pursue the goal of developing robots that can interact zero-shot with generic unseen objects via a diverse repertoire of manipulation skills and show how passive human videos can serve as a rich source of data for learning such generalist robots. Unlike typical robot learning approaches which directly learn how a robot should act from interaction data, we adopt a factorized approach that can leverage large-scale human videos to learn how a human would accomplish a desired task (a human ‘plan’), followed by ‘translating’ this plan to the robot’s embodiment. Specifically, we learn a human ‘plan predictor’ that, given a current image of a scene and a goal image, predicts the future hand and object configurations. We combine this with a ‘translation’ module that learns a plan-conditioned robot manipulation policy, and allows following humans plans for generic manipulation tasks in a zero-shot manner with no deployment-time training. Importantly, while the plan predictor can leverage large-scale human videos for learning, the translation module only requires a small amount of in-domain data, and can generalize to tasks not seen during training. We show that our learned system can perform over 16 manipulation skills that generalize to 40 objects, encompassing 100 real-world tasks for table-top manipulation and diverse in-the-wild manipulation. https://homangab.github.io/hopman/

The grand aim of having a single robot that can manipulate arbitrary objects in diverse settings is at odds with the paucity of robotics datasets. Acquiring and growing such datasets is strenuous due to manual efforts, operational costs, and safety challenges. A path toward such a universal agent requires an efficient framework capable of generalization but within a reasonable data budget. In this paper, we develop an efficient framework (MT-ACT) for training universal agents capable of multi-task manipulation skills using (a) semantic augmentations that can rapidly multiply existing datasets and (b) action representations that can extract performant policies with small yet diverse multi-modal datasets without overfitting. In addition, reliable task conditioning and an expressive policy architecture enables our agent to exhibit a diverse repertoire of skills in novel situations specified using task commands. Using merely 7500 demonstrations, we are able to train a single policy RoboAgent capable of 12 unique skills, and demonstrate its generalization over 38 tasks spread across common daily activities in diverse kitchen scenes. On average, RoboAgent outperforms prior methods by over 40% in unseen situations while being more sample efficient. See https://robopen.github.io/for video results and appendix.

Motivated by the intuitive understanding humans have about the space of possible interactions, and the ease with which they can generalize this understanding to previously unseen scenes, we develop an approach for learning ‘visual affordances’. Given an input image of a scene, we infer a distribution over plausible future states that can be achieved via interactions with it. To allow predicting diverse plausible futures, we discretize the space of continuous images with a VQ-VAE and use a Transformer-based model to learn a conditional distribution in the latent embedding space. We show that these models can be trained using large-scale and diverse passive data, and that the learned models exhibit compositional generalization to diverse objects beyond the training distribution. We evaluate the quality and diversity of the generations, and demonstrate how the trained affordance model can be used for guiding exploration during visual goal-conditioned policy learning in robotic manipulation.

Allowing effective inference of latent vectors while training GANs can greatly increase their applicability in various downstream tasks. Recent approaches, such as ALI and BiGAN frameworks, develop methods of inference of latent variables in GANs by adversarially training an image generator along with an encoder to match two joint distributions of image and latent vector pairs. We generalize these approaches to incorporate multiple layers of feedback on reconstructions, self-supervision, and other forms of supervision based on prior or learned knowledge about the desired solutions. We achieve this by modifying the discriminator's objective to correctly identify more than two joint distributions of tuples of an arbitrary number of random variables consisting of images, latent vectors, and other variables generated through auxiliary tasks, such as reconstruction and inpainting or as outputs of suitable pre-trained models. We design a non-saturating maximization objective for the generator-encoder pair and prove that the resulting adversarial game corresponds to a global optimum that simultaneously matches all the distributions. Within our proposed framework, we introduce a novel set of techniques for providing self-supervised feedback to the model based on properties, such as patch-level correspondence and cycle consistency of reconstructions. Through comprehensive experiments, we demonstrate the efficacy, scalability, and flexibility of the proposed approach for a variety of tasks. The appendix of the paper can be found at the following link: https://drive.google.com/file/d/1i99e682CqYWMEDXlnqkqrctGLVA9viiz/view?usp=sharing

Although deep learning models have achieved state-of-the art performance on a number of vision tasks, generalization over high dimensional multi-modal data, and reliable predictive uncertainty estimation are still active areas of research. Bayesian approaches including Bayesian Neural Nets (BNNs) do not scale well to modern computer vision tasks, as they are difficult to train, and have poor generalization under dataset-shift. This motivates the need for effective ensembles which can generalize and give reliable uncertainty estimates. In this paper, we target the problem of generating effective ensembles of neural networks by encouraging diversity in prediction. We explicitly optimize a diversity inducing adversarial loss for learning the stochastic latent variables and thereby obtain diversity in the output predictions necessary for modeling multi-modal data. We evaluate our method on benchmark datasets: MNIST, CIFAR100, TinyImageNet and MIT Places 2, and compared to the most competitive baselines show significant improvements in classification accuracy, under a shift in the data distribution and in out-of-distribution detection. over 10% relative improvement in classification accuracy, over 5% relative improvement in generalizing under dataset shift, and over 5% better predictive uncertainty estimation as inferred by efficient out-of-distribution (OOD) detection.

