arXiv subjects

Yue Xin

Publications and source records attributed to Yue Xin.

At least 19 recordsLinked to original sources

arXiv2026

ResMerge: Residual Policy Learning and Merging for Continual Adaptation of Pre-trained Robot Policies

Continual adaptation to new tasks is essential for robots deployed in the real world, but adapting pre-trained policies remains challenging due to parameter growth and the need for continual demonstration collection. Parameter-Efficient Continual Fine-Tuning (PECFT) isolates task-specific updates, but adapter-per-task designs require demonstrations and grow with the task set. We propose ResMerge, a two-stage framework that learns mergeable residual policies atop a frozen pre-trained policy: residual reinforcement learning (RL) acquires task-specific corrections without additional expert demonstrations during adaptation, and parameter-space merging consolidates compatible residuals to reduce stored parameters. We introduce Residual-Field Consistency (RFC), a simple, bounded, and interpretable metric that assesses compatibility in residual action-field space by comparing correction direction and magnitude. Experiments on LIBERO across two backbones and on a real-world Franka robot show improved continual learning performance over strong PECFT baselines, while selective merging reduces the residual-parameter footprint. Project website: https://resmerge.github.io/.

cs.RO↗

arXiv2026

The $ϕ$-conjugation of quaternionic matrices and generalized Autonne-Takagi factorization

Let $ϕ$ be a quaternion of modulus $1$. In this article, we study some topics related to $ϕ$-conjugation for quaternionic matrices, including $ϕ$-Hermitian matrices, $ϕ$-conjugate normal matrices, unitary $ϕ$-congruence and $ϕ$-HSH decomposition (decomposition of a $ϕ$-Hermitian matrix and a skew $ϕ$-Hermitian matrix). In particular, we generalize the Autonne-Takagi factorization of quaternion $ϕ$-Hermitian matrices for all unit quaternion $ϕ$. This gives an affirmative answer to a problem proposed by R. Horn and F. Zhang in the paper ``A generalization of the complex Autonne-Takagi factorization to quaternion matrices, Linear Multilinear A. 60: 1239--1244, 2012''.

math.RA↗

arXiv2026

Embodied Skill Bench: Benchmarking How Well Agents with Embodied Skills Perform on Physical Tasks

LLM agents increasingly control robots using VLA policies, perception models, and procedural knowledge. These modules are difficult to reuse because their interfaces, embodiment assumptions, and conditions of use differ, and their benefits on physical tasks remain unclear. We package them as embodied skills that extend agent skills with explicit embodiment assumptions and conditions of use, building 250 action, perception, and experience skills from 1,052 resources. On this basis we introduce EmbodiedSkillBench, which runs an agent in a closed observation-action loop and records success, actions, tokens, and wall-clock time. It supports three modes: native control, skill-only execution, and Agent-Directed Reuse (ADR), a mode we propose in which the agent keeps its native controls and decides for each skill whether to invoke, modify, or ignore it. Across four agents on 100 perturbed LIBERO-Pro tasks, we find that skills help an agent only as far as they exceed what it can already do. Skill-only execution lifts GPT-5.6-Sol from 11% to 63% and Kimi-K3 and Claude-Opus-5 by about 20 points, while cutting their token usage by 58% to 85%. Yet all four agents then end up between 61% and 64%, and GPT-6-Astra falls from 92% under native control to 61%. This suggests that as agents grow stronger, the role of embodied skills shifts from supplying missing capability to reducing cost: GPT-6-Astra already solves most tasks through native control, so restricting it to skills removes more than it adds. Under ADR, which lets it keep native control and call skills selectively, GPT-6-Astra recovers to 85%, within 7 points of native control, while using 14.1% fewer tokens and 11.9% less time. With Jev as a skill choosing motion directions and gripper actions, it reaches 81% with 50% fewer tokens and 57% less time. The code is available at https://github.com/HanZhi306/EmbodiedSkill-Bench.

