arXiv subjects

Xin Xu

Publications and source records attributed to Xin Xu.

At least 163 records · Page 9Linked to original sources

arXiv2025

Skill Expansion and Composition in Parameter Space

Humans excel at reusing prior knowledge to address new challenges and developing skills while solving problems. This paradigm becomes increasingly popular in the development of autonomous agents, as it develops systems that can self-evolve in response to new challenges like human beings. However, previous methods suffer from limited training efficiency when expanding new skills and fail to fully leverage prior knowledge to facilitate new task learning. In this paper, we propose Parametric Skill Expansion and Composition (PSEC), a new framework designed to iteratively evolve the agents' capabilities and efficiently address new challenges by maintaining a manageable skill library. This library can progressively integrate skill primitives as plug-and-play Low-Rank Adaptation (LoRA) modules in parameter-efficient finetuning, facilitating efficient and flexible skill expansion. This structure also enables the direct skill compositions in parameter space by merging LoRA modules that encode different skills, leveraging shared information across skills to effectively program new skills. Based on this, we propose a context-aware module to dynamically activate different skills to collaboratively handle new tasks. Empowering diverse applications including multi-objective composition, dynamics shift, and continual policy shift, the results on D4RL, DSRL benchmarks, and the DeepMind Control Suite show that PSEC exhibits superior capacity to leverage prior knowledge to efficiently tackle new challenges, as well as expand its skill libraries to evolve the capabilities. Project website: https://ltlhuuu.github.io/PSEC/.

cs.LG↗

arXiv2025

BiasEdit: Debiasing Stereotyped Language Models via Model Editing

Previous studies have established that language models manifest stereotyped biases. Existing debiasing strategies, such as retraining a model with counterfactual data, representation projection, and prompting often fail to efficiently eliminate bias or directly alter the models' biased internal representations. To address these issues, we propose BiasEdit, an efficient model editing method to remove stereotypical bias from language models through lightweight networks that act as editors to generate parameter updates. BiasEdit employs a debiasing loss guiding editor networks to conduct local edits on partial parameters of a language model for debiasing while preserving the language modeling abilities during editing through a retention loss. Experiments on StereoSet and Crows-Pairs demonstrate the effectiveness, efficiency, and robustness of BiasEdit in eliminating bias compared to tangental debiasing baselines and little to no impact on the language models' general capabilities. In addition, we conduct bias tracing to probe bias in various modules and explore bias editing impacts on different components of language models.

cs.CL↗

arXiv2025

An adaptive ADMM with regularized spectral penalty for sparse portfolio selection

The mean-variance (MV) model is the core of modern portfolio theory. Nevertheless, it suffers from the over-fitting problem due to the estimation errors of model parameters. We consider the $\ell_{1}$ regularized MV model, which adds an $\ell_{1}$ regularization term in the objective to prevent over-fitting and promote sparsity of solutions. By investigating the relationship between sample size and over-fitting, we propose an initial regularization parameter scheme in the $\ell_{1}$ regularized MV model. Then we propose an adaptive parameter tuning strategy to control the amount of short sales. ADMM is a well established algorithm whose performance is affected by the penalty parameter. In this paper, a penalty parameter scheme based on regularized Barzilai-Borwein step size is proposed, and the modified ADMM is used to solve the $\ell_{1}$ regularized MV problem. Numerical results verify the effectiveness of the two types of parameters proposed in this paper.

math.OC↗

arXiv2025

Discovery of 15 new pulsars at high Galactic Latitudes with FAST

We present the discovery and timing results of 15 pulsars discovered in a high Galactic latitude survey conducted with the Five-hundred-meter Aperture Spherical Telescope (FAST). The survey targeted a region as close as possible to the Galactic Center, encompassing an area near the Galactic Bulge. The newly discovered pulsars consist of eleven normal pulsars and four millisecond pulsars (MSPs). Among the MSPs, three are identified in binary systems with orbital periods of ~3.1, 4.6 and 12.5 days, respectively. We have successfully obtained coherent timing solutions for three of the normal pulsars (PSRs J1745-0059, J1746-0156 and J1800-0059). Furthermore, within our data set we found that four pulsars (three new and one known) show mode-changing and/or subpulse drifting phenomena. Comparing our discoveries with simulations of the Galactic disk and Bulge MSP populations indicates that these new pulsars are most likely located in the disk. Nonetheless, our discoveries demonstrate that deep surveys at high Galactic latitudes have significant potential to enhance our understanding of the MSP population in the direction of the Bulge.

astro-ph.HE↗

arXiv2025

Can LLMs Solve longer Math Word Problems Better?

