Daniel Seita

Ayano Hiranaka, Ya-Chuan Hsu, Stefanos Nikolaidis et al.

AI assistants in human-AI collaboration often correct suboptimal human actions through behavioral feedback (e.g., alerts or steering-wheel nudges in assistive driving). Such interventions can mitigate immediate errors, but long-term improvement requires addressing the underlying misconceptions that cause repeated mistakes. We introduce SENSEI, a framework that infers user misconceptions from interaction behavior and provides targeted, minimal yet sufficient suggestions to correct them. Our approach departs from action- or trajectory-level interventions by operating over a structured knowledge representation to localize and correct the sources of erroneous behavior. Across three long-horizon tasks with diverse misconceptions and corresponding behaviors, SENSEI demonstrates zero-shot compositional generalization, disentangling multiple overlapping misconceptions despite training only on single-misconception cases. A user study further shows that our method identifies real human misconceptions and provides effective guidance that improves long-horizon task performance, successfully correcting $90\%$ of student misconceptions. Code and project page are available at https://misoshiruseijin.github.io/SENSEI/.

Kuanning Wang, Ke Fan, Yuqian Fu et al.

We present OCRA, an Object-Centric framework for video-based human-to-Robot Action transfer that learns directly from human demonstration videos to enable robust manipulation. Object-centric learning emphasizes task-relevant objects and their interactions while filtering out irrelevant background, providing a natural and scalable way to teach robots. OCRA leverages multi-view RGB videos, the state-of-the-art 3D foundation model VGGT, and advanced detection and segmentation models to reconstruct object-centric 3D point clouds, capturing rich interactions between objects. To handle properties not easily perceived by vision alone, we incorporate tactile priors via a large-scale dataset of over one million tactile images. These 3D and tactile priors are fused through a multimodal module (ResFiLM) and fed into a Diffusion Policy to generate robust manipulation actions. Extensive experiments on both vision-only and visuo-tactile tasks show that OCRA significantly outperforms existing baselines and ablations, demonstrating its effectiveness for learning from human demonstration videos.

Kuanning Wang, Ke Fan, Chenhao Qiu et al.

Robust robotic manipulation requires not only predicting how the scene evolves over time, but also recognizing task-relevant objects in complex scenes. However, existing VLA models face two limitations. They typically act only on the current frame, while future prediction and object-aware reasoning are often learned in separate latent spaces. We propose OFlow (injecting Object-Aware Temporal Flow Matching into VLAs), a framework that addresses both limitations by unifying temporal foresight and object-aware reasoning in a shared semantic latent space. Our method forecasts future latents with temporal flow matching, factorizes them into object-aware representations that emphasize physically relevant cues while filtering task-irrelevant variation, and conditions continuous action generation on these predictions. By integrating OFlow into VLA pipelines, our method enables more reliable control under distribution shifts. Extensive experiments across LIBERO, LIBERO-Plus, MetaWorld, and SimplerEnv benchmarks and real-world tasks demonstrate that object-aware foresight consistently enhances robustness and success.

Haozhe Lou, Mingtong Zhang, Haoran Geng et al.

Simulation provides a cost-effective and flexible platform for data generation and policy learning to develop robotic systems. However, bridging the gap between simulation and real-world dynamics remains a significant challenge, especially in physical parameter identification. In this work, we introduce a real-to-sim-to-real engine that leverages the Gaussian Splat representations to build a differentiable engine, enabling object mass identification from real-world visual observations and robot control signals, while enabling grasping policy learning simultaneously. Through optimizing the mass of the manipulated object, our method automatically builds high-fidelity and physically plausible digital twins. Additionally, we propose a novel approach to train force-aware grasping policies from limited data by transferring feasible human demonstrations into simulated robot demonstrations. Through comprehensive experiments, we demonstrate that our engine achieves accurate and robust performance in mass identification across various object geometries and mass values. Those optimized mass values facilitate force-aware policy learning, achieving superior and high performance in object grasping, effectively reducing the sim-to-real gap.

I-Chun Arthur Liu, Sicheng He, Daniel Seita et al.

Bimanual manipulation is critical to many robotics applications. In contrast to single-arm manipulation, bimanual manipulation tasks are challenging due to higher-dimensional action spaces. Prior works leverage large amounts of data and primitive actions to address this problem, but may suffer from sample inefficiency and limited generalization across various tasks. To this end, we propose VoxAct-B, a language-conditioned, voxel-based method that leverages Vision Language Models (VLMs) to prioritize key regions within the scene and reconstruct a voxel grid. We provide this voxel grid to our bimanual manipulation policy to learn acting and stabilizing actions. This approach enables more efficient policy learning from voxels and is generalizable to different tasks. In simulation, we show that VoxAct-B outperforms strong baselines on fine-grained bimanual manipulation tasks. Furthermore, we demonstrate VoxAct-B on real-world $\texttt{Open Drawer}$ and $\texttt{Open Jar}$ tasks using two UR5s. Code, data, and videos are available at https://voxact-b.github.io.

Ethan Foong, Yunshuang Li, Hao Jiang et al.

Dexterous robot hands offer rich opportunities for multifunctional manipulation, where a robot must execute multiple skills in sequence while maintaining control over previously grasped objects. Most prior work in dexterous manipulation focuses on single-object, single-skill tasks. In contrast, our insight is that many sequential tasks require resource-aware grasps that conserve fingers for future actions. In this paper, we study sequential grasp-conditioned dexterous manipulation, where a robot first grasps an object and then performs a second, distinct manipulation subtask while preserving the initial grasp. We introduce HANDFUL, a learning framework that models finger usage as a limited resource and encourages exploration of resource-aware grasps through finger-level contact rewards. These grasps are subsequently selected for downstream tasks via curriculum-based policy learning. We further propose HANDFUL-Bench, a simulation benchmark that introduces sequential dexterous manipulation tasks across multiple secondsubtask objectives, including pushing, pulling, and pressing, under a shared grasp-conditioned setup. Extensive simulation results demonstrate that prioritizing resource-aware grasps improves second-subtask success and robustness compared to a baseline that greedily optimizes the initial grasp before attempting the second subtask. We additionally validate our approach on a real dexterous LEAP hand. Together, this work establishes resource-aware grasp planning as a key principle for multifunctional dexterous manipulation. Supplementary material is available on our website: https://handful-dex.github.io.

