arXiv subjects

Qian Luo

Publications and source records attributed to Qian Luo.

At least 19 recordsLinked to original sources

arXiv2026

The Imitator Game: Benchmarking Robot Imitative Ability Beyond Action Prediction

Humans imitate at the level of intent: given a demonstration, we infer its goal and carry it out with whatever tools, objects, and layouts are at hand. Current robot policies instead learn observation-to-action mappings from visual inputs and language instructions, without explicitly inferring the demonstrated task. Learning from human video thus remains largely trajectory-level: models can replay motions in near-identical scenes, but still struggle to imitate what the demonstrator intends rather than merely what they do. We introduce The Imitator Game, a four-level benchmark (L0-L3) that progressively widens the gap between the human demonstration and the robot's own scene, isolating where trajectory replay ceases to suffice and task understanding becomes necessary. We pair it with IG-10K, the largest environment-aligned paired human-robot dataset to date and the only one instantiated across all four levels in both real and simulated settings (20,000+ paired episodes, 50+ tasks, 6 domains), and Imitator Arena, an open platform for blind A/B human evaluation. Across nine state-of-the-art models, performance is stable from L0 to L2 but collapses at L3, identifying functional substitution - achieving the same intent through a different object affordance - as the decisive barrier to intent-level imitation. Human-video-conditioned models outperform caption-conditioned ones, yet every model falls below 13% zero-shot success on unseen tasks; fine-tuning IG-10K-pretrained models with only $10$ paired human-robot demonstrations yields large gains that grow with pretraining scale. The project website and access to Imitator Arena are available at https://imitator-game.github.io.

cs.RO↗

arXiv2026

RoboLineage: Agent-Native Data Lifecycle Governance Across Robot Policy Iterations

We present RoboLineage, an agent-native data lifecycle governance system for robot policy iteration. Modern robot policies improve through repeated data collection, review, retraining, evaluation, and release decisions, but the evidence connecting these steps is often scattered across local tools, scripts, and expert memory. RoboLineage makes this lifecycle explicit by representing rollouts, reviews, dataset decisions, training runs, policy metadata, evaluations, deployment recommendations, and next-collection plans as typed lineage artifacts. Agents interpret embodied rollout evidence, adapt accepted data to existing training stacks, maintain data health, and summarize cross-iteration state under explicit artifact boundaries. In real-robot manipulation workflows, RoboLineage makes routine policy iteration faster and more auditable while maintaining downstream policy performance. We open source RoboLineage as a lightweight lifecycle layer for different robot embodiments and training families. Project page: https://robolineage.github.io/

cs.RO↗

arXiv2026

Geometric Entropy: When Trajectory Diversity Helps and Hurts in Imitation Learning

We study how trajectory-shape diversity in demonstrations affects imitation learning (IL) performance across models, tasks, and data scales. We introduce Geometric Entropy (H_G), a task-agnostic metric that quantifies the intrinsic diversity of transit trajectories after normalizing away extrinsic variation, such as goal pose and workspace scale, via target-frame alignment. Across multiple IL architectures and both simulated and real-robot contact-rich manipulation tasks, we observe a consistent inverted-U relationship between success and H_G: increasing geometric diversity improves robustness in low-diversity regimes but degrades performance once diversity induces strategy ambiguity. Moreover, the optimal entropy shifts toward lower values as task mastery increases through more data, easier tasks, or stronger priors, and for a pretrained vision-language-action model the trend becomes effectively monotonic decreasing. Practically, H_G enables fast pre-training auditing of demonstration datasets and offers a simple guideline for calibrating demonstrations toward the learnable regime.

cs.RO↗

arXiv2025

MICCAI STSR 2025 Challenge: Semi-Supervised Teeth and Pulp Segmentation and CBCT-IOS Registration