Math Word Problems (MWPs) play a vital role in assessing the capabilities of Large Language Models (LLMs), yet current research primarily focuses on questions with concise contexts. The impact of longer contexts on mathematical reasoning remains under-explored. This study pioneers the investigation of Context Length Generalizability (CoLeG), which refers to the ability of LLMs to solve MWPs with extended narratives. We introduce Extended Grade-School Math (E-GSM), a collection of MWPs featuring lengthy narratives, and propose two novel metrics to evaluate the efficacy and resilience of LLMs in tackling these problems. Our analysis of existing zero-shot prompting techniques with proprietary LLMs along with open-source LLMs reveals a general deficiency in CoLeG. To alleviate these issues, we propose tailored approaches for different categories of LLMs. For proprietary LLMs, we introduce a new instructional prompt designed to mitigate the impact of long contexts. For open-source LLMs, we develop a novel auxiliary task for fine-tuning to enhance CoLeG. Our comprehensive results demonstrate the effectiveness of our proposed methods, showing improved performance on E-GSM. Additionally, we conduct an in-depth analysis to differentiate the effects of semantic understanding and reasoning efficacy, showing that our methods improves the latter. We also establish the generalizability of our methods across several other MWP benchmarks. Our findings highlight the limitations of current LLMs and offer practical solutions correspondingly, paving the way for further exploration of model generalizability and training methodologies.

cs.CL↗

arXiv2025

UGMathBench: A Diverse and Dynamic Benchmark for Undergraduate-Level Mathematical Reasoning with Large Language Models

Large Language Models (LLMs) have made significant strides in mathematical reasoning, underscoring the need for a comprehensive and fair evaluation of their capabilities. However, existing benchmarks often fall short, either lacking extensive coverage of undergraduate-level mathematical problems or probably suffering from test-set contamination. To address these issues, we introduce UGMathBench, a diverse and dynamic benchmark specifically designed for evaluating undergraduate-level mathematical reasoning with LLMs. UGMathBench comprises 5,062 problems across 16 subjects and 111 topics, featuring 10 distinct answer types. Each problem includes three randomized versions, with additional versions planned for release as leading open-source LLMs become saturated in UGMathBench. Furthermore, we propose two key metrics: effective accuracy (EAcc), which measures the percentage of correctly solved problems across all three versions, and reasoning gap ($Δ$), which assesses reasoning robustness by calculating the difference between the average accuracy across all versions and EAcc. Our extensive evaluation of 23 leading LLMs reveals that the highest EAcc achieved is 56.3\% by OpenAI-o1-mini, with large $Δ$ values observed across different models. This highlights the need for future research aimed at developing "large reasoning models" with high EAcc and $Δ= 0$. We anticipate that the release of UGMathBench, along with its detailed evaluation codes, will serve as a valuable resource to advance the development of LLMs in solving mathematical problems. Codes and data are available at https://github.com/YangLabHKUST/UGMathBench

cs.CL↗

arXiv2025

Star Proper Motions Based on Two-epoch Observations from the SDSS and DESI Imaging Surveys

In this study, we present the construction of a new proper motion catalog utilizing the photometric data from the Sloan Digital Sky Survey (SDSS) and Dark Energy Spectroscopic Instrument (DESI) imaging surveys, with a median time baseline of about 13 years. To mitigate systematic errors, the DESI galaxy positions are employed to establish a reference frame and to correct the position-, magnitude-, and color-dependent discrepancies between SDSS and DESI imaging datasets. Spanning 12,589 square degrees, the catalog encompasses about 223.7 million non-Gaia objects down to $m_r \sim$ 23. Based on 734k quasars, the assessment of the global systematic errors in the DESI-SDSS proper motion catalog yields values of 0.06 mas yr$^{-1}$ for $μ_{α_{*}}$ and 0.12 mas yr$^{-1}$ for $μ_δ$. The catalog exhibits precision surpassing 3.4 mas yr$^{-1}$, albeit varying with position, color, and magnitude. An additional evaluation employing approximately 2,644 distant star samples yields an overall precision of approximately 2.5 and 2.9 mas yr$^{-1}$ for $μ_{α_{*}}$ and $μ_δ$, respectively. Further comparisons with proper motions from SDSS Stripe 82 reveal a strong consistency between the two datasets. As a practical application, we utilize fainter non-Gaia objects in our catalog to update the proper motions of 15 star clusters. The resulting proper motions for these clusters exhibit excellent consistency with those derived from Gaia data. Our proper motion measurements, characterized by a deeper limiting magnitude, stand as a valuable complement to the Gaia dataset.