Yunshuang Li, Yiyang Ling, Gaurav S. Sukhatme et al.

Nonprehensile manipulation, such as pushing and pulling, enables robots to move, align, or reposition objects that may be difficult to grasp due to their geometry, size, or relationship to the robot or the environment. Much of the existing work in nonprehensile manipulation relies on parallel-jaw grippers or tools such as rods and spatulas. In contrast, multi-fingered dexterous hands offer richer contact modes and versatility for handling diverse objects to provide stable support over the objects, which compensates for the difficulty of modeling the dynamics of nonprehensile manipulation. Therefore, we propose Geometry-aware Dexterous Pushing and Pulling(GD2P) for nonprehensile manipulation with dexterous robotic hands. We study pushing and pulling by framing the problem as synthesizing and learning pre-contact dexterous hand poses that lead to effective manipulation. We generate diverse hand poses via contact-guided sampling, filter them using physics simulation, and train a diffusion model conditioned on object geometry to predict viable poses. At test time, we sample hand poses and use standard motion planners to select and execute pushing and pulling actions. We perform extensive real-world experiments with an Allegro Hand and a LEAP Hand, demonstrating that GD2P offers a scalable route for generating dexterous nonprehensile manipulation motions with its applicability to different hand morphologies. Our project website is available at: geodex2p.github.io.

Zhonghao Shi, Enyu Zhao, Nathaniel Dennler et al.

Real-time human perception is crucial for effective human-robot interaction (HRI). Large vision-language models (VLMs) offer promising generalizable perceptual capabilities but often suffer from high latency, which negatively impacts user experience and limits VLM applicability in real-world scenarios. To systematically study VLM capabilities in human perception for HRI and performance-latency trade-offs, we introduce HRIBench, a visual question-answering (VQA) benchmark designed to evaluate VLMs across a diverse set of human perceptual tasks critical for HRI. HRIBench covers five key domains: (1) non-verbal cue understanding, (2) verbal instruction understanding, (3) human-robot object relationship understanding, (4) social navigation, and (5) person identification. To construct HRIBench, we collected data from real-world HRI environments to curate questions for non-verbal cue understanding, and leveraged publicly available datasets for the remaining four domains. We curated 200 VQA questions for each domain, resulting in a total of 1000 questions for HRIBench. We then conducted a comprehensive evaluation of both state-of-the-art closed-source and open-source VLMs (N=11) on HRIBench. Our results show that, despite their generalizability, current VLMs still struggle with core perceptual capabilities essential for HRI. Moreover, none of the models within our experiments demonstrated a satisfactory performance-latency trade-off suitable for real-time deployment, underscoring the need for future research on developing smaller, low-latency VLMs with improved human perception capabilities. HRIBench and our results can be found in this Github repository: https://github.com/interaction-lab/HRIBench.

Minjune Hwang, Yigit Korkmaz, Daniel Seita et al.

Preference-based reward learning is widely used for shaping agent behavior to match a user's preference, yet its sparse binary feedback makes it especially vulnerable to causal confusion. The learned reward often latches onto spurious features that merely co-occur with preferred trajectories during training, collapsing when those correlations disappear or reverse at test time. We introduce ReCouPLe, a lightweight framework that uses natural language rationales to provide the missing causal signal. Each rationale is treated as a guiding projection axis in an embedding space, training the model to score trajectories based on features aligned with that axis while de-emphasizing context that is unrelated to the stated reason. Because the same rationales (e.g., "avoids collisions", "completes the task faster") can appear across multiple tasks, ReCouPLe naturally reuses the same causal direction whenever tasks share semantics, and transfers preference knowledge to novel tasks without extra data or language-model fine-tuning. Our learned reward model can ground preferences on the articulated reason, aligning better with user intent and generalizing beyond spurious features. ReCouPLe outperforms baselines by up to 1.5x in reward accuracy under distribution shifts, and 2x in downstream policy performance in novel tasks. We have released our code at https://github.com/mj-hwang/ReCouPLe

Jason Chen, I-Chun Arthur Liu, Gaurav Sukhatme et al.

Bimanual robot learning from demonstrations is fundamentally limited by the cost and narrow visual diversity of real-world data, which constrains policy robustness across viewpoints, object configurations, and embodiments. We present Canny-guided Robot Data Generation using Video Diffusion Transformers (CRAFT), a video diffusion-based framework for scalable bimanual demonstration generation that synthesizes temporally coherent manipulation videos while producing action labels. By conditioning video diffusion on edge-based structural cues extracted from simulator-generated trajectories, CRAFT produces physically plausible trajectory variations and supports a unified augmentation pipeline spanning object pose changes, camera viewpoints, lighting and background variations, cross-embodiment transfer, and multi-view synthesis. We leverage a pre-trained video diffusion model to convert simulated videos, along with action labels from the simulation trajectories, into action-consistent demonstrations. Starting from only a few real-world demonstrations, CRAFT generates a large, visually diverse set of photorealistic training data, bypassing the need to replay demonstrations on the real robot (Sim2Real). Across simulated and real-world bimanual tasks, CRAFT improves success rates over existing augmentation strategies and straightforward data scaling, demonstrating that diffusion-based video generation can substantially expand demonstration diversity and improve generalization for dual-arm manipulation tasks. Our project website is available at: https://craftaug.github.io/

Wei Chow, Jiageng Mao, Boyi Li et al.

