Search

Vision-based robot policy learning, which maps visual inputs to actions, necessitates a holistic understanding of diverse visual tasks beyond single-task needs like classification or segmentation. Inspired by this, we introduce Theia, a vision foundation model for robot learning that distills multiple off-the-shelf vision foundation models trained on varied vision tasks. Theia's rich visual representations encode diverse visual knowledge, enhancing downstream robot learning. Extensive experiments demonstrate that Theia outperforms its teacher models and prior robot learning models using less training data and smaller model sizes. Additionally, we quantify the quality of pre-trained visual representations and hypothesize that higher entropy in feature norm distributions leads to improved robot learning performance. Code and models are available at https://github.com/bdaiinstitute/theia.

With large language models, robots can understand language more flexibly and more capable than ever before. This survey reviews and situates recent literature into a spectrum with two poles: 1) mapping between language and some manually defined formal representation of meaning, and 2) mapping between language and high-dimensional vector spaces that translate directly to low-level robot policy. Using a formal representation allows the meaning of the language to be precisely represented, limits the size of the learning problem, and leads to a framework for interpretability and formal safety guarantees. Methods that embed language and perceptual data into high-dimensional spaces avoid this manually specified symbolic structure and thus have the potential to be more general when fed enough data but require more data and computing to train. We discuss the benefits and tradeoffs of each approach and finish by providing directions for future work that achieves the best of both worlds., Comment: IJCAI 2024 Survey Track

In machine learning for sequential decision-making, an algorithmic agent learns to interact with an environment while receiving feedback in the form of a reward signal. However, in many unstructured real-world settings, such a reward signal is unknown and humans cannot reliably craft a reward signal that correctly captures desired behavior. To solve tasks in such unstructured and open-ended environments, we present Demonstration-Inferred Preference Reinforcement Learning (DIP-RL), an algorithm that leverages human demonstrations in three distinct ways, including training an autoencoder, seeding reinforcement learning (RL) training batches with demonstration data, and inferring preferences over behaviors to learn a reward function to guide RL. We evaluate DIP-RL in a tree-chopping task in Minecraft. Results suggest that the method can guide an RL agent to learn a reward function that reflects human preferences and that DIP-RL performs competitively relative to baselines. DIP-RL is inspired by our previous work on combining demonstrations and pairwise preferences in Minecraft, which was awarded a research prize at the 2022 NeurIPS MineRL BASALT competition, Learning from Human Feedback in Minecraft. Example trajectory rollouts of DIP-RL and baselines are located at https://sites.google.com/view/dip-rl., Comment: Paper accepted at The Many Facets of Preference Learning Workshop at the International Conference on Machine Learning (ICML), Honolulu, Hawaii, USA, 2023

To facilitate research in the direction of fine-tuning foundation models from human feedback, we held the MineRL BASALT Competition on Fine-Tuning from Human Feedback at NeurIPS 2022. The BASALT challenge asks teams to compete to develop algorithms to solve tasks with hard-to-specify reward functions in Minecraft. Through this competition, we aimed to promote the development of algorithms that use human feedback as channels to learn the desired behavior. We describe the competition and provide an overview of the top solutions. We conclude by discussing the impact of the competition and future directions for improvement.

We present two new classes of algorithms for efficient field integration on graphs encoding point clouds. The first class, SeparatorFactorization(SF), leverages the bounded genus of point cloud mesh graphs, while the second class, RFDiffusion(RFD), uses popular epsilon-nearest-neighbor graph representations for point clouds. Both can be viewed as providing the functionality of Fast Multipole Methods (FMMs), which have had a tremendous impact on efficient integration, but for non-Euclidean spaces. We focus on geometries induced by distributions of walk lengths between points (e.g., shortest-path distance). We provide an extensive theoretical analysis of our algorithms, obtaining new results in structural graph theory as a byproduct. We also perform exhaustive empirical evaluation, including on-surface interpolation for rigid and deformable objects (particularly for mesh-dynamics modeling), Wasserstein distance computations for point clouds, and the Gromov-Wasserstein variant.

Despite recent advancement in virtual reality technology, teleoperating a high DoF robot to complete dexterous tasks in cluttered scenes remains difficult. In this work, we propose a system that allows the user to teleoperate a Fetch robot to perform grasping in an easy and intuitive way, through exploiting the rich environment information provided by the virtual reality space. Our system has the benefit of easy transferability to different robots and different tasks, and can be used without any expert knowledge. We tested the system on a real fetch robot, and a video demonstrating the effectiveness of our system can be seen at https://youtu.be/1-xW2Bx_Cms.