Cone-Beam Computed Tomography (CBCT) and Intraoral Scanning (IOS) are essential for digital dentistry, but annotated data scarcity limits automated solutions for pulp canal segmentation and cross-modal registration. To benchmark semi-supervised learning (SSL) in this domain, we organized the STSR 2025 Challenge at MICCAI 2025, featuring two tasks: (1) semi-supervised segmentation of teeth and pulp canals in CBCT, and (2) semi-supervised rigid registration of CBCT and IOS. We provided 60 labeled and 640 unlabeled IOS samples, plus 30 labeled and 250 unlabeled CBCT scans with varying resolutions and fields of view. The challenge attracted strong community participation, with top teams submitting open-source deep learning-based SSL solutions. For segmentation, leading methods used nnU-Net and Mamba-like State Space Models with pseudo-labeling and consistency regularization, achieving a Dice score of 0.967 and Instance Affinity of 0.738 on the hidden test set. For registration, effective approaches combined PointNetLK with differentiable SVD and geometric augmentation to handle modality gaps; hybrid neural-classical refinement enabled accurate alignment despite limited labels. All data and code are publicly available at https://github.com/ricoleehduu/STS-Challenge-2025 to ensure reproducibility.

cs.CV↗

arXiv2025

MICCAI STS 2024 Challenge: Semi-Supervised Instance-Level Tooth Segmentation in Panoramic X-ray and CBCT Images

Orthopantomogram (OPGs) and Cone-Beam Computed Tomography (CBCT) are vital for dentistry, but creating large datasets for automated tooth segmentation is hindered by the labor-intensive process of manual instance-level annotation. This research aimed to benchmark and advance semi-supervised learning (SSL) as a solution for this data scarcity problem. We organized the 2nd Semi-supervised Teeth Segmentation (STS 2024) Challenge at MICCAI 2024. We provided a large-scale dataset comprising over 90,000 2D images and 3D axial slices, which includes 2,380 OPG images and 330 CBCT scans, all featuring detailed instance-level FDI annotations on part of the data. The challenge attracted 114 (OPG) and 106 (CBCT) registered teams. To ensure algorithmic excellence and full transparency, we rigorously evaluated the valid, open-source submissions from the top 10 (OPG) and top 5 (CBCT) teams, respectively. All successful submissions were deep learning-based SSL methods. The winning semi-supervised models demonstrated impressive performance gains over a fully-supervised nnU-Net baseline trained only on the labeled data. For the 2D OPG track, the top method improved the Instance Affinity (IA) score by over 44 percentage points. For the 3D CBCT track, the winning approach boosted the Instance Dice score by 61 percentage points. This challenge confirms the substantial benefit of SSL for complex, instance-level medical image segmentation tasks where labeled data is scarce. The most effective approaches consistently leveraged hybrid semi-supervised frameworks that combined knowledge from foundational models like SAM with multi-stage, coarse-to-fine refinement pipelines. Both the challenge dataset and the participants' submitted code have been made publicly available on GitHub (https://github.com/ricoleehduu/STS-Challenge-2024), ensuring transparency and reproducibility.

eess.IV↗

arXiv2025

Hyper-GoalNet: Goal-Conditioned Manipulation Policy Learning with HyperNetworks

Goal-conditioned policy learning for robotic manipulation presents significant challenges in maintaining performance across diverse objectives and environments. We introduce Hyper-GoalNet, a framework that generates task-specific policy network parameters from goal specifications using hypernetworks. Unlike conventional methods that simply condition fixed networks on goal-state pairs, our approach separates goal interpretation from state processing -- the former determines network parameters while the latter applies these parameters to current observations. To enhance representation quality for effective policy generation, we implement two complementary constraints on the latent space: (1) a forward dynamics model that promotes state transition predictability, and (2) a distance-based constraint ensuring monotonic progression toward goal states. We evaluate our method on a comprehensive suite of manipulation tasks with varying environmental randomization. Results demonstrate significant performance improvements over state-of-the-art methods, particularly in high-variability conditions. Real-world robotic experiments further validate our method's robustness to sensor noise and physical uncertainties. Code is available at: https://github.com/wantingyao/hyper-goalnet.

cs.RO↗

arXiv2025

HiMaCon: Discovering Hierarchical Manipulation Concepts from Unlabeled Multi-Modal Data