cs.RO↗

arXiv2026

Ambiguity-Free Inertial Measurement with Multi-Wavelength Atom Interferometry

White-light interferometry enables ambiguity-free localization by synthesizing interference envelopes from multiple optical wavelengths, but no analogous capability has been realized for coherent matter waves. Here we report the first experimental demonstration of multi-wavelength atom interferometry, establishing the matter-wave counterpart of white-light interferometry. By exploiting counter-propagating atomic beams as multi-wavelength matter wave sources and synthesizing interference envelopes from their spectral components, we realize inertial measurements based on envelope localization rather than conventional fringe-phase estimation. The resulting multi-scale interferometric response provides ambiguity-free operation, a well-defined rotational scale factor, and reduced sensitivity to initial phase bias. As a proof of principle, we demonstrate simultaneous dual-axis rotation and acceleration sensing and directly resolve the phase ambiguity that fundamentally limits conventional open-loop atom interferometers. We further measure the Earth's rotation with a relative error of 4.3% and a long-term stability of 93 ppm at an averaging time of 15,000 s. Our results establish multi-wavelength atom interferometry as a new paradigm for coherent matter-wave sensing, extending the principles of white-light interferometry to atom optics and opening new opportunities for inertial sensing, geodesy, precision metrology, and inertial navigation.

physics.atom-ph↗

arXiv2026

URDF-Anything+: End-to-End Generation for Simulation-Ready Articulated Assets

Articulated objects are fundamental for robotics, simulation of physics, and interactive virtual environments. However, recovering them from visual observations is inherently challenging, as images provide only partial and ambiguous cues about both part geometry and their underlying kinematic structure. Existing approaches typically rely on multi-stage pipelines, retrieval from asset libraries, or explicit part segmentation. We present URDF-Anything+, an end-to-end autoregressive diffusion framework that generates simulation-ready URDF models directly from a single RGB image. Conditioned on visual observations and object geometry, URDF-Anything+ operates in a structured latent space and jointly models part geometry and articulation in a unified generation process. Specifically, the model sequentially predicts each articulated part together with its associated joint parameters, while a termination token dynamically determines the number of parts. This design enables direct generation of fully executable URDFs without external retrieval or post-processing stages. Experiments on large-scale articulated object benchmarks demonstrate that URDF-Anything+ outperforms prior methods in geometric reconstruction quality, joint parameter estimation, and physical executability, while being substantially more efficient than existing multi-stage approaches. Furthermore, the generated URDFs serve as faithful digital twins, enabling the zero-shot transfer of manipulation policies trained purely in simulation.

cs.RO↗

arXiv2026

The right invariant metric on the analytic automorphism group of the unit open disk induced by maximal modulus

In this paper, we study the right invariant metric $d_{H^{\infty}}$ on the analytic automorphism group $\rm{Aut}(\mathbb{D})$ of the unit open disk $\mathbb{D}$ induced by maximal modulus, that is, $d_{H^{\infty}}(φ, ψ)=\sup_{z\in\mathbb{D}}|φ(z)-ψ(z)|$ for any $φ, ψ\in \rm{Aut}(\mathbb{D})$. We give the explicit formula of the right invariant metric $d_{H^{\infty}}$ and characterize the almost regular Finsler geometric structure of $(\rm{Aut}(\mathbb{D}), d_{H^{\infty}})$.

math.CV↗

arXiv2026

Spectral-Domain Coherent Control of Broadband Raman Coupling in Atom Interferometry

The performance of atom interferometers is commonly limited by the finite spectral acceptance of atomic beam splitters and mirrors, which restricts efficient coupling to atoms with large Doppler shifts and reduces the usable atomic flux. Here, we demonstrate spectral-domain coherent control of Raman coupling by engineering its effective two-photon spectrum. By synthesizing multiple frequency components, the Raman interaction simultaneously addresses a broad range of atomic velocities, effectively overcoming the conventional transit-time-limited linewidth. Implemented in a continuous atomic-beam Mach-Zehnder interferometer, where the transverse Doppler broadening is 17 times larger than the intrinsic Raman linewidth, this approach enhances the fringe contrast from 5.9(2)% to 15.1(2)%, indicating a substantial increase in effective atomic participation. Our results establish spectral-domain coherent control as a general strategy for achieving spectrally robust atom interferometry and open new opportunities for quantum sensing in systems with strong inhomogeneous broadening.