Understanding the physical world is a fundamental challenge in embodied AI, critical for enabling agents to perform complex tasks and operate safely in real-world environments. While Vision-Language Models (VLMs) have shown great promise in reasoning and task planning for embodied agents, their ability to comprehend physical phenomena remains extremely limited. To close this gap, we introduce PhysBench, a comprehensive benchmark designed to evaluate VLMs' physical world understanding capability across a diverse set of tasks. PhysBench contains 10,002 entries of interleaved video-image-text data, categorized into four major domains: physical object properties, physical object relationships, physical scene understanding, and physics-based dynamics, further divided into 19 subclasses and 8 distinct capability dimensions. Our extensive experiments, conducted on 75 representative VLMs, reveal that while these models excel in common-sense reasoning, they struggle with understanding the physical world -- likely due to the absence of physical knowledge in their training data and the lack of embedded physical priors. To tackle the shortfall, we introduce PhysAgent, a novel framework that combines the generalization strengths of VLMs with the specialized expertise of vision models, significantly enhancing VLMs' physical understanding across a variety of tasks, including an 18.4\% improvement on GPT-4o. Furthermore, our results demonstrate that enhancing VLMs' physical world understanding capabilities can help embodied agents such as MOKA. We believe that PhysBench and PhysAgent offer valuable insights and contribute to bridging the gap between VLMs and physical world understanding.

Hao Jiang, Yue Wu, Yue Wang et al.

Dexterous hands enable concurrent prehensile and nonprehensile manipulation, such as holding one object while interacting with another, a capability essential for everyday tasks yet underexplored in robotics. Learning such long-horizon, contact-rich multi-stage behaviors is challenging because demonstrations are expensive to collect and end-to-end policies require substantial data to generalize across varied object geometries and placements. We present DexMulti, a sample-efficient approach for real-world dexterous multi-task manipulation that decomposes demonstrations into object-centric skills with well-defined temporal boundaries. Rather than learning monolithic policies, our method retrieves demonstrated skills based on current object geometry, aligns them to the observed object state using an uncertainty-aware estimator that tracks centroid and yaw, and executes them via a retrieve-align-execute paradigm. We evaluate on three multi-stage tasks requiring concurrent manipulation (Grasp + Pull, Grasp + Open, and Grasp + Grasp) across two dexterous hands (Allegro and LEAP) in over 1,000 real-world trials. Our approach achieves an average success rate of 66% on training objects with only 3-4 demonstrations per object, outperforming diffusion policy baselines by 2-3x while requiring far fewer demonstrations. Results demonstrate robust generalization to held-out objects and spatial variations up to +/-25 cm.

Faiz Aladin, Ashwin Balasubramanian, Lars Lindemann et al.

Reachability analysis has become increasingly important in robotics to distinguish safe from unsafe states. Unfortunately, existing reachability and safety analysis methods often fall short, as they typically require known system dynamics or large datasets to estimate accurate system models, are computationally expensive, and assume full state information. A recent method, called MORALS, aims to address these shortcomings by using topological tools to estimate Regions of Attraction (ROA) in a low-dimensional latent space. However, MORALS still relies on full state knowledge and has not been studied when only sensor measurements are available. This paper presents Visual Morse Graph-Aided Estimation of Regions of Attraction in a Learned Latent Space (V-MORALS). V-MORALS takes in a dataset of image-based trajectories of a system under a given controller, and learns a latent space for reachability analysis. Using this learned latent space, our method is able to generate well-defined Morse Graphs, from which we can compute ROAs for various systems and controllers. V-MORALS provides capabilities similar to the original MORALS architecture without relying on state knowledge, and using only high-level sensor data. Our project website is at: https://v-morals.onrender.com.

Enyu Zhao, Vedant Raval, Hejia Zhang et al.

Vision-Language Models (VLMs) have revolutionized artificial intelligence and robotics due to their commonsense reasoning capabilities. In robotic manipulation, VLMs are used primarily as high-level planners, but recent work has also studied their lower-level reasoning ability, which refers to making decisions about precise robot movements. However, the community currently lacks a clear and common benchmark that can evaluate how well VLMs can aid low-level reasoning in robotics. Consequently, we propose a novel benchmark, ManipBench, to evaluate the low-level robot manipulation reasoning capabilities of VLMs across various dimensions, including how well they understand object-object interactions and deformable object manipulation. We extensively test 33 representative VLMs across 10 model families on our benchmark, including variants to test different model sizes. Our evaluation shows that the performance of VLMs significantly varies across tasks, and there is a strong correlation between this performance and trends in our real-world manipulation tasks. It also shows that there remains a significant gap between these models and human-level understanding. See our website at: https://manipbench.github.io.

Yiyang Ling, Karan Owalekar, Oluwatobiloba Adesanya et al.

Motion planning involves determining a sequence of robot configurations to reach a desired pose, subject to movement and safety constraints. Traditional motion planning finds collision-free paths, but this is overly restrictive in clutter, where it may not be possible for a robot to accomplish a task without contact. In addition, contacts range from relatively benign (e.g. brushing a soft pillow) to more dangerous (e.g. toppling a glass vase), making it difficult to characterize which may be acceptable. In this paper, we propose IMPACT, a novel motion planning framework that uses Vision-Language Models (VLMs) to infer environment semantics, identifying which parts of the environment can best tolerate contact based on object properties and locations. Our approach generates an anisotropic cost map that encodes directional push safety. We pair this map with a contact-aware A* planner to find stable contact-rich paths. We perform experiments using 20 simulation and 10 real-world scenes and assess using task success rate, object displacements, and feedback from human evaluators. Our results over 3200 simulation and 200 real-world trials suggest that IMPACT enables efficient contact-rich motion planning in cluttered settings while outperforming alternative methods and ablations. Our project website is available at https://impact-planning.github.io/.

Siheng Zhao, Jiageng Mao, Wei Chow et al.

We introduce RoLA, a framework that transforms any in-the-wild image into an interactive, physics-enabled robotic environment. Unlike previous methods, RoLA operates directly on a single image without requiring additional hardware or digital assets. Our framework democratizes robotic data generation by producing massive visuomotor robotic demonstrations within minutes from a wide range of image sources, including camera captures, robotic datasets, and Internet images. At its core, our approach combines a novel method for single-view physical scene recovery with an efficient visual blending strategy for photorealistic data collection. We demonstrate RoLA's versatility across applications like scalable robotic data generation and augmentation, robot learning from Internet images, and single-image real-to-sim-to-real systems for manipulators and humanoids. Video results are available at https://sihengz02.github.io/RoLA .

I-Chun Arthur Liu, Jason Chen, Gaurav Sukhatme et al.

