arXiv subjects

Xiaoxu Xu

Publications and source records attributed to Xiaoxu Xu.

18 recordsLinked to original sources

arXiv2026

InternVLA-A1.5: Unifying Understanding, Latent Foresight, and Action for Compositional Generalization

Unified models for robot manipulation aim to equip one policy with both the semantic priors of pretrained VLMs and the physical dynamics learned through future prediction. In practice, existing designs tend to erode the semantics of the pretrained backbone, suffer interference among heterogeneous objectives, and learn future prediction from scratch in pixel space, leaving the dynamics priors of pretrained video generators unexploited. We present InternVLA-A1.5, which builds the policy on a native VLM backbone that keeps training on VQA and subtask prediction, and attaches a lightweight unified expert for continuous action generation. Future prediction is recast as a latent-querying problem, where a small set of learnable foresight tokens condenses the task-relevant future into a compact latent code under the supervision of a frozen pretrained video generation model, so the policy inherits world-model dynamics priors without ever learning pixel-level generation. The video branch is discarded at inference, keeping real-time control. Pretrained on 1.2M robot episodes and 3M multimodal samples, InternVLA-A1.5 achieves the best overall results on all six simulation benchmarks. In the real world, the preserved semantics deliver the strongest compositional generalization on held-out instruction bindings, and the two designs together sustain long-horizon execution.

cs.RO↗

arXiv2026

FutureVLA: Joint Visuomotor Prediction for Vision-Language-Action Model

Predictive foresight is important to intelligent embodied agents. Since the motor execution of a robot is intrinsically constrained by its visual perception of environmental geometry, effectively anticipating the future requires capturing this tightly coupled visuomotor interplay. While recent vision-language-action models attempt to incorporate future guidance, they struggle with this joint modeling. Existing explicit methods divert capacity to task-irrelevant visual details, whereas implicit methods relying on sparse frame pairs disrupt temporal continuity. By heavily relying on visual reconstruction, these methods become visually dominated, entangling static scene context with dynamic action intent. We argue that effective joint visuomotor predictive modeling requires both temporal continuity and visually-conditioned supervision decoupling. To this end, we propose FutureVLA, featuring a novel Joint Visuomotor Predictive Architecture. FutureVLA is designed to extract joint visuomotor embeddings by first decoupling visual and motor information, and then jointly encoding generalized physical priors. Specifically, in the pretraining stage, we leverage heterogeneous manipulation datasets and introduce a Joint Visuomotor Gating mechanism to structurally separate visual state preservation from temporal action modeling. It allows the motor stream to focus on continuous physical dynamics while explicitly querying visual tokens for environmental constraints, yielding highly generalizable joint visuomotor embeddings. Subsequently, in the post-training stage, we employ a latent embeddings alignment strategy, enabling diverse downstream VLA models to internalize these temporal priors without modifying their inference architectures. Extensive experiments demonstrate that FutureVLA consistently improves VLA frameworks.

cs.RO↗

arXiv2026

Asymptotic formulas for phase recovering from phaseless data of biharmonic waves at a fixed frequency

This paper focuses on phase retrieval from phaseless total-field data in biharmonic scattering problems. We prove that a phased biharmonic wave can be uniquely determined by the modulus of the total biharmonic wave within a nonempty domain. As a direct corollary, the uniqueness for the inverse biharmonic scattering problem with phaseless total-field data is established. Moreover, using the Atkinson-type asymptotic expansion, we derive explicit asymptotic formulas for the problem of phase retrieval.

math.AP↗

arXiv2025

Simultaneously recover two constant coefficients and a polygon with a single pair of Cauchy data for the Helmholtz equation

This paper is concerned with an inverse boundary value problem for the Helmholtz equation over a bounded domain. The aim is to reconstruct two constant coefficients together with the location and shape of a Dirichlet polygonal obstacle from a single pair of Cauchy data. Uniqueness results are verified under some a priori assumptions and the one-wave factorization method has been adapted to recover the polygonal obstacle as well as the two coefficients. A modified factorization using the Dirichlet-to-Neumann operator is employed to overcome difficulties arising from possible eigenvalues. Intensive numerical examples indicate that our method is efficient.

