arXiv subjects

Yuanbin Wu

Publications and source records attributed to Yuanbin Wu.

At least 19 recordsLinked to original sources

arXiv2026

Nuclear excitation by radiative electron-ion recombination

A nuclear excitation mechanism, nuclear excitation by radiative electron-ion recombination (NERER), is put forward theoretically here. NERER is a third-order process that proceeds via a virtual electronic state: an electron recombines into an atomic vacancy of an ion with the simultaneous emission of a real photon and excitation of the nucleus. The photon emission compensates the energy mismatch between the free-bound electronic transition and the nuclear transition energies, thus there is no resonant condition imposed to the incident electron. We develop here the theoretical framework for NERER, and investigate the case of the $8.4$ eV isomeric excitation of $^{229}$Th for the production of the nuclear clock isomer $^{229m}$Th. Our results show that, with the coupling to the inner atomic shells for highly-charged ions, the NERER cross section can exceed the one of the known lower-order process of nuclear excitation by inelastic electron scattering by more than one order of magnitude. Our findings offer a new pathway for nuclear excitation and efficient isomer production, and support further investigations for high-order effects in the interplay between the atomic and nuclear systems.

nucl-th↗

arXiv2026

Nuclear excitation via inelastic scattering of low-energy vortex electrons

Vortex particles carrying orbital angular momenta (OAMs) have found important applications in broad fields. However, the experimental verification of OAM transfer at the nuclear scale remains a great challenge. Here, we put forward a novel method to probe such OAM transfer through nuclear excitation via inelastic scattering of low-energy vortex electrons. We develop a Dirac distorted-wave Born approximation framework that incorporates the incident-electron OAM and a nonperturbative treatment of the Coulomb field, and apply it to $^{229}\mathrm{Th}$. We find that the vortex and non-vortex electrons yield opposite angular distributions, attributed to the OAM-modified selection rule and the Coulomb-induced redistribution of partial-wave strengths, providing an angle-resolved signature. Moreover, the vortex electron exhibits topological protection in the nuclear Coulomb field. Our method offers a route to probing nuclear-scale OAM transfer and deepens our understanding of the topological properties of vortex particles.

nucl-th↗

arXiv2026

Isomer depletion via nuclear excitation by inelastic electron scattering

Isomer depletion via the process of nuclear excitation by inelastic electron scattering is investigated theoretically. A comprehensive study on low-energy nuclear excitations by inelastic electron scattering is performed to analyze the impact of the nuclear and ion charge, the nuclear transition energy, and the nuclear transition multipolarity on the cross section of the process. We apply the analysis to the case of isomer depletion, in which an excitation from the isomeric state to a nuclear level above the isomeric state can lead to decay to a nuclear level below the isomer itself and hence lead to the release of the energy stored in the isomer. For this purpose, the isomer depletion of $\mathrm{{}^{93m}{Mo}}$, $\mathrm{{}^{152m}{Eu}}$, and $\mathrm{{}^{178m}{Hf}}$, which represent the most important scenarios of isomer depletion, are studied. Our results demonstrate the capability of the process of nuclear excitation by inelastic electron scattering for isomer depletion.

nucl-th↗

arXiv2026

Logic-Regularized Verifier Elicits Reasoning from LLMs

Verifiers are crucial components for enhancing modern LLMs' reasoning capability. Typicalverifiers require resource-intensive superviseddataset construction, which is costly and faceslimitations in data diversity. In this paper, wepropose LOVER, an unsupervised verifier regularized by logical rules. LOVER treats theverifier as a binary latent variable, utilizinginternal activations and enforcing three logical constraints on multiple reasoning paths:negation consistency, intra-group consistency,and inter-group consistency (grouped by thefinal answer). By incorporating logical rulesas priors, LOVER can leverage unlabeled examples and is directly compatible with any offthe-shelf LLMs. Experiments on 10 datasetsdemonstrate that LOVER significantly outperforms unsupervised baselines, achieving performance comparable to the supervised verifier(reaching its 95% level on average). The sourcecode is publicly available at https://github.com/wangxinyufighting/llm-lover.