Self-supervised goal proposal and reaching is a key component for exploration and efficient policy learning algorithms. Such a self-supervised approach without access to any oracle goal sampling distribution requires deep exploration and commitment so that long horizon plans can be efficiently discovered. In this paper, we propose an exploration framework, which learns a dynamics-aware manifold of reachable states. For a goal, our proposed method deterministically visits a state at the current frontier of reachable states (commit- ment/reaching) and then stochastically explores to reach the goal (exploration). This allocates exploration budget near the frontier of the reachable region instead of its interior. We target the challenging problem of policy learning from initial and goal states specified as images, and do not assume any access to the underlying ground-truth states of the robot and the environment. To keep track of reachable latent states, we propose a distance-conditioned reachability network that is trained to infer whether one state is reachable from another within the specified latent space distance. Given an initial state, we obtain a frontier of reachable states from that state. By incorporating a curriculum for sampling easier goals (closer to the start state) before more difficult goals, we demonstrate that the proposed self-supervised exploration algorithm, superior performance compared to existing baselines on a set of challenging robotic environments. https://sites.google.com/view/leaf-exploration

Effective planning in model-based reinforcement learning (MBRL) and model-predictive control (MPC) relies on the accuracy of the learned dynamics model. In many instances of MBRL and MPC, this model is assumed to be stationary and is periodically re-trained from scratch on state transition experience collected from the beginning of environment interactions. This implies that the time required to train the dynamics model - and the pause required between plan executions - grows linearly with the size of the collected experience. We argue that this is too slow for lifelong robot learning and propose HyperCRL, a method that continually learns the encountered dynamics in a sequence of tasks using task-conditional hypernetworks. Our method has three main attributes: first, it includes dynamics learning sessions that do not revisit training data from previous tasks, so it only needs to store the most recent fixed-size portion of the state transition experience; second, it uses fixed-capacity hypernetworks to represent non-stationary and task-aware dynamics; third, it outperforms existing continual learning alternatives that rely on fixed-capacity networks, and does competitively with baselines that remember an ever increasing coreset of past experience. We show that HyperCRL is effective in continual model-based reinforcement learning in robot locomotion and manipulation scenarios, such as tasks involving pushing and door opening. Our project website with videos is at this link http://rvl.cs.toronto.edu/blog/2020/hypercrl/

Learning from visual data opens the potential to accrue a large range of manipulation behaviors by leveraging human demonstrations without specifying each of them mathe-matically, but rather through natural task specification. In this paper, we present Learning by Watching (LbW), an algorithmic framework for policy learning through imitation from a single video specifying the task. The key insights of our method are two-fold. First, since the human arms may not have the same morphology as robot arms, our framework learns unsupervised human to robot translation to overcome the morphology mis-match issue. Second, to capture the details in salient regions that are crucial for learning state representations, our model performs unsupervised keypoint detection on the translated robot videos. The detected keypoints form a structured representation that contains semantically meaningful information and can be used directly for computing reward and policy learning. We evaluate the effectiveness of our LbW framework on five robot manipulation tasks, including reaching, pushing, sliding, coffee making, and drawer closing. Extensive experimental evaluations demonstrate that our method performs favorably against the state-of-the-art approaches. More results and analysis are available at pair.toronto.edu/lbw-kp/.

In this paper we target the problem of transferring policies across multiple environments with different dynamics parameters and motor noise variations, by introducing a framework that decouples the processes of policy learning and system identification. Efficiently transferring learned policies to an unknown environment with changes in dynamics configurations in the presence of motor noise is very important for operating robots in the real world, and our work is a novel attempt in that direction. We introduce MANGA: Method Agnostic Neural-policy Generalization and Adaptation, that trains dynamics conditioned policies and efficiently learns to estimate the dynamics parameters of the environment given off-policy state-transition rollouts in the environment. Our scheme is agnostic to the type of training method used - both reinforcement learning (RL) and imitation learning (IL) strategies can be used. We demonstrate the effectiveness of our approach by experimenting with four different MuJoCo agents and comparing against previously proposed transfer baselines.