We propose the Multiple View Performer (MVP) - a new architecture for 3D shape completion from a series of temporally sequential views. MVP accomplishes this task by using linear-attention Transformers called Performers. Our model allows the current observation of the scene to attend to the previous ones for more accurate infilling. The history of past observations is compressed via the compact associative memory approximating modern continuous Hopfield memory, but crucially of size independent from the history length. We compare our model with several baselines for shape completion over time, demonstrating the generalization gains that MVP provides. To the best of our knowledge, MVP is the first multiple view voxel reconstruction method that does not require registration of multiple depth views and the first causal Transformer based model for 3D shape completion., Comment: 6 pages, 2 pages of references, 6 figures, 3 tables

Cryptocurrencies have garnered a lot of attention by governments and internet enthusiasts over the past three years. These currencies are celebrated for their security and speedy transactions in a modern era of digital commerce. Bitcoin was the first of these currencies to gain a large advantage over subsequent iterations. Bitcoin was first conceived by Satoshi Nakamoto who mentioned the concept of a cryptocurrency in his paper titled Bitcoin. It featured new concepts such as proof of work and transactions which utilized hash based encryption. One particular alternative cryptocurrency is known as Litecoin. Backed by a memory intensive algorithm known as Scrypt, many cryptocurrency enthusiasts have decided to celebrate this particular coin. Scrypt expands on Bitcoin's proof of work algorithm by adding the amount of work it takes to commit a transaction within the Litecoin network. Scrypt forces more work on the device that is being used to perform the algorithm by making frequent memory requests. This makes it difficult to create specialized hardware to create new coins and to commit transactions due to the nature of memory intensive applications., Comment: Published in 2014

Providing mobile robots with the ability to manipulate objects has, despite decades of research, remained a challenging problem. The problem is approachable in constrained environments where there is ample prior knowledge of the environment layout and manipulatable objects. The challenge is in building systems that scale beyond specific situational instances and gracefully operate in novel conditions. In the past, researchers used heuristic and simple rule-based strategies to accomplish tasks such as scene segmentation or reasoning about occlusion. These heuristic strategies work in constrained environments where a roboticist can make simplifying assumptions about everything from the geometries of the objects to be interacted with, level of clutter, camera position, lighting, and a myriad of other relevant variables. The work in this thesis will demonstrate how to build a system for robotic mobile manipulation that is robust to changes in these variables. This robustness will be enabled by recent simultaneous advances in the fields of big data, deep learning, and simulation. The ability of simulators to create realistic sensory data enables the generation of massive corpora of labeled training data for various grasping and navigation-based tasks. It is now possible to build systems that work in the real world trained using deep learning entirely on synthetic data. The ability to train and test on synthetic data allows for quick iterative development of new perception, planning and grasp execution algorithms that work in many environments., Comment: PhD thesis

While both navigation and manipulation are challenging topics in isolation, many tasks require the ability to both navigate and manipulate in concert. To this end, we propose a mobile manipulation system that leverages novel navigation and shape completion methods to manipulate an object with a mobile robot. Our system utilizes uncertainty in the initial estimation of a manipulation target to calculate a predicted next-best-view. Without the need of localization, the robot then uses the predicted panoramic view at the next-best-view location to navigate to the desired location, capture a second view of the object, create a new model that predicts the shape of object more accurately than a single image alone, and uses this model for grasp planning. We show that the system is highly effective for mobile manipulation tasks through simulation experiments using real world data, as well as ablations on each component of our system., Comment: 6 pages, 2 pages of references, 5 figures, 5 tables

Pole-swapping algorithms are generalizations of bulge-chasing algorithms for the generalized eigenvalue problem. Structure-preserving pole-swapping algorithms for the palindromic and alternating eigenvalue problems, which arise in control theory, are derived. A refinement step that guarantees backward stability of the algorithms is included. This refinement can also be applied to bulge-chasing algorithms that had been introduced previously, thereby guaranteeing their backward stability in all cases.

Pole-swapping algorithms, which are generalizations of the QZ algorithm for the generalized eigenvalue problem, are studied. A new modular (and therefore more flexible) convergence theory that applies to all pole-swapping algorithms is developed. A key component of all such algorithms is a procedure that swaps two adjacent eigenvalues in a triangular pencil. An improved swapping routine is developed, and its superiority over existing methods is demonstrated by a backward error analysis and numerical tests. The modularity of the new convergence theory and the generality of the pole-swapping approach shed new light on bi-directional chasing algorithms, optimally packed shifts, and bulge pencils, and allow the design of novel algorithms.

This work describes a new human-in-the-loop (HitL) assistive grasping system for individuals with varying levels of physical capabilities. We investigated the feasibility of using four potential input devices with our assistive grasping system interface, using able-bodied individuals to define a set of quantitative metrics that could be used to assess an assistive grasping system. We then took these measurements and created a generalized benchmark for evaluating the effectiveness of any arbitrary input device into a HitL grasping system. The four input devices were a mouse, a speech recognition device, an assistive switch, and a novel sEMG device developed by our group that was connected either to the forearm or behind the ear of the subject. These preliminary results provide insight into how different interface devices perform for generalized assistive grasping tasks and also highlight the potential of sEMG based control for severely disabled individuals., Comment: 8 pages, 21 figures

This work provides an architecture that incorporates depth and tactile information to create rich and accurate 3D models useful for robotic manipulation tasks. This is accomplished through the use of a 3D convolutional neural network (CNN). Offline, the network is provided with both depth and tactile information and trained to predict the object's geometry, thus filling in regions of occlusion. At runtime, the network is provided a partial view of an object. Tactile information is acquired to augment the captured depth information. The network can then reason about the object's geometry by utilizing both the collected tactile and depth information. We demonstrate that even small amounts of additional tactile information can be incredibly helpful in reasoning about object geometry. This is particularly true when information from depth alone fails to produce an accurate geometric prediction. Our method is benchmarked against and outperforms other visual-tactile approaches to general geometric reasoning. We also provide experimental results comparing grasping success with our method.