cs.CL↗

arXiv2026

Stabilizing LLM Supervised Fine-Tuning via Explicit Distributional Control

Post-training large language models (LLMs) often suffers from catastrophic forgetting, where improvements on a target objective degrade previously acquired capabilities. Recent evidence suggests that this phenomenon is primarily driven by excessive distributional drift during optimization. Motivated by this perspective, we propose Anchored Learning, a simple framework that explicitly controls distributional updates during offline fine-tuning via a dynamically evolving moving anchor. Instead of matching a fixed reference distribution, the anchor interpolates between the current model and a frozen reference to construct an intermediate target that the model distills toward, transforming global fine-tuning into a sequence of local trust-region updates in distribution space. Theoretically, we prove this anchor-based update admits a linear KL-divergence upper bound per iteration, ensuring a stable transition between model distributions. Extensive experiments on iGSM, MedCalc, and IFEval show that Anchored Learning consistently lies on the Pareto frontier of gain-stability trade-offs, achieving near-optimal performance improvements while substantially reducing degradation compared to strong baselines. For example, while standard SFT suffers from over 53% performance degradation on iGSM and MedCalc, Anchored Learning slashes this drop to under 5% while maintaining near-optimal gains (e.g., 75.2% on iGSM).

cs.LG↗

arXiv2026

Nuclear non-resonant photoexcitation assisted by electron recombination

We investigate theoretically a nuclear excitation mechanism involving absorption of non-resonant photons leveraged by the coupling to the atomic shell. The nuclear non-resonant photoexcitation is assisted by electron recombination which compensates the energy mismatch between photon and nuclear transition energies, reminiscent of parametric up-conversion in non-linear media. This third-order process proceeds via a virtual nuclear state rather than virtual electronic states, distinguishing this mechanism from the electronic bridge. We investigate the process on the example of a so-far not observed 14.2 keV hard x-ray transition in 193Pt driven by an x-ray free-electron laser. Although the calculated cross section is small, it can be compensated by the vast number of non-resonant photons from the x-ray laser pulse. By enabling nuclear excitation through non-resonant photons, this up-conversion-like mechanism suggests new directions for non-linear x-ray interactions mediated by nuclear transitions.

nucl-th↗

arXiv2026

Off-axis vortex scattering of electron-positron annihilation into a photon pair

The off-axis triple-vortex scattering process of $e^-e^+\to\gamma\gamma$ is studied theoretically, in which the positron is in a plane-wave state and the electron and photons are in vortex states. We develop a theoretical formalism for the process, which allows us to study the effects of various vortex parameters and scattering angle. We adopt a Bessel-Gaussian type wave packet for the initial vortex electron for the purpose of normalization. Numerical calculations are performed for an electron and a positron with a moderate energy around $1~\textrm{MeV}$. Our results demonstrate strong impacts of the scattering angle and the topological charges on the cross section and distributions in the energy and cone angles of the vortex photons. This could provide insight into off-axis vortex scattering and also a possible approach to distinguishing and detecting vortex electrons by off-axis vortex scattering.

hep-ph↗

arXiv2025

DVAGen: Dynamic Vocabulary Augmented Generation

Language models trained with a fixed vocabulary struggle to generalize to novel or out-of-vocabulary words, limiting their flexibility in handling diverse token combinations. Existing dynamic vocabulary approaches attempt to address this limitation but face challenges such as fragmented codebases, lack of support for modern LLMs, and limited inference scalability. To overcome these issues, we introduce DVAGen, a fully open-source, unified framework designed for training, evaluation, and visualization of dynamic vocabulary-augmented language models. Our framework modularizes the pipeline for ease of customization, integrates seamlessly with open-source LLMs, and is the first to provide both CLI and WebUI tools for real-time result inspection. We validate the effectiveness of dynamic vocabulary methods on modern LLMs and demonstrate support for batch inference, significantly improving inference throughput.

cs.CL↗

arXiv2025

Protein Design with Dynamic Protein Vocabulary

Protein design is a fundamental challenge in biotechnology, aiming to design novel sequences with specific functions within the vast space of possible proteins. Recent advances in deep generative models have enabled function-based protein design from textual descriptions, yet struggle with structural plausibility. Inspired by classical protein design methods that leverage natural protein structures, we explore whether incorporating fragments from natural proteins can enhance foldability in generative models. Our empirical results show that even random incorporation of fragments improves foldability. Building on this insight, we introduce ProDVa, a novel protein design approach that integrates a text encoder for functional descriptions, a protein language model for designing proteins, and a fragment encoder to dynamically retrieve protein fragments based on textual functional descriptions. Experimental results demonstrate that our approach effectively designs protein sequences that are both functionally aligned and structurally plausible. Compared to state-of-the-art models, ProDVa achieves comparable function alignment using less than 0.04% of the training data, while designing significantly more well-folded proteins, with the proportion of proteins having pLDDT above 70 increasing by 7.38% and those with PAE below 10 increasing by 9.6%.

