arXiv subjects

Benjamin Yang

Publications and source records attributed to Benjamin Yang.

8 recordsLinked to original sources

arXiv2026

NovaPlan: Zero-Shot Long-Horizon Manipulation via Closed-Loop Video Language Planning

Solving complex long-horizon robotic tasks requires joint reasoning over abstract task structure and low-level physical interaction. While combining Vision-Language Models (VLMs) and video generation models offers a promising path for zero-shot planning, their individual tendencies to hallucinate physics or violate geometric consistency often compound over time, preventing reliable real-world execution. We introduce NovaPlan, a hierarchical framework that enables robust, zero-shot long-horizon manipulation by systematically proposing, verifying, and repairing visual plans. At the high level, a VLM planner decomposes tasks and filters out dynamically inconsistent futures by verifying multiple candidate video rollouts. To translate these imagined futures into reliable physical actions, NovaPlan utilizes a hybrid geometric representation that adaptively switches between object-centric flow and human hand flow. Finally, NovaPlan closes the loop by continuously monitoring execution to verify outcomes and synthesize local, non-prehensile corrective behaviors, such as fingertip poking, when failures occur. Across diverse multi-stage tasks, NovaPlan substantially outperforms prior zero-shot systems, achieving complex assembly and dexterous error recovery entirely without task-specific training or demonstrations. Please visit our project website for additional results: https://nova-plan.github.io/

cs.RO↗

arXiv2026

From Pretraining to Proficiency: Real-World Subtask RL for Long-Horizon Manipulation with Minimal Human Intervention

A pretrained robot foundation policy may execute most of a long-horizon task yet repeatedly fail at a few critical subtasks. Collecting additional full-task demonstrations for supervised fine-tuning (SFT) requires operators to repeat behaviors the policy already performs well. Reinforcement learning (RL) fine-tuning offers a promising path to bridge this gap, but existing approaches struggle to solve long-horizon tasks using only sparse rewards. We present PARTS (Policy Adaptation with RL on Targeted Subtasks), a real-world subtask RL framework that concentrates practice at these bottlenecks while allowing training rollouts to proceed with minimal human intervention. The frozen pretrained policy supplies nominal actions throughout execution, while agent-generated selectors and success verifiers activate residual corrections and provide local outcome rewards. These rewards support learning from successful subtasks even when complete-task successes are scarce. Training combines online RL with success-reweighted retraining, and each retrained residual policy is redeployed to collect further experience. Humans identify bottlenecks during setup and perform physical resets when needed. On bimanual YAM and single-arm Franka tasks, PARTS improves complete-task success from 32% to 61% and from 50% to 95%, respectively, using tens of minutes of real-world RL rollouts per task on average. Compared with existing real-world RL fine-tuning methods, PARTS raises full-task success by more than 25% under the same robot-rollout budget while requiring less human involvement.

cs.RO↗

arXiv2026

Kinematic Interface for the Wild: Modular Bimanual Loco-Manipulation Capture from 360$^{\circ}$ Cameras Alone

A wrist-mounted camera for UMI-style data collection must do two jobs: record the manipulation and localize in the scene. Most handheld devices localize online from workspace-facing views crowded by hands and objects, or add dedicated tracking hardware. Room-scale bimanual capture therefore still tends to instrument the operator or the scene for accurate localization. We present KIWI (Kinematic Interface for the Wild), a capture kit whose only electronics are off-the-shelf cameras. Our core system splits the two jobs across the two lenses of a 360-degree camera. The rear lens faces the room and builds a shared metric map that registers both hands, and an optional head camera, in one frame without workspace co-visibility; the front lens records the manipulation, and offline IMU fusion bridges front-lens tracking loss. Through our quick-release plate, the camera module attaches to chopstick grippers, parallel-jaw grippers, hand-wrist mounts, or robot flanges. Across six bimanual recordings, combining the rear and front lenses failed to localize only 0.1% of query frames, whereas front-only bimanual feature alignment failed on 24.8% of frames and lost one recording entirely; against evaluation fiducials, localization error stayed within 4.5 mm. KIWI's recovered poses were sufficiently consistent for the four wrist streams alone to reconstruct the scene as a 3D Gaussian splat.

cs.RO↗

arXiv2026

Augmenting Human Performance with an XR Agent Learning from Online Behavior and BCI Evidence