math.AP↗

arXiv2025

DBGroup: Dual-Branch Point Grouping for Weakly Supervised 3D Semantic Instance Segmentation

Weakly supervised 3D instance segmentation is essential for 3D scene understanding, especially as the growing scale of data and high annotation costs associated with fully supervised approaches. Existing methods primarily rely on two forms of weak supervision: one-thing-one-click annotations and bounding box annotations, both of which aim to reduce labeling efforts. However, these approaches still encounter limitations, including labor-intensive annotation processes, high complexity, and reliance on expert annotators. To address these challenges, we propose \textbf{DBGroup}, a two-stage weakly supervised 3D instance segmentation framework that leverages scene-level annotations as a more efficient and scalable alternative. In the first stage, we introduce a Dual-Branch Point Grouping module to generate pseudo labels guided by semantic and mask cues extracted from multi-view images. To further improve label quality, we develop two refinement strategies: Granularity-Aware Instance Merging and Semantic Selection and Propagation. The second stage involves multi-round self-training on an end-to-end instance segmentation network using the refined pseudo-labels. Additionally, we introduce an Instance Mask Filter strategy to address inconsistencies within the pseudo labels. Extensive experiments demonstrate that DBGroup achieves competitive performance compared to sparse-point-level supervised 3D instance segmentation methods, while surpassing state-of-the-art scene-level supervised 3D semantic segmentation approaches. Code is available at https://github.com/liuxuexun/DBGroup.

cs.CV↗

arXiv2025

3D Weakly Supervised Semantic Segmentation via Class-Aware and Geometry-Guided Pseudo-Label Refinement

3D weakly supervised semantic segmentation (3D WSSS) aims to achieve semantic segmentation by leveraging sparse or low-cost annotated data, significantly reducing reliance on dense point-wise annotations. Previous works mainly employ class activation maps or pre-trained vision-language models to address this challenge. However, the low quality of pseudo-labels and the insufficient exploitation of 3D geometric priors jointly create significant technical bottlenecks in developing high-performance 3D WSSS models. In this paper, we propose a simple yet effective 3D weakly supervised semantic segmentation method that integrates 3D geometric priors into a class-aware guidance mechanism to generate high-fidelity pseudo labels. Concretely, our designed methodology first employs Class-Aware Label Refinement module to generate more balanced and accurate pseudo labels for semantic categrories. This initial refinement stage focuses on enhancing label quality through category-specific optimization. Subsequently, the Geometry-Aware Label Refinement component is developed, which strategically integrates implicit 3D geometric constraints to effectively filter out low-confidence pseudo labels that fail to comply with geometric plausibility. Moreover, to address the challenge of extensive unlabeled regions, we propose a Label Update strategy that integrates Self-Training to propagate labels into these areas. This iterative process continuously enhances pseudo-label quality while expanding label coverage, ultimately fostering the development of high-performance 3D WSSS models. Comprehensive experimental validation reveals that our proposed methodology achieves state-of-the-art performance on both ScanNet and S3DIS benchmarks while demonstrating remarkable generalization capability in unsupervised settings, maintaining competitive accuracy through its robust design.

cs.CV↗

arXiv2025

Weakly-Supervised 3D Visual Grounding based on Visual Language Alignment

Learning to ground natural language queries to target objects or regions in 3D point clouds is quite essential for 3D scene understanding. Nevertheless, existing 3D visual grounding approaches require a substantial number of bounding box annotations for text queries, which is time-consuming and labor-intensive to obtain. In this paper, we propose 3D-VLA, a weakly supervised approach for 3D visual grounding based on Visual Linguistic Alignment. Our 3D-VLA exploits the superior ability of current large-scale vision-language models (VLMs) on aligning the semantics between texts and 2D images, as well as the naturally existing correspondences between 2D images and 3D point clouds, and thus implicitly constructs correspondences between texts and 3D point clouds with no need for fine-grained box annotations in the training procedure. During the inference stage, the learned text-3D correspondence will help us ground the text queries to the 3D target objects even without 2D images. To the best of our knowledge, this is the first work to investigate 3D visual grounding in a weakly supervised manner by involving large scale vision-language models, and extensive experiments on ReferIt3D and ScanRefer datasets demonstrate that our 3D-VLA achieves comparable and even superior results over the fully supervised methods.