Learning bimanual manipulation is challenging due to its high dimensionality and tight coordination required between two arms. Eye-in-hand imitation learning, which uses wrist-mounted cameras, simplifies perception by focusing on task-relevant views. However, collecting diverse demonstrations remains costly, motivating the need for scalable data augmentation. While prior work has explored visual augmentation in single-arm settings, extending these approaches to bimanual manipulation requires generating viewpoint-consistent observations across both arms and producing corresponding action labels that are both valid and feasible. In this work, we propose Diffusion for COordinated Dual-arm Data Augmentation (D-CODA), a method for offline data augmentation tailored to eye-in-hand bimanual imitation learning that trains a diffusion model to synthesize novel, viewpoint-consistent wrist-camera images for both arms while simultaneously generating joint-space action labels. It employs constrained optimization to ensure that augmented states involving gripper-to-object contacts adhere to constraints suitable for bimanual coordination. We evaluate D-CODA on 5 simulated and 3 real-world tasks. Our results across 2250 simulation trials and 300 real-world trials demonstrate that it outperforms baselines and ablations, showing its potential for scalable data augmentation in eye-in-hand bimanual manipulation. Our project website is at: https://dcodaaug.github.io/D-CODA/.

Zeyu Shangguan, Daniel Seita, Mohammad Rostami

Cross-modal feature extraction and integration have led to steady performance improvements in few-shot learning tasks due to generating richer features. However, existing multi-modal object detection (MM-OD) methods degrade when facing significant domain-shift and are sample insufficient. We hypothesize that rich text information could more effectively help the model to build a knowledge relationship between the vision instance and its language description and can help mitigate domain shift. Specifically, we study the Cross-Domain few-shot generalization of MM-OD (CDMM-FSOD) and propose a meta-learning based multi-modal few-shot object detection method that utilizes rich text semantic information as an auxiliary modality to achieve domain adaptation in the context of FSOD. Our proposed network contains (i) a multi-modal feature aggregation module that aligns the vision and language support feature embeddings and (ii) a rich text semantic rectify module that utilizes bidirectional text feature generation to reinforce multi-modal feature alignment and thus to enhance the model's language understanding capability. We evaluate our model on common standard cross-domain object detection datasets and demonstrate that our approach considerably outperforms existing FSOD methods.

Siddharth Srikanth, Freddie Liang, Sophie Hsu et al.

Vision-Language-Action (VLA) models have significant potential to enable general-purpose robotic systems for a range of vision-language tasks. However, the performance of VLA-based robots is highly sensitive to the precise wording of language instructions, and it remains difficult to predict when such robots will fail. To improve the robustness of VLAs to different wordings, we present Q-DIG (Quality Diversity for Diverse Instruction Generation), which performs red-teaming by scalably identifying diverse natural language task descriptions that induce failures while remaining task-relevant. Q-DIG integrates Quality Diversity (QD) techniques with Vision-Language Models (VLMs) to generate a broad spectrum of adversarial instructions that expose meaningful vulnerabilities in VLA behavior. Our results across multiple simulation benchmarks show that Q-DIG finds more diverse and meaningful failure modes compared to baseline methods, and that fine-tuning VLAs on the generated instructions improves task success rates. Furthermore, results from a user study highlight that Q-DIG generates prompts judged to be more natural and human-like than those from baselines. Finally, real-world evaluations of Q-DIG prompts show results consistent with simulation, and fine-tuning VLAs on the generated prompts further success rates on unseen instructions. Together, these findings suggest that Q-DIG is a promising approach for identifying vulnerabilities and improving the robustness of VLA-based robots. Our anonymous project website is at qdigvla.github.io.

Toan Nguyen, Weiduo Yuan, Songlin Wei et al.

In-context imitation learning enables robots to adapt to new tasks from a small number of demonstrations without additional training. However, existing approaches typically condition only on state-action trajectories and lack explicit representations of task intent. This limitation hinders performance in complex and ambiguous task settings where the same actions may be consistent with different objectives. To address this, we present In-Context Imitation Learning with Visual Reasoning (ICLR), a novel framework that augments demonstration prompts with structured visual reasoning traces representing anticipated future robot trajectories in image space. ICLR also jointly learns to generate reasoning traces and low-level actions within a unified autoregressive transformer, enabling the model to mimic not only action prediction but also the reasoning process that leads to those actions. We extensively evaluate ICLR in both simulation and real-world manipulation tasks and demonstrate consistent improvements in success rates and generalization to unseen tasks and novel object configurations compared to other in-context imitation learning methods. These results suggest that incorporating embodied visual reasoning represents a promising direction for enhancing the robustness and generalization of robotic in-context learning systems.

Nikita Sarawgi, Omey M. Manyar, Fan Wang et al.

Robotic bin packing is widely deployed in warehouse automation, with current systems achieving robust performance through heuristic and learning-based strategies. These systems must balance compact placement with rapid execution, where selecting alternative items or reorienting them can improve space utilization but introduce additional time. We propose a selection-based formulation that explicitly reasons over this trade-off: at each step, the robot evaluates multiple candidate actions, weighing expected packing benefit against estimated operational time. This enables time-aware strategies that selectively accept increased operational time when it yields meaningful spatial improvements. Our method, STEP (Space-Time Efficient Packing), uses a preference-conditioned, Transformer-based reinforcement learning policy, and allows generalization across candidate set sizes and integration with standard placement modules. It achieves a 44% reduction in operational time without compromising packing density. Additional material is available at https://step-packing.github.io.

Kuanning Wang, Yongchong Gu, Yuqian Fu et al.

Scooping items with tools such as spoons and ladles is common in daily life, ranging from assistive feeding to retrieving items from environmental disaster sites. However, developing a general and autonomous robotic scooping policy is challenging since it requires reasoning about complex tool-object interactions. Furthermore, scooping often involves manipulating deformable objects, such as granular media or liquids, which is challenging due to their infinite-dimensional configuration spaces and complex dynamics. We propose a method, SCOOP'D, which uses simulation from OmniGibson (built on NVIDIA Omniverse) to collect scooping demonstrations using algorithmic procedures that rely on privileged state information. Then, we use generative policies via diffusion to imitate demonstrations from observational input. We directly apply the learned policy in diverse real-world scenarios, testing its performance on various item quantities, item characteristics, and container types. In zero-shot deployment, our method demonstrates promising results across 465 trials in diverse scenarios, including objects of different difficulty levels that we categorize as "Level 1" and "Level 2." SCOOP'D outperforms all baselines and ablations, suggesting that this is a promising approach to acquiring robotic scooping skills. Project page is at https://scoopdiff.github.io/.