astro-ph.GA↗

arXiv2025

Reconstructing Pristine Molecular Orbitals from Scanning Tunneling Microscopy Images via Artificial Intelligence Approaches

Molecular orbital (MO) is one of the most fundamental concepts for molecules, relating to all branches of chemistry, while scanning tunneling microscopy (STM) has been widely recognized for its potential to measure the spatial distribution of MOs. However, the precise characterization of MO with high resolution in real space is a long-standing challenge owing to the inevitable interference of high-angular-momentum contributions from functionalized tips in STM. Here, leveraging advances in artificial intelligence for image recognition, we establish a physics-driven deep-learning network, named STM-Net, to reconstruct MOs from high-resolution STM images with a functionalized tip, taking advantage of the separable characteristics of different angular momentum contributions. We demonstrate that STM-Net can be directly applied to a variety of experimental observations, successfully reconstructing pristine MO features for molecules under diverse conditions. Moreover, STM-Net can adapt to various states of the functionalized tip and the substrate, illustrating the broad applicability of our physics-driven framework. These results pave the way for accurate characterization of MO with high resolution, potentially leading to new insights and applications for this fundamental concept in chemistry.

physics.chem-ph↗

arXiv2025

Healing of the edge magnetic island in the island divertor configuration on J-TEXT

The phenomena of island healing and configuration transition induced by high-power electron cyclotron resonance heating (ECRH) have been investigated in the island divertor configuration on the J-TEXT tokamak. Experimental results reveal that the size of the edge open magnetic island with mode number m/n = 3/1 decreases substantially under specific ECRH conditions. This process, referred to as island healing, occurs when ECRH with a power of 500~600 kW is deposited in the plasma core or when 250 kW of ECRH is deposited at r = 0.5 a, where a is the minor radius. The reduction of the island width makes the island divertor ineffective and transition into the limiter configuration. A model incorporating the influence of ECRH on the scrape-off layer (SOL) thermoelectric current is proposed to explain the observed changes in the edge magnetic topology of the island divertor configuration. These findings suggest that ECRH should be deposited at the plasma core with carefully controlled power to ensure the stable and compatible operation of ECRH and the island divertor configuration in tokamaks. The results can provide insights into achieving robust operation of an island divertor in tokamaks.

physics.plasm-ph↗

arXiv2025

VAGeo: View-specific Attention for Cross-View Object Geo-Localization

Cross-view object geo-localization (CVOGL) aims to locate an object of interest in a captured ground- or drone-view image within the satellite image. However, existing works treat ground-view and drone-view query images equivalently, overlooking their inherent viewpoint discrepancies and the spatial correlation between the query image and the satellite-view reference image. To this end, this paper proposes a novel View-specific Attention Geo-localization method (VAGeo) for accurate CVOGL. Specifically, VAGeo contains two key modules: view-specific positional encoding (VSPE) module and channel-spatial hybrid attention (CSHA) module. In object-level, according to the characteristics of different viewpoints of ground and drone query images, viewpoint-specific positional codings are designed to more accurately identify the click-point object of the query image in the VSPE module. In feature-level, a hybrid attention in the CSHA module is introduced by combining channel attention and spatial attention mechanisms simultaneously for learning discriminative features. Extensive experimental results demonstrate that the proposed VAGeo gains a significant performance improvement, i.e., improving acc@0.25/acc@0.5 on the CVOGL dataset from 45.43%/42.24% to 48.21%/45.22% for ground-view, and from 61.97%/57.66% to 66.19%/61.87% for drone-view.

cs.CV↗

arXiv2025

Event-based Video Person Re-identification via Cross-Modality and Temporal Collaboration

Video-based person re-identification (ReID) has become increasingly important due to its applications in video surveillance applications. By employing events in video-based person ReID, more motion information can be provided between continuous frames to improve recognition accuracy. Previous approaches have assisted by introducing event data into the video person ReID task, but they still cannot avoid the privacy leakage problem caused by RGB images. In order to avoid privacy attacks and to take advantage of the benefits of event data, we consider using only event data. To make full use of the information in the event stream, we propose a Cross-Modality and Temporal Collaboration (CMTC) network for event-based video person ReID. First, we design an event transform network to obtain corresponding auxiliary information from the input of raw events. Additionally, we propose a differential modality collaboration module to balance the roles of events and auxiliaries to achieve complementary effects. Furthermore, we introduce a temporal collaboration module to exploit motion information and appearance cues. Experimental results demonstrate that our method outperforms others in the task of event-based video person ReID.