cs.CV↗

arXiv2024

LESS: Label-Efficient and Single-Stage Referring 3D Segmentation

Referring 3D Segmentation is a visual-language task that segments all points of the specified object from a 3D point cloud described by a sentence of query. Previous works perform a two-stage paradigm, first conducting language-agnostic instance segmentation then matching with given text query. However, the semantic concepts from text query and visual cues are separately interacted during the training, and both instance and semantic labels for each object are required, which is time consuming and human-labor intensive. To mitigate these issues, we propose a novel Referring 3D Segmentation pipeline, Label-Efficient and Single-Stage, dubbed LESS, which is only under the supervision of efficient binary mask. Specifically, we design a Point-Word Cross-Modal Alignment module for aligning the fine-grained features of points and textual embedding. Query Mask Predictor module and Query-Sentence Alignment module are introduced for coarse-grained alignment between masks and query. Furthermore, we propose an area regularization loss, which coarsely reduces irrelevant background predictions on a large scale. Besides, a point-to-point contrastive loss is proposed concentrating on distinguishing points with subtly similar features. Through extensive experiments, we achieve state-of-the-art performance on ScanRefer dataset by surpassing the previous methods about 3.7% mIoU using only binary labels. Code is available at https://github.com/mellody11/LESS.

cs.CV↗

arXiv2024

3D Weakly Supervised Semantic Segmentation with 2D Vision-Language Guidance

In this paper, we propose 3DSS-VLG, a weakly supervised approach for 3D Semantic Segmentation with 2D Vision-Language Guidance, an alternative approach that a 3D model predicts dense-embedding for each point which is co-embedded with both the aligned image and text spaces from the 2D vision-language model. Specifically, our method exploits the superior generalization ability of the 2D vision-language models and proposes the Embeddings Soft-Guidance Stage to utilize it to implicitly align 3D embeddings and text embeddings. Moreover, we introduce the Embeddings Specialization Stage to purify the feature representation with the help of a given scene-level label, specifying a better feature supervised by the corresponding text embedding. Thus, the 3D model is able to gain informative supervisions both from the image embedding and text embedding, leading to competitive segmentation performances. To the best of our knowledge, this is the first work to investigate 3D weakly supervised semantic segmentation by using the textual semantic information of text category labels. Moreover, with extensive quantitative and qualitative experiments, we present that our 3DSS-VLG is able not only to achieve the state-of-the-art performance on both S3DIS and ScanNet datasets, but also to maintain strong generalization capability.

cs.CV↗

arXiv2024

An inverse obstacle problem with a single pair of Cauchy data: Laplace's equation case

This paper is concerned with an inverse obstacle problem for the Laplace's equation. The aim is to recover the constant conductivity coefficient in the equation and the boundary of a Dirichlet polygonal obstacle from a single pair of Cauchy data. Uniqueness results are established under some a priori assumptions on the input boundary value data. A domain-defined sampling method, based on the factorization method originating from inverse acoustic scattering, has been proposed to recover both the constant conductivity coefficient and the polygonal obstacle. A hybrid strategy, which combines the sampling method and iterative scheme, is employed {\color{hgh}to reconstruct} the location and shape of the obstacle. Numerical examples indicate that our method is efficient.

math.AP↗

arXiv2024

Detection of a piecewise linear crack with one incident wave

This paper is concerned with inverse crack scattering problems for time-harmonic acoustic waves. We prove that a piecewise linear crack with the sound-soft boundary condition in two dimensions can be uniquely determined by the far-field data corresponding to a single incident plane wave or point source. We propose two non-iterative methods for imaging the location and shape of a crack. The first one is a contrast sampling method, while the second one is a variant of the classical factorization method but only with one incoming wave. Newton's iteration method is then employed for getting a more precise reconstruction result. Numerical examples are presented to show the effectiveness of the proposed hybrid method.

math.NA↗

arXiv2023

Uniqueness in inverse scattering with phaseless near-field data generated by superpositions of two incident plane waves