Jason Chen, I-Chun Arthur Liu, Gaurav Sukhatme et al.

Training robust bimanual manipulation policies via imitation learning requires demonstration data with broad coverage over robot poses, contacts, and scene contexts. However, collecting diverse and precise real-world demonstrations is costly and time-consuming, which hinders scalability. Prior works have addressed this with data augmentation, typically for either eye-in-hand (wrist camera) setups with RGB inputs or for generating novel images without paired actions, leaving augmentation for eye-to-hand (third-person) RGB-D training with new action labels less explored. In this paper, we propose Synthetic Robot Pose Generation for RGB-D Bimanual Data Augmentation (ROPA), an offline imitation learning data augmentation method that fine-tunes Stable Diffusion to synthesize third-person RGB and RGB-D observations of novel robot poses. Our approach simultaneously generates corresponding joint-space action labels while employing constrained optimization to enforce physical consistency through appropriate gripper-to-object contact constraints in bimanual scenarios. We evaluate our method on 5 simulated and 3 real-world tasks. Our results across 2625 simulation trials and 300 real-world trials demonstrate that ROPA outperforms baselines and ablations, showing its potential for scalable RGB and RGB-D data augmentation in eye-to-hand bimanual manipulation. Our project website is available at: https://ropaaug.github.io/.

Zeyu Shangguan, Daniel Seita, Mohammad Rostami

Advancements in cross-modal feature extraction and integration have significantly enhanced performance in few-shot learning tasks. However, current multi-modal object detection (MM-OD) methods often experience notable performance degradation when encountering substantial domain shifts. We propose that incorporating rich textual information can enable the model to establish a more robust knowledge relationship between visual instances and their corresponding language descriptions, thereby mitigating the challenges of domain shift. Specifically, we focus on the problem of Cross-Domain Multi-Modal Few-Shot Object Detection (CDMM-FSOD) and introduce a meta-learning-based framework designed to leverage rich textual semantics as an auxiliary modality to achieve effective domain adaptation. Our new architecture incorporates two key components: (i) A multi-modal feature aggregation module, which aligns visual and linguistic feature embeddings to ensure cohesive integration across modalities. (ii) A rich text semantic rectification module, which employs bidirectional text feature generation to refine multi-modal feature alignment, thereby enhancing understanding of language and its application in object detection. We evaluate the proposed method on common cross-domain object detection benchmarks and demonstrate that it significantly surpasses existing few-shot object detection approaches.

Vincent Lim, Huang Huang, Lawrence Yunliang Chen et al.

This paper introduces the task of {\em Planar Robot Casting (PRC)}: where one planar motion of a robot arm holding one end of a cable causes the other end to slide across the plane toward a desired target. PRC allows the cable to reach points beyond the robot workspace and has applications for cable management in homes, warehouses, and factories. To efficiently learn a PRC policy for a given cable, we propose Real2Sim2Real, a self-supervised framework that automatically collects physical trajectory examples to tune parameters of a dynamics simulator using Differential Evolution, generates many simulated examples, and then learns a policy using a weighted combination of simulated and physical data. We evaluate Real2Sim2Real with three simulators, Isaac Gym-segmented, Isaac Gym-hybrid, and PyBullet, two function approximators, Gaussian Processes and Neural Networks (NNs), and three cables with differing stiffness, torsion, and friction. Results with 240 physical trials suggest that the PRC policies can attain median error distance (as % of cable length) ranging from 8% to 14%, outperforming baselines and policies trained on only real or only simulated examples. Code, data, and videos are available at https://tinyurl.com/robotcast.

Daniel Seita, Abhinav Gopal, Zhao Mandi et al.

Deep reinforcement learning (RL) has shown great empirical successes, but suffers from brittleness and sample inefficiency. A potential remedy is to use a previously-trained policy as a source of supervision. In this work, we refer to these policies as teachers and study how to transfer their expertise to new student policies by focusing on data usage. We propose a framework, Data CUrriculum for Reinforcement learning (DCUR), which first trains teachers using online deep RL, and stores the logged environment interaction history. Then, students learn by running either offline RL or by using teacher data in combination with a small amount of self-generated data. DCUR's central idea involves defining a class of data curricula which, as a function of training time, limits the student to sampling from a fixed subset of the full teacher data. We test teachers and students using state-of-the-art deep RL algorithms across a variety of data curricula. Results suggest that the choice of data curricula significantly impacts student learning, and that it is beneficial to limit the data during early training stages while gradually letting the data availability grow over time. We identify when the student can learn offline and match teacher performance without relying on specialized offline RL algorithms. Furthermore, we show that collecting a small fraction of online data provides complementary benefits with the data curriculum. Supplementary material is available at https://tinyurl.com/teach-dcur.

Ryan Hoque, Ashwin Balakrishna, Carl Putterman et al.

Corrective interventions while a robot is learning to automate a task provide an intuitive method for a human supervisor to assist the robot and convey information about desired behavior. However, these interventions can impose significant burden on a human supervisor, as each intervention interrupts other work the human is doing, incurs latency with each context switch between supervisor and autonomous control, and requires time to perform. We present LazyDAgger, which extends the interactive imitation learning (IL) algorithm SafeDAgger to reduce context switches between supervisor and autonomous control. We find that LazyDAgger improves the performance and robustness of the learned policy during both learning and execution while limiting burden on the supervisor. Simulation experiments suggest that LazyDAgger can reduce context switches by an average of 60% over SafeDAgger on 3 continuous control tasks while maintaining state-of-the-art policy performance. In physical fabric manipulation experiments with an ABB YuMi robot, LazyDAgger reduces context switches by 60% while achieving a 60% higher success rate than SafeDAgger at execution time.

Ryan Hoque, Daniel Seita, Ashwin Balakrishna et al.