Effective generalization in robotic manipulation requires representations that capture invariant patterns of interaction across environments and tasks. We present a self-supervised framework for learning hierarchical manipulation concepts that encode these invariant patterns through cross-modal sensory correlations and multi-level temporal abstractions without requiring human annotation. Our approach combines a cross-modal correlation network that identifies persistent patterns across sensory modalities with a multi-horizon predictor that organizes representations hierarchically across temporal scales. Manipulation concepts learned through this dual structure enable policies to focus on transferable relational patterns while maintaining awareness of both immediate actions and longer-term goals. Empirical evaluation across simulated benchmarks and real-world deployments demonstrates significant performance improvements with our concept-enhanced policies. Analysis reveals that the learned concepts resemble human-interpretable manipulation primitives despite receiving no semantic supervision. This work advances both the understanding of representation learning for manipulation and provides a practical approach to enhancing robotic performance in complex scenarios.

cs.RO↗

arXiv2025

Solving time-fractional diffusion equations with Robin boundary conditions via fractional Hamiltonian boundary value methods

In this paper, we propose a novel numerical scheme for solving time-fractional reaction-diffusion problems with Robin boundary conditions, where the time derivative is in the Caputo sense of order $α\in(0,1)$. The existence and uniqueness of the solution is proved. Our proposed method is based on the spectral collocation method in space and Fractional Hamiltonian boundary value methods in time. For the considered spectral collocation method, the basis functions used are not the standard polynomial basis functions, but rather adapt to Robin boundary conditions, and the exponential convergence property is provided. The proposed procedure achieves spectral accuracy in space and is also capable of getting spectral accuracy in time. Some numerical examples are provided to support the theoretical results.

math.NA↗

arXiv2025

How flexible do we need to be? Using electricity systems models to identify optimal designs for flexible carbon capture storage system for gas-fired power plants

As the share of variable renewable energy in power systems grows, enhancing the operational flexibility of combined cycle gas turbines with carbon capture and storage (CCGT-CCS) becomes increasingly valuable. This study integrates techno-economic analysis with capacity expansion modeling to quantify the value of improved CCGT-CCS flexibility-such as lower start-up costs, reduced minimum generation, faster ramping, and shorter up/down times-at both plant and system levels. Using the Texas power system as a case study, we find that increased flexibility raises CCGT-CCS generation profits and installed capacity. Under various policy scenarios, CCGT-CCS benefits most from a CO2 tax (or equivalent emissions cap), more so than from clean energy standards or capture subsidies like the federal 45Q tax credit. However, electricity system cost savings remain modest, reducing total costs by only 0.3-0.5%. Thus, flexibility improvements should be pursued only if they entail limited increases in capital and maintenance costs.

eess.SY↗

arXiv2025

Are EVs Cleaner Than We Think? Evaluating Consequential Greenhouse Gas Emissions from EV Charging

While electrifying transportation eliminates tailpipe greenhouse gas (GHG) emissions, electric vehicle (EV) adoption can create additional electricity sector emissions. To quantify this emissions impact, prior work typically employs short-run marginal emissions or average emissions rates calculated from historical data or power systems models that do not consider changes in installed capacity. In this work, we use an electricity system capacity expansion model to consider the full consequential GHG emissions impact from large-scale EV adoption in the western United States, accounting for induced changes in generation and storage capacity. We find that the metrics described above do not accurately reflect the true emissions impact of EV adoption-average emissions rates can either under- or over-estimate emission impacts, and short-run marginal emissions rates can significantly underestimate emission reductions, especially when charging timing is flexible. Our results also show that using short-run marginal emission rates as signals to coordinate EV charging could increase emissions relative to price-based charging signals, indicating the need for alternative control strategies to minimize consequential emissions.

eess.SY↗

arXiv2025

Process and Policy Insights from an Intercomparison of Open Electricity System Capacity Expansion Models

This study performs a detailed intercomparison of four open-source electricity capacity expansion models - Temoa, Switch, GenX, and USENSYS - to evaluate 1) how closely the results of these models align when inputs and configurations are harmonized, and 2) the degree to which varying model configurations affect outputs. We harmonize the inputs to each model using PowerGenome and use clearly defined scenarios (policy conditions) and configurations (model setup choices). This allows us to isolate how differences in model structure affect policy outcomes and investment decisions. Our framework allows each model to be tested on identical assumptions for policy, technology costs, and operational constraints, allowing us to focus on differences that arise from inherent model structures. Key findings highlight that, when harmonized, models produce very similar capacity portfolios under current policies and net-zero scenarios, with less than 1% difference in system costs for most configurations. This agreement among models allows us to focus on how configuration choices affect model results. For instance, configurations with unit commitment constraints or economic retirement yield different investments and system costs compared to simpler configurations. Our findings underscore the importance of aligning input data and transparently defining scenarios and configurations to provide robust policy insights.