physics.atom-ph↗

arXiv2026

Closed-loop dual-channel atomic beam interferometry beyond the half-fringe limit

Atom interferometric inertial sensors offer exceptional sensitivity but are fundamentally constrained by the periodic phase response of matter-wave interference, which imposes an intrinsic half-fringe dynamic-range limit and prevents continuous inertial tracking. In multi-axis configurations, additional cross coupling between acceleration and rotation further complicates closed-loop operation. Here we demonstrate the first dual-channel closed-loop operation of an atomic beam interferometer, realizing decoupled feedback control of acceleration- and rotation-induced phases and overcoming the half-fringe limitation. Using continuous, transversely cooled $^{87}$Rb atomic beams, the interferometric phases associated with rotation and acceleration are independently extracted, tracked across multiple fringes, and actively compensated through Raman frequency modulation. This closed-loop scheme enables unambiguous measurements up to $\pm1\,\mathrm{^{\circ}/s}$ in rotation and $\pm0.17\,\mathrm{g}$ in acceleration while maintaining high fringe contrast, corresponding to nearly two orders-of-magnitude extension beyond the conventional half-fringe limit. The sensor achieves a long-term stability of $4\times10^{-4}\,\mathrm{^{\circ}/h}$ for rotation and $4\,\mathrm{μg}$ for acceleration at an averaging time of $1000\,\mathrm{s}$. By converting the intrinsically periodic interferometric response into stabilized phase-encoded inertial channels, this work establishes a new operating regime for atomic beam interferometry and advances matter-wave sensors toward practical quantum inertial navigation under dynamic conditions.

physics.atom-ph↗

arXiv2025

Persistence modules induced by inner functions

As well-known, inner functions play an important role in the study of bounded analytic function theory. In recent years, persistence module theory, as a main tool applied to Topological Data Analysis, has received widespread attention. In this paper, we aim to use persistence module theory to study inner functions. We introduce the persistence modules arised from the level sets of inner functions. Some properties of these persistence modules are shown. In particular, we prove that the persistence modules (potentially not of locally finite type) induced by a class of inner functions have interval module decompositions. Furthermore, we demonstrate that the interleaving distance of the persistence modules is continuous with respect to the supremum norm for a class of Blaschke products, which could be used to discuss the path-connectedness of Blaschke products. As an example, we provide an explicit formula for the interleaving distance of the persistence modules induced by the Blaschke products with order two.

math.AT↗

arXiv2025

SalaMAnder: Shapley-based Mathematical Expression Attribution and Metric for Chain-of-Thought Reasoning

Chain-of-Thought (CoT) prompting enhances the math reasoning capability of large language models (LLMs) to a large margin. However, the mechanism underlying such improvements remains unexplored. In this paper, we present \textbf{SalaMAnder} (\textbf{S}h\textbf{a}p\textbf{l}ey-b\textbf{a}sed \textbf{M}athematical Expression \textbf{A}ttribution a\textbf{nd} M\textbf{e}t\textbf{r}ic), a theoretically grounded methodology as well as a mathematically rigorous evaluation metric for quantifying component-level contributions in few-shot CoT reasoning. Concretely, we leverage the Shapley value for mathematical expression attribution and develop an efficient stratified sampling algorithm that significantly reduces the computational complexity. Besides, we develop the \textbf{CoSP} (\textbf{C}ardinality \textbf{o}f \textbf{S}hapley \textbf{P}ositives) metric through covariance analysis. Comprehensive validation across popular LLM models and diverse mathematical benchmarks demonstrates that the CoSP metric within our SalaMAnder framework exhibits a robust monotonic correlation with model performance, not only providing theoretical explanations for the empirical success of existing few-shot CoT but also establishing mathematically rigorous principles for prompt construction optimization. Furthermore, we verify the reliability of the explanation, based on which we unify the insights of previous work.

cs.AI↗

arXiv2025

Generalizable Geometric Image Caption Synthesis

Multimodal large language models have various practical applications that demand strong reasoning abilities. Despite recent advancements, these models still struggle to solve complex geometric problems. A key challenge stems from the lack of high-quality image-text pair datasets for understanding geometric images. Furthermore, most template-based data synthesis pipelines typically fail to generalize to questions beyond their predefined templates. In this paper, we bridge this gap by introducing a complementary process of Reinforcement Learning with Verifiable Rewards (RLVR) into the data generation pipeline. By adopting RLVR to refine captions for geometric images synthesized from 50 basic geometric relations and using reward signals derived from mathematical problem-solving tasks, our pipeline successfully captures the key features of geometry problem-solving. This enables better task generalization and yields non-trivial improvements. Furthermore, even in out-of-distribution scenarios, the generated dataset enhances the general reasoning capabilities of multimodal large language models, yielding accuracy improvements of $2.8\%\text{-}4.8\%$ in statistics, arithmetic, algebraic, and numerical tasks with non-geometric input images of MathVista and MathVerse, along with $2.4\%\text{-}3.9\%$ improvements in Art, Design, Tech, and Engineering tasks in MMMU.