Robotic fabric manipulation has applications in home robotics, textiles, senior care and surgery. Existing fabric manipulation techniques, however, are designed for specific tasks, making it difficult to generalize across different but related tasks. We build upon the Visual Foresight framework to learn fabric dynamics that can be efficiently reused to accomplish different sequential fabric manipulation tasks with a single goal-conditioned policy. We extend our earlier work on VisuoSpatial Foresight (VSF), which learns visual dynamics on domain randomized RGB images and depth maps simultaneously and completely in simulation. In this earlier work, we evaluated VSF on multi-step fabric smoothing and folding tasks against 5 baseline methods in simulation and on the da Vinci Research Kit (dVRK) surgical robot without any demonstrations at train or test time. A key finding was that depth sensing significantly improves performance: RGBD data yields an 80% improvement in fabric folding success rate in simulation over pure RGB data. In this work, we vary 4 components of VSF, including data generation, visual dynamics model, cost function, and optimization procedure. Results suggest that training visual dynamics models using longer, corner-based actions can improve the efficiency of fabric folding by 76% and enable a physical sequential fabric folding task that VSF could not previously perform with 90% reliability. Code, data, videos, and supplementary material are available at https://sites.google.com/view/fabric-vsf/.

Minho Hwang, Jeffrey Ichnowski, Brijen Thananjeyan et al.

Peg transfer is a well-known surgical training task in the Fundamentals of Laparoscopic Surgery (FLS). While human sur-geons teleoperate robots such as the da Vinci to perform this task with high speed and accuracy, it is challenging to automate. This paper presents a novel system and control method using a da Vinci Research Kit (dVRK) surgical robot and a Zivid depth sensor, and a human subjects study comparing performance on three variants of the peg-transfer task: unilateral, bilateral without handovers, and bilateral with handovers. The system combines 3D printing, depth sensing, and deep learning for calibration with a new analytic inverse kinematics model and a time-minimized motion controller. In a controlled study of 3384 peg transfer trials performed by the system, an expert surgical resident, and 9 volunteers, results suggest that the system achieves accuracy on par with the experienced surgical resident and is significantly faster and more consistent than the surgical resident and volunteers. The system also exhibits the highest consistency and lowest collision rate. To our knowledge, this is the first autonomous system to achieve superhuman performance on a standardized surgical task.

Daniel Seita, Pete Florence, Jonathan Tompson et al.

Rearranging and manipulating deformable objects such as cables, fabrics, and bags is a long-standing challenge in robotic manipulation. The complex dynamics and high-dimensional configuration spaces of deformables, compared to rigid objects, make manipulation difficult not only for multi-step planning, but even for goal specification. Goals cannot be as easily specified as rigid object poses, and may involve complex relative spatial relations such as "place the item inside the bag". In this work, we develop a suite of simulated benchmarks with 1D, 2D, and 3D deformable structures, including tasks that involve image-based goal-conditioning and multi-step deformable manipulation. We propose embedding goal-conditioning into Transporter Networks, a recently proposed model architecture for learning robotic manipulation that rearranges deep features to infer displacements that can represent pick and place actions. In simulation and in physical experiments, we demonstrate that goal-conditioned Transporter Networks enable agents to manipulate deformable structures into flexibly specified configurations without test-time visual anchors for target locations. We also significantly extend prior results using Transporter Networks for manipulating deformable objects by testing on tasks with 2D and 3D deformables. Supplementary material is available at https://berkeleyautomation.github.io/bags/.

Samuel Paradis, Minho Hwang, Brijen Thananjeyan et al.

Automation of surgical tasks using cable-driven robots is challenging due to backlash, hysteresis, and cable tension, and these issues are exacerbated as surgical instruments must often be changed during an operation. In this work, we propose a framework for automation of high-precision surgical tasks by learning sample efficient, accurate, closed-loop policies that operate directly on visual feedback instead of robot encoder estimates. This framework, which we call intermittent visual servoing (IVS), intermittently switches to a learned visual servo policy for high-precision segments of repetitive surgical tasks while relying on a coarse open-loop policy for the segments where precision is not necessary. To compensate for cable-related effects, we apply imitation learning to rapidly train a policy that maps images of the workspace and instrument from a top-down RGB camera to small corrective motions. We train the policy using only 180 human demonstrations that are roughly 2 seconds each. Results on a da Vinci Research Kit suggest that combining the coarse policy with half a second of corrections from the learned policy during each high-precision segment improves the success rate on the Fundamentals of Laparoscopic Surgery peg transfer task from 72.9% to 99.2%, 31.3% to 99.2%, and 47.2% to 100.0% for 3 instruments with differing cable-related effects. In the contexts we studied, IVS attains the highest published success rates for automated surgical peg transfer and is significantly more reliable than previous techniques when instruments are changed. Supplementary material is available at https://tinyurl.com/ivs-icra.

Harry Zhang, Jeffrey Ichnowski, Daniel Seita et al.

We explore how high-speed robot arm motions can dynamically manipulate cables to vault over obstacles, knock objects from pedestals, and weave between obstacles. In this paper, we propose a self-supervised learning framework that enables a UR5 robot to perform these three tasks. The framework finds a 3D apex point for the robot arm, which, together with a task-specific trajectory function, defines an arcing motion that dynamically manipulates the cable to perform tasks with varying obstacle and target locations. The trajectory function computes minimum-jerk motions that are constrained to remain within joint limits and to travel through the 3D apex point by repeatedly solving quadratic programs to find the shortest and fastest feasible motion. We experiment with 5 physical cables with different thickness and mass and compare performance against two baselines in which a human chooses the apex point. Results suggest that a baseline with a fixed apex across the three tasks achieves respective success rates of 51.7%, 36.7%, and 15.0%, and a baseline with human-specified, task-specific apex points achieves 66.7%, 56.7%, and 15.0% success rate respectively, while the robot using the learned apex point can achieve success rates of 81.7% in vaulting, 65.0% in knocking, and 60.0% in weaving. Code, data, and supplementary materials are available at https: //sites.google.com/berkeley.edu/dynrope/home.

Aditya Ganapathi, Priya Sundaresan, Brijen Thananjeyan et al.