cs.CV↗

arXiv2024

Deep Searches for Radio Pulsations and Bursts from Four Magnetar and a Magnetar-like pulsar with FAST

We report on radio observations of four magnetars SGR 0501+4516, Swift 1834.9-0846, 1E 1841-045, SGR 1900+14 and a magnetar-like pulsar PSR J1846-0258 with the Five-hundred-meter Aperture Spherical radio Telescope (FAST) at 1250 MHz. Notably, PSR J1846-0258 was observed one month after its 2020 X-ray outburst. The data from these observations were searched for periodic emissions and single pulses. No radio emission was detected for any of our targets. After accounting for the effect of red noise, the non-detections yield stringent upper limits on the radio flux density, with $S_{1250} \leq 16.9\, μ$Jy for the four magnetars and the magnetar-like pulsar, along with constraints on single-pulse flux densities. Our deep radio observations suggest that these magnetars and the magnetar-like pulsar are indeed radio-quiet sources or unfavorably beamed. The resulting flux upper limits, along with previous findings, are discussed, highlighting the significance of further radio observations of radio-quiet magnetars and the high-B magnetar-like pulsar.

astro-ph.HE↗

arXiv2024

Toward Scalable Multirobot Control: Fast Policy Learning in Distributed MPC

Distributed model predictive control (DMPC) is promising in achieving optimal cooperative control in multirobot systems (MRS). However, real-time DMPC implementation relies on numerical optimization tools to periodically calculate local control sequences online. This process is computationally demanding and lacks scalability for large-scale, nonlinear MRS. This article proposes a novel distributed learning-based predictive control (DLPC) framework for scalable multirobot control. Unlike conventional DMPC methods that calculate open-loop control sequences, our approach centers around a computationally fast and efficient distributed policy learning algorithm that generates explicit closed-loop DMPC policies for MRS without using numerical solvers. The policy learning is executed incrementally and forward in time in each prediction interval through an online distributed actor-critic implementation. The control policies are successively updated in a receding-horizon manner, enabling fast and efficient policy learning with the closed-loop stability guarantee. The learned control policies could be deployed online to MRS with varying robot scales, enhancing scalability and transferability for large-scale MRS. Furthermore, we extend our methodology to address the multirobot safe learning challenge through a force field-inspired policy learning approach. We validate our approach's effectiveness, scalability, and efficiency through extensive experiments on cooperative tasks of large-scale wheeled robots and multirotor drones. Our results demonstrate the rapid learning and deployment of DMPC policies for MRS with scales up to 10,000 units.

cs.RO↗

arXiv2024

Mix-Modality Person Re-Identification: A New and Practical Paradigm

Current visible-infrared cross-modality person re-identification research has only focused on exploring the bi-modality mutual retrieval paradigm, and we propose a new and more practical mix-modality retrieval paradigm. Existing Visible-Infrared person re-identification (VI-ReID) methods have achieved some results in the bi-modality mutual retrieval paradigm by learning the correspondence between visible and infrared modalities. However, significant performance degradation occurs due to the modality confusion problem when these methods are applied to the new mix-modality paradigm. Therefore, this paper proposes a Mix-Modality person re-identification (MM-ReID) task, explores the influence of modality mixing ratio on performance, and constructs mix-modality test sets for existing datasets according to the new mix-modality testing paradigm. To solve the modality confusion problem in MM-ReID, we propose a Cross-Identity Discrimination Harmonization Loss (CIDHL) adjusting the distribution of samples in the hyperspherical feature space, pulling the centers of samples with the same identity closer, and pushing away the centers of samples with different identities while aggregating samples with the same modality and the same identity. Furthermore, we propose a Modality Bridge Similarity Optimization Strategy (MBSOS) to optimize the cross-modality similarity between the query and queried samples with the help of the similar bridge sample in the gallery. Extensive experiments demonstrate that compared to the original performance of existing cross-modality methods on MM-ReID, the addition of our CIDHL and MBSOS demonstrates a general improvement.

cs.CV↗

arXiv2024

Photometric and Spectroscopic Investigations of Three Large Amplitude Contact Binaries