cs.LG↗

arXiv2025

Towards Economical Inference: Enabling DeepSeek's Multi-Head Latent Attention in Any Transformer-based LLMs

Multi-head Latent Attention (MLA) is an innovative architecture proposed by DeepSeek, designed to ensure efficient and economical inference by significantly compressing the Key-Value (KV) cache into a latent vector. Compared to MLA, standard LLMs employing Multi-Head Attention (MHA) and its variants such as Grouped-Query Attention (GQA) exhibit significant cost disadvantages. Enabling well-trained LLMs (e.g., Llama) to rapidly adapt to MLA without pre-training from scratch is both meaningful and challenging. This paper proposes the first data-efficient fine-tuning method for transitioning from MHA to MLA (MHA2MLA), which includes two key components: for partial-RoPE, we remove RoPE from dimensions of queries and keys that contribute less to the attention scores, for low-rank approximation, we introduce joint SVD approximations based on the pre-trained parameters of keys and values. These carefully designed strategies enable MHA2MLA to recover performance using only a small fraction (0.3% to 0.6%) of the data, significantly reducing inference costs while seamlessly integrating with compression techniques such as KV cache quantization. For example, the KV cache size of Llama2-7B is reduced by 92.19%, with only a 0.5% drop in LongBench performance.

cs.CL↗

arXiv2025

TASO: Task-Aligned Sparse Optimization for Parameter-Efficient Model Adaptation

LoRA has become one of the most widely used parameter-efficient fine-tuning methods due to its simplicity and effectiveness. However, numerous studies have shown that LoRA often introduces substantial parameter redundancy, which not only increases the number of trainable parameters but also hinders the effectiveness of fine-tuning. Since identifying redundant parameters in LoRA is inherently difficult, how to eliminate them efficiently and accurately remains a challenging problem. In this paper, we propose TASO, a redundancy reduction method that leverages importance information from the pretrained model's weights to mitigate LoRA redundancy. Specifically, we estimate parameter importance on downstream tasks and identify task-specific core regions based on the distribution of importance scores. The location information of these core regions is then used to determine the sparse structure of LoRA modules, enabling redundancy removal before fine-tuning. Our approach significantly reduces the number of trainable parameters required for task adaptation, while providing a novel task-aligned perspective for LoRA redundancy reduction. Experimental results demonstrate that, with a parameter budget comparable to LoRA with rank $r = 1$, TASO consistently outperforms standard LoRA across multiple tasks, achieving strong fine-tuning performance while effectively eliminating redundant parameters.

cs.CL↗

arXiv2025

Microscopic nuclear structure study of $^{229}$Th by Projected Shell Model

$^{229}$Th, a crucial candidate for nuclear clocks and many other applications, is a typical heavy nucleus with an extremely low-energy isomeric state $^{229m}$Th. A detailed study of the nuclear structure of $^{229}$Th is performed here by the microscopic model of state-of-the-art projected shell model. Our calculation describes well low-energy levels of $^{229}$Th, and provides a reduced transition probability $B(M1)$ of $0.0240$ W.u. for the isomeric transition which agrees well with the radiative lifetime of $^{229m}$Th measured recently. Our result supports a small multipole mixing for the cross-band transition of the second-excited state of $^{229}$Th, suggesting that further investigations on the inconsistencies in the decay of the second-excited state should be necessary. The physics behind these properties is revealed by the analysis of the nuclear wave functions. Our findings provide a deep insight into $^{229}$Th from the microscopic nuclear structure point of view, and offer the chance for further studies for nuclear clocks and relevant topics by microscopic nuclear structure theory.

nucl-th↗

arXiv2025

The CoNLL-2013 Shared Task on Grammatical Error Correction

The CoNLL-2013 shared task was devoted to grammatical error correction. In this paper, we give the task definition, present the data sets, and describe the evaluation metric and scorer used in the shared task. We also give an overview of the various approaches adopted by the participating teams, and present the evaluation results.