We present a domain adaptation based generative framework for zero-shot learning. Our framework addresses the problem of domain shift between the seen and unseen class distributions in zero-shot learning and minimizes the shift by developing a generative model trained via adversarial domain adaptation. Our approach is based on end-to-end learning of the class distributions of seen classes and unseen classes. To enable the model to learn the class distributions of unseen classes, we parameterize these class distributions in terms of the class attribute information (which is available for both seen and unseen classes). This provides a very simple way to learn the class distribution of any unseen class, given only its class attribute information, and no labeled training data. Training this model with adversarial domain adaptation further provides robustness against the distribution mismatch between the data from seen and unseen classes. Our approach also provides a novel way for training neural net based classifiers to overcome the hubness problem in zero-shot learning. Through a comprehensive set of experiments, we show that our model yields superior accuracies as compared to various state-of-the-art zero shot learning models, on a variety of benchmark datasets. Code for the experiments is available at github.com/vkkhare/ZSL-ADA.

Learning effective visuomotor policies for robots purely from data is challenging, but also appealing since a learning-based system should not require manual tuning or calibration. In the case of a robot operating in a real environment the training process can be costly, time-consuming, and even dangerous since failures are common at the start of training. For this reason, it is desirable to be able to leverage simulation and off-policy data to the extent possible to train the robot. In this work, we introduce a robust framework that plans in simulation and transfers well to the real environment. Our model incorporates a gradient-descent based planning module, which, given the initial image and goal image, encodes the images to a lower dimensional latent state and plans a trajectory to reach the goal. The model, consisting of the encoder and planner modules, is first trained through a meta-learning strategy in simulation. We subsequently perform adversarial domain transfer on the encoder by using a bank of unlabelled but random images from the simulation and real environments to enable the encoder to map images from the real and simulated environments to a similarly distributed latent representation. By fine tuning the entire model (encoder + planner) with only a few real world expert demonstrations, we show successful planning performances in different navigation tasks.

Recent studies in recommender systems emphasize the importance of dealing with the cold-start problem i.e. the modeling of new users or items in the recommendation system. Meta-learning approaches have gained popularity recently in the Machine Learning (ML) community for learning representations useful for a wide-range of tasks. Inspired by the generalizable modeling prowess of Model-Agnostic Meta Learning, we design a recommendation framework that is trained to be reasonably good enough for a wide range of users. During testing, to adapt to a specific user, the model parameters are updated by a few gradient steps. We evaluate our approach on three different benchmark datasets, from Movielens, Netflix, and MyFitnessPal. Through detailed simulation studies, we show that this framework handles the user cold-start model much better than state-of-the art benchmark recommender systems. We also show that the proposed approach performs well on the task of general recommendations to non cold-start users and effectively takes care of routine and eclectic preference trends of users.

Recent advances in deep convolutional neural networks have enabled convenient diet tracking exploiting photos captured with smartphone cameras. However, most of the current diet tracking apps focus on recognizing solid foods while omitting drinks despite their negative impacts on our health when consumed without moderation. After an extensive analysis of drink images, we found that such an absence is due to the following challenges that conventional convolutional neural networks trained under the single-task learning framework cannot easily handle. First, drinks are amorphous. Second, visual cues of the drinks are often occluded and distorted by their container properties. Third, ingredients are inconspicuous because they often blend into the drink. In this work, we present a healthy drink classifier trained under a hierarchical multi-task learning framework composed of a shared residual network with hierarchically shared convolutional layers between similar tasks and task-specific fully-connected layers. The proposed structure includes two main tasks, namely sugar level classification and alcoholic drink recognition, and six auxiliary tasks, such as classification and recognition of drink name, drink type, branding logo, container transparency, container shape, and container material. We also curated a drink dataset, Drink101, composed of 101 different drinks including 11,445 images overall. Our experimental results demonstrate improved classification precision compared to single-task learning and baseline multi-task learning approaches.

Recent studies in recommendation systems emphasize the significance of modeling latent features behind temporal evolution of user preference and item state to make relevant suggestions. However, static and dynamic behaviors and trends of users and items, which highly influence the feasibility of recommendations, were not adequately addressed in previous works. In this work, we leverage the temporal and latent feature modelling capabilities of Recurrent Neural Network (RNN) and Generative Adversarial Network (GAN), respectively, to propose a Recurrent Generative Adversarial Network (RecGAN). We use customized Gated Recurrent Unit (GRU) cells to capture latent features of users and items observable from short-term and long-term temporal profiles. The modification also includes collaborative filtering mechanisms to improve the relevance of recommended items. We evaluate RecGAN using two datasets on food and movie recommendation. Results indicate that our model outperforms other baseline models irrespective of user behavior and density of training data.