arXiv subjects

Jiayao Ma

Publications and source records attributed to Jiayao Ma.

12 recordsLinked to original sources

arXiv2026

BioPro: Towards Difference-Aware Gender Fairness for Vision-Language Models

Vision-Language Models (VLMs) inherit significant social biases from their training data, notably in gender representation. Current fairness interventions often adopt a difference-unaware perspective that enforces uniform treatment across demographic groups. These approaches, however, fail to distinguish between contexts where neutrality is required and those where group-specific attributes are legitimate and must be preserved. Building upon recent advances in difference-aware fairness for text-only models, we extend this concept to the multimodal domain and formalize the problem of difference-aware gender fairness for image captioning and text-to-image generation. We advocate for selective debiasing, which aims to mitigate unwanted bias in neutral contexts while preserving valid distinctions in explicit ones. To achieve this, we propose BioPro (Bias Orthogonal Projection), an entirely training-free framework. BioPro identifies a low-dimensional gender-variation subspace through counterfactual embeddings and applies projection to selectively neutralize gender-related information. Experiments show that BioPro effectively reduces gender bias in neutral cases while maintaining gender faithfulness in explicit ones, thus providing a promising direction toward achieving selective fairness in VLMs. Beyond gender bias, we further demonstrate that BioPro can effectively generalize to continuous bias variables, such as scene brightness, highlighting its broader applicability.

cs.AI↗

arXiv2026

Affine Rota-Baxter groups and affine skew braces

Rota-Baxter groups and skew braces are closely related algebraic structures, both providing set-theoretical solutions to the Yang-Baxter equation. In this paper, we extend these structures to the setting of affine schemes. First, we introduce affine Rota-Baxter groups and, by leveraging the duality between affine groups and Hopf algebras via their coordinate rings, prove the equivalence between affine Rota-Baxter groups and co-Rota-Baxter Hopf algebras. Next, we show affine Rota-Baxter groups can naturally give rise to the affine skew braces defined by Angiono, Galindo, and Vendramin. Conversely, any affine skew brace can be embedded into an affine Rota-Baxter group. By linking these to the relationship between affine skew braces and Hopf co-braces, we give new connections between co-Rota-Baxter Hopf algebras and Hopf co-braces. Finally, we propose the study of solutions to the Yang-Baxter equation within the framework of affine schemes, demonstrating that affine skew braces naturally give rise to such solutions.

math.GR↗

arXiv2026

On the Robustness of Machine Unlearning for Vision-Language Models

Vision-language models (VLMs) may memorize undesirable information from training data, motivating growing interest in machine unlearning. In this work, we present the first systematic survey and robustness analysis of VLM unlearning. We provide a comprehensive taxonomy and review of existing VLM unlearning methods, together with unified evaluations under multiple prompt settings. We then propose three attack paradigms to examine whether forgotten multimodal knowledge can be reactivated through contextual prompting or downstream retraining. Extensive experiments show that many existing methods remain vulnerable under these attacks, indicating that current approaches often hide rather than fully remove target knowledge. Our study provides new insights into the robustness and limitations of current VLM unlearning methods and highlights the need for more reliable multimodal unlearning strategies. Code is available at https://github.com/XMUDeepLIT/VLM-UnL-Attack.

cs.CV↗

arXiv2026

Object Hallucination-Free Reinforcement Unlearning for Vision-Language Models

Vision-language models (VLMs) raise growing concerns about privacy, copyright, and bias, motivating machine unlearning to remove sensitive knowledge. However, existing methods primarily fine-tune the language decoder, leading to superficial forgetting that fails to erase underlying visual representations and often introduces object hallucination. We propose HFRU, a reinforcement unlearning framework that operates on the vision encoder for deep semantic removal. Our two-stage approach combines alignment disruption with GRPO-based optimization using a composite reward, including an abstraction reward that encourages semantically valid substitutions and mitigates hallucinations. Experiments on object recognition and face identity tasks show that HFRU achieves over 98% forgetting and retention performance, while introducing negligible object hallucination, significantly outperforming prior methods.Our code and implementation details are available at https://github.com/XMUDeepLIT/HFRU.