We present OLIVE, a framework for adapting a foundation model to provide real-time assistance in temporally demanding, high-stakes, and dynamic tasks. We show that passive EEG, fused online with behavioral evidence, can meaningfully extend the number of targets users detect and engage beyond their unaided action bandwidth. OLIVE learns from both explicit behavioral signals (the targets the user shoots down in an XR first-person shooter game) and implicit physiological signals (fixation-locked EEG) to provide timely guidance, continuously adapting a frozen vision-language model's inference on which items are task-relevant by jointly estimating per-source reliability without manual labels or offline training. Through three user studies, including two live deployments of an assistive agent driven by OLIVE in XR, we show that OLIVE Pareto-dominates prior test-time adaptation frameworks, achieving the highest convergence rate at comparable convergence speed. Combining implicit physiological and explicit behavioral signals, the OLIVE agent produces the largest and most reliable within-session improvement to a user's ability to detect and engage targets, largely independent of the individual's skill. When the target switches silently, the agent that uses both behavioral and physiological signals reconverges significantly faster than the behavior-only agent (1.27 times faster on average, p = .008), restoring trustworthy guidance at the moment the task changes, precisely when reliable assistance matters most.

cs.AI↗

arXiv2026

Agents of ViTAL: Ethics Missions -- A Narrative-Centered Learning Environment with a Co-Designed Conversational Agent for Middle School AI Ethics

Agents of ViTAL: Ethics Missions is a browser-based, narrative-centered learning environment in which middle school students collaboratively evaluate whether a fictional school should adopt an AI-powered classroom feedback tool. Students investigate stakeholder perspectives, weigh tradeoffs across three AI ethics dimensions (privacy, bias, and environmental impact), and negotiate group consensus through a shared Ranking Challenge interface. The environment embeds EthicsBot, a conversational agent designed as a peer-like thought partner that scaffolds ethical reasoning and collaborative discussion. The project is being iteratively co-designed with high school students, whose feedback shapes EthicsBot's role, behavior, and guardrails. Initial classroom implementations with four ninth-grade classes demonstrated strong engagement and substantive ethical reasoning grounded in students' lived experiences with AI. The demo invites attendees to explore the learning environment, participate in the collaborative Ranking Challenge, and interact with EthicsBot.

cs.CY↗

arXiv2026

SwEYEpinch: Exploring Intuitive, Efficient Text Entry for Extended Reality via Eye and Hand Tracking

Despite steady progress, text entry in Extended Reality (XR) often remains slower and more effortful than typing on a physical keyboard or touchscreen. We explore a simple idea: use gaze to swipe through a virtual keyboard for the fast, low-effort where and a manual pinch held throughout the swipe for the when, extending and validating it through a series of user studies. We first show that a basic version including a low-latency decoder with spatiotemporal Dynamic Time Warping and fixation filtering outperforms selecting individual keys sequentially, either by finger tapping each or gazing at each while pinching. We then add mid-swipe prediction and in-gesture cancellation, improving words per minute (WPM) without hurting accuracy. We show that this approach is faster and more preferred than previous gaze-swipe approaches, finger tapping with prediction, or hand swiping with the same additions. Furthermore, a seven-day, 30-session study demonstrates sustained learning, with peak performance reaching 64.7 WPM.

cs.HC↗

arXiv2026

Bimanual XR Specification of Relative and Absolute Assembly Hierarchies for Teleoperation

We present a bimanual XR interaction approach for specifying remote assembly tasks as hierarchies of relative and absolute object constraints that specify high-level teleoperation goals for robots. Grabbing one object in each hand creates a constraint group (visualized as a hull) and groups can be nested into hierarchies. Each group can be relative (with a robot-specifiable 6DoF pose) or absolute (with an author-specified fixed 6DoF pose) in relation to its parent. A relative group specifies a subassembly that can be constructed at a location chosen by the robot software for efficiency rather than mandated by the user.

cs.HC↗

arXiv2019

Prediction of Silicate Glasses' Stiffness by High-Throughput Molecular Dynamics Simulations and Machine Learning

The development by machine learning of models predicting materials' properties usually requires the use of a large number of consistent data for training. However, quality experimental datasets are not always available or self-consistent. Here, as an alternative route, we combine machine learning with high-throughput molecular dynamics simulations to predict the Young's modulus of silicate glasses. We demonstrate that this combined approach offers excellent predictions over the entire compositional domain. By comparing the performance of select machine learning algorithms, we discuss the nature of the balance between accuracy, simplicity, and interpretability in machine learning.

cond-mat.mtrl-sci↗