cs.AI↗

arXiv2025

A Continuous Dual-Axis Atomic Interferometric Inertial Sensor

We present an interferometric inertial sensor that utilizes two counter-propagating atomic beams with transverse two-dimensional cooling. By employing three parallel and spatially aligned Raman laser beams for Doppler-sensitive Raman transitions, we successfully generate inertia-sensitive Mach-Zehnder interference fringes with an interrogation length of $2L=54\,\rm{cm}$. The sensor's capability to measure rotation and acceleration simultaneously in dynamic environments is validated through comparative analysis with classical sensors under force oscillation in different directions. Additionally, we conduct experiments on a turntable to calibrate the gyroscope's scaling factor and address nonlinearity. The angular random walk (ARW) and velocity random walk (VRW) of the senor are $3\times10^{-4}\,^\circ/\rm{\sqrt{h}}$ and $107\,\mathrmμg/\rm{\sqrt{Hz}}$, respectively, with the long-term stability reaching $9\times10^{-4}\,\rm{^\circ/h}$ for rotation and $10\,\rm{μg}$ for acceleration at an integration time of 1000s.

physics.atom-ph↗

arXiv2025

Enhancing Chain-of-Thought Reasoning with Critical Representation Fine-tuning

Representation Fine-tuning (ReFT), a recently proposed Parameter-Efficient Fine-Tuning (PEFT) method, has attracted widespread attention for significantly improving parameter efficiency by editing representation space alone. In this work, we investigate applying ReFT to complex reasoning tasks. However, directly using the native ReFT method, which modifies fixed representations at the beginning and end of each layer, yields suboptimal performance, as these fixed-position representations have uncertain impact on the outputs. We observe that, in complex reasoning tasks, there often exist certain critical representations. These representations either integrate significant information from preceding layers or regulate subsequent layer representations. Through layer-by-layer propagation, they exert a substantial influence on the final output. Naturally, fine-tuning these critical representations has the potential to greatly enhance reasoning performance. Building upon these insights, we propose Critical Representation Fine-Tuning (CRFT), a novel method that identifies and optimizes these critical representations through information flow analysis. CRFT operates within a supervised learning framework, dynamically optimizing critical representations in a low-rank linear subspace while freezing the base model. The effectiveness and efficiency of our method are validated across eight benchmarks for arithmetic and commonsense reasoning, using LLaMA and Mistral model families. Furthermore, our method also adapts effectively to few-shot settings, boosting one-shot accuracy by 16.4%. Our work highlights the untapped potential of representation-level optimization for CoT reasoning, offering a lightweight yet powerful alternative to traditional PEFT methods.

cs.CL↗

arXiv2025

Topologically conjugate classification of diagonal operators

Let $\ell^{p}$, $1\leq p<\infty$, be the Banach space of absolutely $p$-th power summable sequences and let $π_{n}$ be the natural projection to the $n$-th coordinate for $n\in\mathbb{N}$. Let $\mathfrak{W}=\{w_{n}\}_{n=1}^{\infty}$ be a bounded sequence of complex numbers. Define the operator $D_{\mathfrak{W}}: \ell^{p}\rightarrow\ell^{p}$ by, for any $x=(x_{1},x_{2},\ldots)\in \ell^p$, $π_{n}\circ D_{\mathfrak{W}}(x)=w_{n}x_{n}$ for all $n\geq1$. We call $D_{\mathfrak{W}}$ a diagonal operator on $\ell^{p}$. In this article, we study the topological conjugate classification of the diagonal operators on $\ell^{p}$. More precisely, we obtained the following results. $D_{\mathfrak{W}}$ and $D_{\vert\mathfrak{W}\vert}$ are topologically conjugate, where $\vert\mathfrak{W}\vert=\{\vert w_{n}\vert\}_{n=1}^{\infty}$. If $\inf_{n}\vert w_n\vert>1$, then $D_{\mathfrak{W}}$ is topologically conjugate to $2\mathbf{I}$, where $\mathbf{I}$ means the identity operator. Similarly, if $\inf_{n}\vert w_n\vert>0$ and $\sup_{n}\vert w_n\vert<1$, then $D_{\mathfrak{W}}$ is topologically conjugate to $\frac{1}{2}\mathbf{I}$. In addition, if $\inf_{n}\vert w_n\vert=1$ and $\inf_{n}\vert t_n\vert>1$, then $D_{\mathfrak{W}}$ and $D_{\mathfrak{T}}$ are not topologically conjugate.