cs.CV↗

arXiv2026

RoboClaw: An Agentic Framework for Scalable Long-Horizon Robotic Tasks

Vision-Language-Action (VLA) systems have shown strong potential for language-driven robotic manipulation. However, scaling them to long-horizon tasks remains challenging. Existing pipelines typically separate data collection, policy learning, and deployment, resulting in heavy reliance on manual environment resets and brittle multi-policy execution. We present RoboClaw, an agentic robotics framework that unifies data collection, policy learning, and task execution under a single VLM-driven controller. At the policy level, RoboClaw introduces Entangled Action Pairs (EAP), which couple forward manipulation behaviors with inverse recovery actions to form self-resetting loops for autonomous data collection. This mechanism enables continuous on-policy data acquisition and iterative policy refinement with minimal human intervention. During deployment, the same agent performs high-level reasoning and dynamically orchestrates learned policy primitives to accomplish long-horizon tasks. By maintaining consistent contextual semantics across collection and execution, RoboClaw reduces mismatch between the two phases and improves multi-policy robustness. Experiments in real-world manipulation tasks demonstrate improved stability and scalability compared to conventional open-loop pipelines, while significantly reducing human effort throughout the robot lifecycle, achieving a 25% improvement in success rate over baseline methods on long-horizon tasks and reducing human time investment by 53.7%.

cs.RO↗

arXiv2026

ExPosST: Explicit Positioning with Adaptive Masking for LLM-Based Simultaneous Machine Translation

Large language models (LLMs) have recently demonstrated promising performance in simultaneous machine translation (SimulMT). However, applying decoder-only LLMs to SimulMT introduces a positional mismatch, which leads to a dilemma between decoding efficiency and positional consistency. Existing approaches often rely on specific positional encodings or carefully designed prompting schemes, and thus fail to simultaneously achieve inference efficiency, positional consistency, and broad model compatibility. In this work, we propose ExPosST, a general framework that resolves this dilemma through explicit position allocation. ExPosST reserves fixed positional slots for incoming source tokens, enabling efficient decoding with KV cache across different positional encoding methods. To further bridge the gap between fine-tuning and inference, we introduce a policy-consistent fine-tuning strategy that aligns training with inference-time decoding behavior. Experiments across multiple language pairs demonstrate that ExPosST effectively supports simultaneous translation under diverse policies.

cs.CL↗

arXiv2026

Genie Centurion: Accelerating Scalable Real-World Robot Training with Human Rewind-and-Refine Guidance

While Vision-Language-Action (VLA) models show strong generalizability in various tasks, real-world deployment of robotic policy still requires large-scale, high-quality human expert demonstrations. However, data collection via human teleoperation requires continuous operator attention, which is costly, hard to scale. To address this, we propose Genie Centurion (GCENT), a scalable and general data collection paradigm based on human rewind-and-refine guidance, enabling robots' interactive learning in deployment. GCENT starts at an imperfect policy and improves over time. When the robot execution failures occur, GCENT allows robots to revert to a previous state with a rewind mechanism, after which a teleoperator provides corrective demonstrations to refine the policy. This framework supports a one-human-to-many-robots supervision scheme with a Task Sentinel module, which autonomously predicts task success and solicits human intervention when necessary. Empirical results show that GCENT achieves up to 40% higher task success rates than state-of-the-art data collection methods, and reaches comparable performance using less than half the data in long-horizon and precise tasks. We also quantify the data yield-to-effort ratio under multi-robot scenarios, demonstrating GCENT's potential for scalable and cost-efficient robot policy training in real-world environments.

cs.RO↗

arXiv2025

Text-based Aerial-Ground Person Retrieval

This work introduces Text-based Aerial-Ground Person Retrieval (TAG-PR), which aims to retrieve person images from heterogeneous aerial and ground views with textual descriptions. Unlike traditional Text-based Person Retrieval (T-PR), which focuses solely on ground-view images, TAG-PR introduces greater practical significance and presents unique challenges due to the large viewpoint discrepancy across images. To support this task, we contribute: (1) TAG-PEDES dataset, constructed from public benchmarks with automatically generated textual descriptions, enhanced by a diversified text generation paradigm to ensure robustness under view heterogeneity; and (2) TAG-CLIP, a novel retrieval framework that addresses view heterogeneity through a hierarchically-routed mixture of experts module to learn view-specific and view-agnostic features and a viewpoint decoupling strategy to decouple view-specific features for better cross-modal alignment. We evaluate the effectiveness of TAG-CLIP on both the proposed TAG-PEDES dataset and existing T-PR benchmarks. The dataset and code are available at https://github.com/Flame-Chasers/TAG-PR.

cs.CV↗

arXiv2025

Multimodal Large Language Models for Text-rich Image Understanding: A Comprehensive Review