econ.GN↗

arXiv2025

Impacts of EPA's Finalized Power Plant Greenhouse Gas Standards

The Inflation Reduction Act subsidizes the deployment of clean electricity, hydrogen production, and carbon capture and storage, which could enable additional actions by other federal, state, and local policymakers to reduce emissions. Power plant rules finalized by the Environmental Protection Agency (EPA) in 2024 are one such example of complementary policies. The rules establish emissions intensity standards, not technology mandates, meaning power plant owners can choose from a range of technologies and control options provided that emissions standards are met. This flexibility makes electricity systems modeling important to understand the potential effects of these regulations. We report below a multi-model analysis of the EPA power plant rules that can provide timely information, including for other countries and states, on emissions impacts, policy design for electricity decarbonization, power sector investments and retirements, cost impacts, and load growth. We also discuss related technical, political, and legal uncertainties.

physics.soc-ph↗

arXiv2024

Impacts of EPA Power Plant Emissions Regulations on the US Electricity Sector

Taking aim at one of the largest greenhouse gas emitting sectors, the US Environmental Protection Agency (EPA) finalized new regulations on power plant greenhouse gas emissions in May 2024. These rules take the form of different emissions performance standards for different classes of power plant technologies, creating a complex set of regulations that make it difficult to understand their consequential impacts on power system capacity, operations, and emissions without dedicated and sophisticated modeling. Here, we enhance a state-of-the-art power system capacity expansion model by incorporating new detailed operational constraints tailored to different technologies to represent the EPA's rules. Our results show that adopting these new regulations could reduce US power sector emissions in 2040 to 51% below the 2022 level (vs 26% without the rules). Regulations on coal-fired power plants drive the largest share of reductions. Regulations on new gas turbines incrementally reduce emissions but lower overall efficiency of the gas fleet, increasing the average cost of carbon mitigation. Therefore, we explore several alternative emission mitigation strategies. By comparing these alternatives with regulations finalized by EPA, we highlight the importance of accelerating the retirement of inefficient fossil fuel-fired generators and applying consistent and strict emissions regulations to all gas generators, regardless of their vintage, to cost-effectively achieve deep decarbonization and avoid biasing investment decisions towards less efficient generators.

eess.SY↗

arXiv2024

Text2Reward: Reward Shaping with Language Models for Reinforcement Learning

Designing reward functions is a longstanding challenge in reinforcement learning (RL); it requires specialized knowledge or domain data, leading to high costs for development. To address this, we introduce Text2Reward, a data-free framework that automates the generation and shaping of dense reward functions based on large language models (LLMs). Given a goal described in natural language, Text2Reward generates shaped dense reward functions as an executable program grounded in a compact representation of the environment. Unlike inverse RL and recent work that uses LLMs to write sparse reward codes or unshaped dense rewards with a constant function across timesteps, Text2Reward produces interpretable, free-form dense reward codes that cover a wide range of tasks, utilize existing packages, and allow iterative refinement with human feedback. We evaluate Text2Reward on two robotic manipulation benchmarks (ManiSkill2, MetaWorld) and two locomotion environments of MuJoCo. On 13 of the 17 manipulation tasks, policies trained with generated reward codes achieve similar or better task success rates and convergence speed than expert-written reward codes. For locomotion tasks, our method learns six novel locomotion behaviors with a success rate exceeding 94%. Furthermore, we show that the policies trained in the simulator with our method can be deployed in the real world. Finally, Text2Reward further improves the policies by refining their reward functions with human feedback. Video results are available at https://text-to-reward.github.io/ .