We explore learning pixelwise correspondences between images of deformable objects in different configurations. Traditional correspondence matching approaches such as SIFT, SURF, and ORB can fail to provide sufficient contextual information for fine-grained manipulation. We propose Multi-Modal Gaussian Shape Descriptor (MMGSD), a new visual representation of deformable objects which extends ideas from dense object descriptors to predict all symmetric correspondences between different object configurations. MMGSD is learned in a self-supervised manner from synthetic data and produces correspondence heatmaps with measurable uncertainty. In simulation, experiments suggest that MMGSD can achieve an RMSE of 32.4 and 31.3 for square cloth and braided synthetic nylon rope respectively. The results demonstrate an average of 47.7% improvement over a provided baseline based on contrastive learning, symmetric pixel-wise contrastive loss (SPCL), as opposed to MMGSD which enforces distributional continuity.

Aditya Ganapathi, Priya Sundaresan, Brijen Thananjeyan et al.

Robotic fabric manipulation is challenging due to the infinite dimensional configuration space, self-occlusion, and complex dynamics of fabrics. There has been significant prior work on learning policies for specific deformable manipulation tasks, but comparatively less focus on algorithms which can efficiently learn many different tasks. In this paper, we learn visual correspondences for deformable fabrics across different configurations in simulation and show that this representation can be used to design policies for a variety of tasks. Given a single demonstration of a new task from an initial fabric configuration, the learned correspondences can be used to compute geometrically equivalent actions in a new fabric configuration. This makes it possible to robustly imitate a broad set of multi-step fabric smoothing and folding tasks on multiple physical robotic systems. The resulting policies achieve 80.3% average task success rate across 10 fabric manipulation tasks on two different robotic systems, the da Vinci surgical robot and the ABB YuMi. Results also suggest robustness to fabrics of various colors, sizes, and shapes. See https://tinyurl.com/fabric-descriptors for supplementary material and videos.

Ryan Hoque, Daniel Seita, Ashwin Balakrishna et al.

Robotic fabric manipulation has applications in home robotics, textiles, senior care and surgery. Existing fabric manipulation techniques, however, are designed for specific tasks, making it difficult to generalize across different but related tasks. We extend the Visual Foresight framework to learn fabric dynamics that can be efficiently reused to accomplish different fabric manipulation tasks with a single goal-conditioned policy. We introduce VisuoSpatial Foresight (VSF), which builds on prior work by learning visual dynamics on domain randomized RGB images and depth maps simultaneously and completely in simulation. We experimentally evaluate VSF on multi-step fabric smoothing and folding tasks against 5 baseline methods in simulation and on the da Vinci Research Kit (dVRK) surgical robot without any demonstrations at train or test time. Furthermore, we find that leveraging depth significantly improves performance. RGBD data yields an 80% improvement in fabric folding success rate over pure RGB data. Code, data, videos, and supplementary material are available at https://sites.google.com/view/fabric-vsf/.

Minho Hwang, Brijen Thananjeyan, Samuel Paradis et al.

Automation of surgical subtasks using cable-driven robotic surgical assistants (RSAs) such as Intuitive Surgical's da Vinci Research Kit (dVRK) is challenging due to imprecision in control from cable-related effects such as cable stretching and hysteresis. We propose a novel approach to efficiently calibrate such robots by placing a 3D printed fiducial coordinate frames on the arm and end-effector that is tracked using RGBD sensing. To measure the coupling and history-dependent effects between joints, we analyze data from sampled trajectories and consider 13 approaches to modeling. These models include linear regression and LSTM recurrent neural networks, each with varying temporal window length to provide compensatory feedback. With the proposed method, data collection of 1800 samples takes 31 minutes and model training takes under 1 minute. Results on a test set of reference trajectories suggest that the trained model can reduce the mean tracking error of the physical robot from 2.96 mm to 0.65 mm. Results on the execution of open-loop trajectories of the FLS peg transfer surgeon training task suggest that the best model increases success rate from 39.4 % to 96.7 %, producing performance comparable to that of an expert surgical resident. Supplementary materials, including code and 3D-printable models, are available at https://sites.google.com/berkeley.edu/surgical-calibration

Minho Hwang, Daniel Seita, Brijen Thananjeyan et al.

Recent advances in depth-sensing have significantly increased accuracy, resolution, and frame rate, as shown in the 1920x1200 resolution and 13 frames per second Zivid RGBD camera. In this study, we explore the potential of depth sensing for efficient and reliable automation of surgical subtasks. We consider a monochrome (all red) version of the peg transfer task from the Fundamentals of Laparoscopic Surgery training suite implemented with the da Vinci Research Kit (dVRK). We use calibration techniques that allow the imprecise, cable-driven da Vinci to reduce error from 4-5 mm to 1-2 mm in the task space. We report experimental results for a handover-free version of the peg transfer task, performing 20 and 5 physical episodes with single- and bilateral-arm setups, respectively. Results over 236 and 49 total block transfer attempts for the single- and bilateral-arm peg transfer cases suggest that reliability can be attained with 86.9 % and 78.0 % for each individual block, with respective block transfer speeds of 10.02 and 5.72 seconds. Supplementary material is available at https://sites.google.com/view/peg-transfer.

Daniel Seita, David Chan, Roshan Rao et al.

Learning from demonstrations is a popular tool for accelerating and reducing the exploration requirements of reinforcement learning. When providing expert demonstrations to human students, we know that the demonstrations must fall within a particular range of difficulties called the "Zone of Proximal Development (ZPD)". If they are too easy the student learns nothing, but if they are too difficult the student is unable to follow along. This raises the question: Given a set of potential demonstrators, which among them is best suited for teaching any particular learner? Prior work, such as the popular Deep Q-learning from Demonstrations (DQfD) algorithm has generally focused on single demonstrators. In this work we consider the problem of choosing among multiple demonstrators of varying skill levels. Our results align with intuition from human learners: it is not always the best policy to draw demonstrations from the best performing demonstrator (in terms of reward). We show that careful selection of teaching strategies can result in sample efficiency gains in the learner's environment across nine Atari games

Daniel Seita, Aditya Ganapathi, Ryan Hoque et al.