The recent emergence of Multi-modal Large Language Models (MLLMs) has introduced a new dimension to the Text-rich Image Understanding (TIU) field, with models demonstrating impressive and inspiring performance. However, their rapid evolution and widespread adoption have made it increasingly challenging to keep up with the latest advancements. To address this, we present a systematic and comprehensive survey to facilitate further research on TIU MLLMs. Initially, we outline the timeline, architecture, and pipeline of nearly all TIU MLLMs. Then, we review the performance of selected models on mainstream benchmarks. Finally, we explore promising directions, challenges, and limitations within the field.

cs.CV↗

arXiv2023

Residual Stress-Driven Non-Euclidean Morphing in Origami Structures

Non-Euclidean surfaces are ubiquitous in numerous engineering fields, such as automotive, aerospace, and biomedical engineering domains. Morphing origami has numerous potential engineering applications, including soft robots, mechanical metamaterials, antennas, aerospace structures, and biomedical devices, owing to its intrinsic morphing features from two-dimensional (2D) planes to three-dimensional (3D) surfaces. However, the current one-dimensional (1D) hinge deformation-driven transformation of foldable origami with rigid or slightly deformable panels cannot achieve a 3D complex and large curvilinear morphing. Moreover, most active origami structures use thin hinges with soft materials on their creases, thus resulting in a lower load capability. This study proposes a novel origami morphing method that demonstrates large free-form surface morphing, e.g., Euclidean to non-Euclidean surface morphing with shape-locking. We embedded tensorial anisotropic stress in origami panels during the extrusion-based 3D printing of shape memory polymers. The extrusion-based 3D printing of isotropic shape memory polymers can produce tensorial anisotropic stress in origami panels during fabrication, which can realize large non-Euclidean surface morphing with multiple deformation modes. The connecting topology of the origami unit cells influences the global morphing behavior owing to the interaction of the deformation of adjacent panels. Non-Euclidean morphing integrated with four-dimensional (4D) printing can provide multimodal shape locking at material and structural levels. The non-Euclidean surface morphing caused by tensorial residual stress in the panel during 3D printing expands the design space of origami and kirigami structures.

cond-mat.mtrl-sci↗

arXiv2021

Deformation characteristics and mechanical properties of a non-rigid square-twist origami structure with rotational symmetry

Non-rigid origami patterns could provide more versatile performance than their rigid counterparts in the design of mechanical metamaterials owing to the simultaneous deformation of facets and creases, but their complex deformation modes make quantitative characterization and programmability of mechanical properties a challenging task. Here, we investigated the tensile behavior of a non-rigid square-twist origami structure with rotational symmetry by combining biaxial tension experiments and finite element modeling. A three-stage deformation process, including tightening, unlocking, and flattening, of the structure was unveiled through a detailed analysis of facet distortion and crease rotation, and the relationship between structure deformation and several key features in the energy, force, and stiffness curve was obtained. Based on the analysis, an empirical model was built to correlate the geometric and material parameters of the structure and its deformation energy, initial peak force, and maximum stiffness, which were further validated through experiments. Using the model, the mechanical properties of the structure can be accurately predicted and programmed based on specific engineering requirements, thereby serving the development of new programmable mechanical metamaterials based on the family of square-twist origami.

cond-mat.soft↗

arXiv2020

Rigid Foldability and Mountain-Valley Crease Assignments of Square-Twist Origami Pattern

Rigid foldability allows an origami pattern to fold about crease lines without twisting or stretching component panels. It enables folding of rigid materials, facilitating the design of foldable structures. Recent study shows that rigid foldability is affected by the mountain-valley crease (M-V) assignment of an origami pattern. In this paper, we investigate the rigid foldability of the square-twist origami pattern with diverse M-V assignments by a kinematic method based on the motion transmission path. Four types of square-twist origami patterns are analyzed, among which two are found rigidly foldable, while the other two are not. The explicit kinematic equations of the rigid cases are derived based on the kinematic equivalence between the rigid origami pattern and the closed-loop network of spherical 4R linkages. We also propose a crease-addition method to convert the rigid foldability of the non-rigid patterns. The motion compatibility conditions of the modified patterns are checked, which verify the rigid foldability of the modified patterns. The kinematic analysis reveals the bifurcation behaviour of the modified patterns. This work not only helps to deepen our understanding on the rigid foldability of origami patterns and its relationship with the M-V assignments, but also provides us an effective way to create more rigidly foldable origami patterns from non rigid ones.

physics.app-ph↗