This paper is concerned with the uniqueness in inverse acoustic and electromagnetic scattering with phaseless near-field data generated by superpositions of two incident plane waves at a fixed frequency. It can be proved that the unknown scatterer can be uniquely determined by the phaseless near-field data. The proof is based on the analysis of the phase information and the application of Rellich's lemma together with the Green's formula for the radiating solutions to the Helmholtz equation or the Stratton--Chu formula for the radiating solutions to the Maxwell equations.

math.AP↗

arXiv2022

Uniqueness in inverse diffraction grating problems with infinitely many plane waves at a fixed frequency

This paper is concerned with the inverse diffraction problems by a periodic curve with Dirichlet boundary condition in two dimensions. It is proved that the periodic curve can be uniquely determined by the near-field measurement data corresponding to infinitely many incident plane waves with distinct directions at a fixed frequency. Our proof is based on Schiffer's idea which consists of two ingredients: i) the total fields for incident plane waves with distinct directions are linearly independent, and ii) there exist only finitely many linearly independent Dirichlet eigenfunctions in a bounded domain or in a closed waveguide under additional assumptions on the waveguide boundary. Based on the Rayleigh expansion, we prove that the phased near-field data can be uniquely determined by the phaseless near-field data in a bounded domain, with the exception of a finite set of incident angles. Such a phase retrieval result leads to a new uniqueness result for the inverse grating diffraction problem with phaseless near-field data at a fixed frequency. Since the incident direction determines the quasi-periodicity of the boundary value problem, our inverse issues are different from the existing results of [Htttlich & Kirsch, Inverse Problems 13 (1997): 351-361] where fixed-direction plane waves at multiple frequencies were considered.

math.AP↗

arXiv2019

Uniqueness in inverse electromagnetic scattering problem with phaseless far-field data at a fixed frequency

This paper is concerned with uniqueness in inverse electromagnetic scattering with phaseless far-field pattern at a fixed frequency. In our previous work [{\em SIAM J. Appl. Math.} {\bf 78} (2018), 3024-3039], by adding a known reference ball into the acoustic scattering system, it was proved that the impenetrable obstacle and the index of refraction of an inhomogeneous medium can be uniquely determined by the acoustic phaseless far-field patterns generated by infinitely many sets of superpositions of two plane waves with different directions at a fixed frequency. In this paper, we extend these uniqueness results to the inverse electromagnetic scattering case. The phaseless far-field data are the modulus of the tangential component in the orientations $\mathbf{e}_ϕ$ and $\mathbf{e}_θ$, respectively, of the electric far-field pattern measured on the unit sphere and generated by infinitely many sets of superpositions of two electromagnetic plane waves with different directions and polarizations. Our proof is mainly based on Rellich's lemma and the Stratton--Chu formula for radiating solutions to the Maxwell equations.

math.AP↗

arXiv2019

Uniqueness in inverse acoustic and electromagnetic scattering with phaseless near-field data at a fixed frequency

This paper is concerned with uniqueness results in inverse acoustic and electromagnetic scattering problems with phaseless total-field data at a fixed frequency. Motivated by our previous work ({\em SIAM J. Appl. Math. \bf78} (2018), 1737-1753), where uniqueness results were proved for inverse acoustic scattering with phaseless far-field data generated by superpositions of two plane waves as the incident waves at a fixed frequency, in this paper, we use superpositions of two point sources as the incident fields at a fixed frequency and measure the modulus of the acoustic total-field (called phaseless acoustic near-field data) on two spheres enclosing the scatterers generated by such incident fields on the two spheres. Based on this idea, we prove that the impenetrable bounded obstacle or the index of refraction of an inhomogeneous medium can be uniquely determined from the phaseless acoustic near-field data at a fixed frequency. Moreover, the idea is also extended to the electromagnetic case, and it is proved that the impenetrable bounded obstacle or the index of refraction of an inhomogeneous medium can be uniquely determined by the phaseless electric near-field data at a fixed frequency, that is, the modulus of the tangential component with the orientations $\mathbf e_ϕ$ and $\mathbf e_θ$, respectively, of the electric total-field measured on a sphere enclosing the scatters and generated by superpositions of two electric dipoles at a fixed frequency located on the measurement sphere and another bigger sphere with the polarization vectors $\mathbf e_ϕ$ and $\mathbf e_θ$, respectively. As far as we know, this is the first uniqueness result for three-dimensional inverse electromagnetic scattering with phaseless near-field data.