Sequential pulling policies to flatten and smooth fabrics have applications from surgery to manufacturing to home tasks such as bed making and folding clothes. Due to the complexity of fabric states and dynamics, we apply deep imitation learning to learn policies that, given color (RGB), depth (D), or combined color-depth (RGBD) images of a rectangular fabric sample, estimate pick points and pull vectors to spread the fabric to maximize coverage. To generate data, we develop a fabric simulator and an algorithmic supervisor that has access to complete state information. We train policies in simulation using domain randomization and dataset aggregation (DAgger) on three tiers of difficulty in the initial randomized configuration. We present results comparing five baseline policies to learned policies and report systematic comparisons of RGB vs D vs RGBD images as inputs. In simulation, learned policies achieve comparable or superior performance to analytic baselines. In 180 physical experiments with the da Vinci Research Kit (dVRK) surgical robot, RGBD policies trained in simulation attain coverage of 83% to 95% depending on difficulty tier, suggesting that effective fabric smoothing policies can be learned from an algorithmic supervisor and that depth sensing is a valuable addition to color alone. Supplementary material is available at https://sites.google.com/view/fabric-smoothing.

Xinlei Pan, Daniel Seita, Yang Gao et al.

Deep reinforcement learning has recently made significant progress in solving computer games and robotic control tasks. A known problem, though, is that policies overfit to the training environment and may not avoid rare, catastrophic events such as automotive accidents. A classical technique for improving the robustness of reinforcement learning algorithms is to train on a set of randomized environments, but this approach only guards against common situations. Recently, robust adversarial reinforcement learning (RARL) was developed, which allows efficient applications of random and systematic perturbations by a trained adversary. A limitation of RARL is that only the expected control objective is optimized; there is no explicit modeling or optimization of risk. Thus the agents do not consider the probability of catastrophic events (i.e., those inducing abnormally large negative reward), except through their effect on the expected objective. In this paper we introduce risk-averse robust adversarial reinforcement learning (RARARL), using a risk-averse protagonist and a risk-seeking adversary. We test our approach on a self-driving vehicle controller. We use an ensemble of policy networks to model risk as the variance of value functions. We show through experiments that a risk-averse agent is better equipped to handle a risk-seeking adversary, and experiences substantially fewer crashes compared to agents trained without an adversary.

Daniel Seita, Nawid Jamali, Michael Laskey et al.

A fundamental challenge in manipulating fabric for clothes folding and textiles manufacturing is computing "pick points" to effectively modify the state of an uncertain manifold. We present a supervised deep transfer learning approach to locate pick points using depth images for invariance to color and texture. We consider the task of bed-making, where a robot sequentially grasps and pulls at pick points to increase blanket coverage. We perform physical experiments with two mobile manipulator robots, the Toyota HSR and the Fetch, and three blankets of different colors and textures. We compare coverage results from (1) human supervision, (2) a baseline of picking at the uppermost blanket point, and (3) learned pick points. On a quarter-scale twin bed, a model trained with combined data from the two robots achieves 92% blanket coverage compared with 83% for the baseline and 95% for human supervisors. The model transfers to two novel blankets and achieves 93% coverage. Average coverage results of 92% for 193 beds suggest that transfer-invariant robot pick points on fabric can be effectively learned.

Daniel Seita, Sanjay Krishnan, Roy Fox et al.

Automating precision subtasks such as debridement (removing dead or diseased tissue fragments) with Robotic Surgical Assistants (RSAs) such as the da Vinci Research Kit (dVRK) is challenging due to inherent non-linearities in cable-driven systems. We propose and evaluate a novel two-phase coarse-to-fine calibration method. In Phase I (coarse), we place a red calibration marker on the end effector and let it randomly move through a set of open-loop trajectories to obtain a large sample set of camera pixels and internal robot end-effector configurations. This coarse data is then used to train a Deep Neural Network (DNN) to learn the coarse transformation bias. In Phase II (fine), the bias from Phase I is applied to move the end-effector toward a small set of specific target points on a printed sheet. For each target, a human operator manually adjusts the end-effector position by direct contact (not through teleoperation) and the residual compensation bias is recorded. This fine data is then used to train a Random Forest (RF) to learn the fine transformation bias. Subsequent experiments suggest that without calibration, position errors average 4.55mm. Phase I can reduce average error to 2.14mm and the combination of Phase I and Phase II can reduces average error to 1.08mm. We apply these results to debridement of raisins and pumpkin seeds as fragment phantoms. Using an endoscopic stereo camera with standard edge detection, experiments with 120 trials achieved average success rates of 94.5%, exceeding prior results with much larger fragments (89.4%) and achieving a speedup of 2.1x, decreasing time per fragment from 15.8 seconds to 7.3 seconds. Source code, data, and videos are available at https://sites.google.com/view/calib-icra/.

Daniel Seita, Xinlei Pan, Haoyu Chen et al.

We present a novel Metropolis-Hastings method for large datasets that uses small expected-size minibatches of data. Previous work on reducing the cost of Metropolis-Hastings tests yield variable data consumed per sample, with only constant factor reductions versus using the full dataset for each sample. Here we present a method that can be tuned to provide arbitrarily small batch sizes, by adjusting either proposal step size or temperature. Our test uses the noise-tolerant Barker acceptance test with a novel additive correction variable. The resulting test has similar cost to a normal SGD update. Our experiments demonstrate several order-of-magnitude speedups over previous work.

Daniel Seita, Haoyu Chen, John Canny

A fundamental task in machine learning and related fields is to perform inference on Bayesian networks. Since exact inference takes exponential time in general, a variety of approximate methods are used. Gibbs sampling is one of the most accurate approaches and provides unbiased samples from the posterior but it has historically been too expensive for large models. In this paper, we present an optimized, parallel Gibbs sampler augmented with state replication (SAME or State Augmented Marginal Estimation) to decrease convergence time. We find that SAME can improve the quality of parameter estimates while accelerating convergence. Experiments on both synthetic and real data show that our Gibbs sampler is substantially faster than the state of the art sampler, JAGS, without sacrificing accuracy. Our ultimate objective is to introduce the Gibbs sampler to researchers in many fields to expand their range of feasible inference problems.