cs.LG↗

arXiv2024

InfoCon: Concept Discovery with Generative and Discriminative Informativeness

We focus on the self-supervised discovery of manipulation concepts that can be adapted and reassembled to address various robotic tasks. We propose that the decision to conceptualize a physical procedure should not depend on how we name it (semantics) but rather on the significance of the informativeness in its representation regarding the low-level physical state and state changes. We model manipulation concepts (discrete symbols) as generative and discriminative goals and derive metrics that can autonomously link them to meaningful sub-trajectories from noisy, unlabeled demonstrations. Specifically, we employ a trainable codebook containing encodings (concepts) capable of synthesizing the end-state of a sub-trajectory given the current state (generative informativeness). Moreover, the encoding corresponding to a particular sub-trajectory should differentiate the state within and outside it and confidently predict the subsequent action based on the gradient of its discriminative score (discriminative informativeness). These metrics, which do not rely on human annotation, can be seamlessly integrated into a VQ-VAE framework, enabling the partitioning of demonstrations into semantically consistent sub-trajectories, fulfilling the purpose of discovering manipulation concepts and the corresponding sub-goal (key) states. We evaluate the effectiveness of the learned concepts by training policies that utilize them as guidance, demonstrating superior performance compared to other baselines. Additionally, our discovered manipulation concepts compare favorably to human-annotated ones while saving much manual effort.

cs.RO↗

arXiv2024

BBSEA: An Exploration of Brain-Body Synchronization for Embodied Agents

Embodied agents capable of complex physical skills can improve productivity, elevate life quality, and reshape human-machine collaboration. We aim at autonomous training of embodied agents for various tasks involving mainly large foundation models. It is believed that these models could act as a brain for embodied agents; however, existing methods heavily rely on humans for task proposal and scene customization, limiting the learning autonomy, training efficiency, and generalization of the learned policies. In contrast, we introduce a brain-body synchronization ({\it BBSEA}) scheme to promote embodied learning in unknown environments without human involvement. The proposed combines the wisdom of foundation models (``brain'') with the physical capabilities of embodied agents (``body''). Specifically, it leverages the ``brain'' to propose learnable physical tasks and success metrics, enabling the ``body'' to automatically acquire various skills by continuously interacting with the scene. We carry out an exploration of the proposed autonomous learning scheme in a table-top setting, and we demonstrate that the proposed synchronization can generate diverse tasks and develop multi-task policies with promising adaptability to new tasks and configurations. We will release our data, code, and trained models to facilitate future studies in building autonomously learning agents with large foundation models in more complex scenarios. More visualizations are available at \href{https://bbsea-embodied-ai.github.io}{https://bbsea-embodied-ai.github.io}

cs.RO↗

arXiv2023

Benchmarking Augmentation Methods for Learning Robust Navigation Agents: the Winning Entry of the 2021 iGibson Challenge

Recent advances in deep reinforcement learning and scalable photorealistic simulation have led to increasingly mature embodied AI for various visual tasks, including navigation. However, while impressive progress has been made for teaching embodied agents to navigate static environments, much less progress has been made on more dynamic environments that may include moving pedestrians or movable obstacles. In this study, we aim to benchmark different augmentation techniques for improving the agent's performance in these challenging environments. We show that adding several dynamic obstacles into the scene during training confers significant improvements in test-time generalization, achieving much higher success rates than baseline agents. We find that this approach can also be combined with image augmentation methods to achieve even higher success rates. Additionally, we show that this approach is also more robust to sim-to-sim transfer than image augmentation methods. Finally, we demonstrate the effectiveness of this dynamic obstacle augmentation approach by using it to train an agent for the 2021 iGibson Challenge at CVPR, where it achieved 1st place for Interactive Navigation. Video link: https://www.youtube.com/watch?v=HxUX2HeOSE4

cs.RO↗

arXiv2023

Grounding Object Relations in Language-Conditioned Robotic Manipulation with Semantic-Spatial Reasoning

Grounded understanding of natural language in physical scenes can greatly benefit robots that follow human instructions. In object manipulation scenarios, existing end-to-end models are proficient at understanding semantic concepts, but typically cannot handle complex instructions involving spatial relations among multiple objects. which require both reasoning object-level spatial relations and learning precise pixel-level manipulation affordances. We take an initial step to this challenge with a decoupled two-stage solution. In the first stage, we propose an object-centric semantic-spatial reasoner to select which objects are relevant for the language instructed task. The segmentation of selected objects are then fused as additional input to the affordance learning stage. Simply incorporating the inductive bias of relevant objects to a vision-language affordance learning agent can effectively boost its performance in a custom testbed designed for object manipulation with spatial-related language instructions.

cs.RO↗