math.AP↗

arXiv2018

Uniqueness and direct imaging method for inverse scattering by locally rough surfaces with phaseless near-field data

This paper is concerned with inverse scattering of plane waves by a locally perturbed infinite plane (which is called a locally rough surface) with the modulus of the total-field data (also called the phaseless near-field data) at a fixed frequency in two dimensions. We consider the case where a Dirichlet boundary condition is imposed on the locally rough surface. This problem models inverse scattering of plane acoustic waves by a one-dimensional sound-soft, locally rough surface; it also models inverse scattering of plane electromagnetic waves by a locally perturbed, perfectly reflecting, infinite plane in the TE polarization case. We prove that the locally rough surface is uniquely determined by the phaseless near-field data generated by a countably infinite number of plane waves and measured on an open domain above the locally rough surface. Further, a direct imaging method is proposed to reconstruct the locally rough surface from the phaseless near-field data generated by plane waves and measured on the upper part of the circle with a sufficiently large radius. Theoretical analysis of the imaging algorithm is derived by making use of properties of the scattering solution and results from the theory of oscillatory integrals (especially the method of stationary phase). Moreover, as a by-product of the theoretical analysis, a similar direct imaging method with full far-field data is also proposed to reconstruct the locally rough surface. Finally, numerical experiments are carried out to demonstrate that the imaging algorithm with phaseless near-field data and full far-field data are fast, accurate and very robust with respect to noise in the data.

math.NA↗

arXiv2018

Uniqueness in inverse scattering problems with phaseless far-field data at a fixed frequency. II

This paper is concerned with uniqueness in inverse acoustic scattering with phaseless far-field data at a fixed frequency. In our previous work ({\em SIAM J. Appl. Math. \bf78} (2018), 1737-1753), by utilizing spectral properties of the far-field operator we proved for the first time that the obstacle and the index of refraction of an inhomogeneous medium can be uniquely determined by the phaseless far-field patterns generated by infinitely many sets of superpositions of two plane waves with different directions at a fixed frequency under the a priori assumption that the obstacle is known to be a sound-soft or non-absorbing impedance obstacle and the index of refraction $n$ of the inhomogeneous medium is real-valued and satisfies that either $n-1\ge c_1$ or $n-1\le-c_1$ in the support of $n-1$ for some positive constant $c_1$. In this paper, we remove the a priori assumption on the obstacle and the index of refraction of the inhomogeneous medium by adding a reference ball to the scattering system together with a simpler method of using Rellich's lemma and Green's representation formula for the scattering solutions. Further, our new method is also used to prove uniqueness in determining a locally rough surface from the phaseless far-field patterns corresponding to infinitely many sets of superpositions of two plane waves with different directions as the incident fields at a fixed frequency.

math.AP↗

arXiv2017

Uniqueness in inverse scattering problems with phaseless far-field data at a fixed frequency

This paper is concerned with uniqueness in inverse acoustic scattering with phaseless far-field data at a fixed frequency. The main difficulty of this problem is the so-called translation invariance property of the modulus of the far-field pattern generated by one plane wave as the incident field. Based on our previous work (J. Comput. Phys. 345 (2017), 58-73), the translation invariance property of the phaseless far-field pattern can be broken by using infinitely many sets of superpositions of two plane waves as the incident fields at a fixed frequency. In this paper, we prove that the obstacle and the index of refraction of an inhomogeneous medium can be uniquely determined by the phaseless far-field patterns generated by infinitely many sets of superpositions of two plane waves with different directions at a fixed frequency under the condition that the obstacle is a priori known to be a sound-soft or non-absorbing impedance obstacle and the index of refraction $n$ of the inhomogeneous medium is real-valued and satisfies that either $n-1\ge c_1$ or $n-1\le-c_1$ in the support of $n-1$ for some positive constant $c_1$. To the best of our knowledge, this is the first uniqueness result in inverse scattering with phaseless far-field data. Our proofs are based essentially on the limit of the normalized eigenvalues of the far-field operators which is also established in this paper by using a factorization of the far-field operators.

math.AP↗