math.DS↗

arXiv2024

Mix-GENEO: A Flexible Filtration for Multiparameter Persistent Homology Detects Digital Images

Two important tasks in the field of Topological Data Analysis are building practical multifiltrations on objects and using TDA to detect the geometry. Motivated by the tasks, we build multiparameter filtrations by operators on images named multi-GENEO, multi-DGENEO and mix-GENEO, and we prove the stability of both the interleaving distance and multiparameter persistence landscape of multi-GENEO with respect to the pseudometric on bounded functions. We also give the estimations of upper bound for multi-DGENEO and mix-GENEO. In practical applications, we regard image as a discrete function space, and then we build multifiltrations on the discrete function space. Finally, we construct comparable experiment on MNIST dataset to demonstrate our bifiltrations are superior to 1-parameter filtrations including lower-star filtration and upper-star filtration. For instance, 6 and 9 can be distinguished by our bifiltrations, while they cannot be distinguished by 1-parameter filtrations. The experiment results demonstrate our bifiltrations have ability to detect geometric and topological differences of digital images.

cs.CV↗

arXiv2024

Towards the Dynamics of a DNN Learning Symbolic Interactions

This study proves the two-phase dynamics of a deep neural network (DNN) learning interactions. Despite the long disappointing view of the faithfulness of post-hoc explanation of a DNN, a series of theorems have been proven in recent years to show that for a given input sample, a small set of interactions between input variables can be considered as primitive inference patterns that faithfully represent a DNN's detailed inference logic on that sample. Particularly, Zhang et al. have observed that various DNNs all learn interactions of different complexities in two distinct phases, and this two-phase dynamics well explains how a DNN changes from under-fitting to over-fitting. Therefore, in this study, we mathematically prove the two-phase dynamics of interactions, providing a theoretical mechanism for how the generalization power of a DNN changes during the training process. Experiments show that our theory well predicts the real dynamics of interactions on different DNNs trained for various tasks.

cs.LG↗

arXiv2024

Optimized Cryo-CMOS Technology with VTH<0.2V and Ion>1.2mA/um for High-Peformance Computing

We report the design-technology co-optimization (DTCO) scheme to develop a 28-nm cryogenic CMOS (Cryo-CMOS) technology for high-performance computing (HPC). The precise adjustment of halo implants manages to compensate the threshold voltage (VTH) shift at low temperatures. The optimized NMOS and PMOS transistors, featured by VTH<0.2V, sub-threshold swing (SS)<30 mV/dec, and on-state current (Ion)>1.2mA/um at 77K, warrant a reliable sub-0.6V operation. Moreover, the enhanced driving strength of Cryo-CMOS inherited from a higher transconductance leads to marked improvements in elevating the ring oscillator frequency by 20%, while reducing the power consumption of the compute-intensive cryogenic IC system by 37% at 77K.

eess.SY↗

arXiv2024

DIIT: A Domain-Invariant Information Transfer Method for Industrial Cross-Domain Recommendation

Cross-Domain Recommendation (CDR) have received widespread attention due to their ability to utilize rich information across domains. However, most existing CDR methods assume an ideal static condition that is not practical in industrial recommendation systems (RS). Therefore, simply applying existing CDR methods in the industrial RS environment may lead to low effectiveness and efficiency. To fill this gap, we propose DIIT, an end-to-end Domain-Invariant Information Transfer method for industrial cross-domain recommendation. Specifically, We first simulate the industrial RS environment that maintains respective models in multiple domains, each of them is trained in the incremental mode. Then, for improving the effectiveness, we design two extractors to fully extract domain-invariant information from the latest source domain models at the domain level and the representation level respectively. Finally, for improving the efficiency, we design a migrator to transfer the extracted information to the latest target domain model, which only need the target domain model for inference. Experiments conducted on one production dataset and two public datasets verify the effectiveness and efficiency of DIIT.

cs.IR↗