We performed photometric and spectroscopic studies of three large amplitude contact binaries, NSVS 2418361, ATLAS J057.1170+31.2384 and NSVS 7377875. The amplitudes of three systems' light curves are more than 0.7 magnitude. We analyzed the light curves using Wilson-Devinney code to yield physical parameters. The photometric solutions suggested that NSVS 7377875 belongs to an A-subtype contact binary, while the others are classified as W-subtype ones. Furthermore, the mass ratio of NSVS 7377875 is higher than 0.72, so it belongs to H-subtype contact binaries. Since their light curves have unequal height at two maxima which is called O'Connell effect, a dark spot on the primary component for each target was required to get a better fit of light curves. The orbital period investigation shows that the period of NSVS 2418361 is increasing, indicating a mass transfer from the less massive component to the more massive one, while the other targets exhibit no long-term variation. Our spectral subtraction analysis of LAMOST spectra revealed excess emissions in the $H_α$ line, indicating chromospheric activity in all the three targets. The Gaia distance was applied to estimate the absolute parameters of the three targets, and we obtained their evolutionary state. The relationships between the energy transfer parameter of 76 H-subtype contact binaries and their bolometric luminosity ratios, as well as their contact degree, were presented. We discovered that H-subtype systems have less efficient energy transfer rate, which is corresponding to the conclusion proposed by Csizmadia \& Klagyivik.

astro-ph.SR↗

arXiv2024

Provably Robust and Secure Steganography in Asymmetric Resource Scenario

To circumvent the unbridled and ever-encroaching surveillance and censorship in cyberspace, steganography has garnered attention for its ability to hide private information in innocent-looking carriers. Current provably secure steganography approaches require a pair of encoder and decoder to hide and extract private messages, both of which must run the same model with the same input to obtain identical distributions. These requirements pose significant challenges to the practical implementation of steganography, including limited access to powerful hardware and the intolerance of any changes to the shared input. To relax the limitation of hardware and solve the challenge of vulnerable shared input, a novel and practically significant scenario with asymmetric resource should be considered, where only the encoder is high-resource and accessible to powerful models while the decoder can only read the steganographic carriers without any other model's input. This paper proposes a novel provably robust and secure steganography framework for the asymmetric resource setting. Specifically, the encoder uses various permutations of distribution to hide secret bits, while the decoder relies on a sampling function to extract the hidden bits by guessing the permutation used. Further, the sampling function only takes the steganographic carrier as input, which makes the decoder independent of model's input and model itself. A comprehensive assessment of applying our framework to generative models substantiates its effectiveness. Our implementation demonstrates robustness when transmitting over binary symmetric channels with errors.

cs.CR↗

arXiv2024

A Comprehensive Study of Knowledge Editing for Large Language Models

Large Language Models (LLMs) have shown extraordinary capabilities in understanding and generating text that closely mirrors human communication. However, a primary limitation lies in the significant computational demands during training, arising from their extensive parameterization. This challenge is further intensified by the dynamic nature of the world, necessitating frequent updates to LLMs to correct outdated information or integrate new knowledge, thereby ensuring their continued relevance. Note that many applications demand continual model adjustments post-training to address deficiencies or undesirable behaviors. There is an increasing interest in efficient, lightweight methods for on-the-fly model modifications. To this end, recent years have seen a burgeoning in the techniques of knowledge editing for LLMs, which aim to efficiently modify LLMs' behaviors within specific domains while preserving overall performance across various inputs. In this paper, we first define the knowledge editing problem and then provide a comprehensive review of cutting-edge approaches. Drawing inspiration from educational and cognitive research theories, we propose a unified categorization criterion that classifies knowledge editing methods into three groups: resorting to external knowledge, merging knowledge into the model, and editing intrinsic knowledge. Furthermore, we introduce a new benchmark, KnowEdit, for a comprehensive empirical evaluation of representative knowledge editing approaches. Additionally, we provide an in-depth analysis of knowledge location, which can give a deeper understanding of the knowledge structures inherent within LLMs. Finally, we discuss several potential applications of knowledge editing, outlining its broad and impactful implications.

cs.CL↗

arXiv2024

Acceleration for Deep Reinforcement Learning using Parallel and Distributed Computing: A Survey

Deep reinforcement learning has led to dramatic breakthroughs in the field of artificial intelligence for the past few years. As the amount of rollout experience data and the size of neural networks for deep reinforcement learning have grown continuously, handling the training process and reducing the time consumption using parallel and distributed computing is becoming an urgent and essential desire. In this paper, we perform a broad and thorough investigation on training acceleration methodologies for deep reinforcement learning based on parallel and distributed computing, providing a comprehensive survey in this field with state-of-the-art methods and pointers to core references. In particular, a taxonomy of literature is provided, along with a discussion of emerging topics and open issues. This incorporates learning system architectures, simulation parallelism, computing parallelism, distributed synchronization mechanisms, and deep evolutionary reinforcement learning. Further, we compare 16 current open-source libraries and platforms with criteria of facilitating rapid development. Finally, we extrapolate future directions that deserve further research.

cs.LG↗