cs.CL↗

arXiv2025

On Support Samples of Next Word Prediction

Language models excel in various tasks by making complex decisions, yet understanding the rationale behind these decisions remains a challenge. This paper investigates \emph{data-centric interpretability} in language models, focusing on the next-word prediction task. Using representer theorem, we identify two types of \emph{support samples}-those that either promote or deter specific predictions. Our findings reveal that being a support sample is an intrinsic property, predictable even before training begins. Additionally, while non-support samples are less influential in direct predictions, they play a critical role in preventing overfitting and shaping generalization and representation learning. Notably, the importance of non-support samples increases in deeper layers, suggesting their significant role in intermediate representation formation. These insights shed light on the interplay between data and model decisions, offering a new dimension to understanding language model behavior and interpretability.

cs.CL↗

arXiv2025

The Role of Visual Modality in Multimodal Mathematical Reasoning: Challenges and Insights

Recent research has increasingly focused on multimodal mathematical reasoning, particularly emphasizing the creation of relevant datasets and benchmarks. Despite this, the role of visual information in reasoning has been underexplored. Our findings show that existing multimodal mathematical models minimally leverage visual information, and model performance remains largely unaffected by changes to or removal of images in the dataset. We attribute this to the dominance of textual information and answer options that inadvertently guide the model to correct answers. To improve evaluation methods, we introduce the HC-M3D dataset, specifically designed to require image reliance for problem-solving and to challenge models with similar, yet distinct, images that change the correct answer. In testing leading models, their failure to detect these subtle visual differences suggests limitations in current visual perception capabilities. Additionally, we observe that the common approach of improving general VQA capabilities by combining various types of image encoders does not contribute to math reasoning performance. This finding also presents a challenge to enhancing visual reliance during math reasoning. Our benchmark and code would be available at \href{https://github.com/Yufang-Liu/visual_modality_role}{https://github.com/Yufang-Liu/visual\_modality\_role}.

cs.CV↗

arXiv2025

PDFBench: A Benchmark for De novo Protein Design from Function

Function-guided protein design is a crucial task with significant applications in drug discovery and enzyme engineering. However, the field lacks a unified and comprehensive evaluation framework. Current models are assessed using inconsistent and limited subsets of metrics, which prevents fair comparison and a clear understanding of the relationships between different evaluation criteria. To address this gap, we introduce PDFBench, the first comprehensive benchmark for function-guided denovo protein design. Our benchmark systematically evaluates eight state-of-the-art models on 16 metrics across two key settings: description-guided design, for which we repurpose the Mol-Instructions dataset, originally lacking quantitative benchmarking, and keyword-guided design, for which we introduce a new test set, SwissTest, created with a strict datetime cutoff to ensure data integrity. By benchmarking across a wide array of metrics and analyzing their correlations, PDFBench enables more reliable model comparisons and provides key insights to guide future research.

cs.LG↗

arXiv2025

Isomer production by multi-photon excitation

The multi-photon excitation to the $8$-eV nuclear isomeric state $^{229\text{m}}$Th in the direct laser-nucleus interaction is investigated theoretically. We solve the time-dependent Schrödinger equation with the method which allows us to study the $n$-photon absorption in the nuclear excitation in the direct laser-nucleus interaction. Based on the laser facilities available currently or in the near future, we analyze the impact of the laser parameters on the excitation probability of the multi-photon excitation. The possibilities of the $2$-, $3$- and $4$-photon excitations to the isomeric state $^{229\text{m}}$Th from the ground state are discussed in details. Our results show the strong impact of the laser intensity and pulse duration on the multi-photon excitation probability. The onset of high-order effects in the multi-photon excitation in the direct laser-nucleus interaction is also revealed. Our findings open new possibilities to study the multi-photon laser-nucleus interaction in high-power laser facilities.

nucl-th↗

arXiv2025

All-vortex nonlinear Compton scattering in a polarized laser field

The process of all-vortex nonlinear Compton scattering in an intense and polarized laser field, in which the initial and final electrons and the emitted $γ$ photon are all in vortex states, is studied theoretically. We develop a formalism for the process, which allows us to study the exchanges of the orbital angular momentum (OAM) and spin angular momentum among the electron, $γ$ photon, and laser. A wave packet of the Bessel-Gaussian type is adopted to describe the initial vortex electron for the purpose of normalization. Both circularly and linearly polarized lasers are examined. Substantial numerical calculations are performed to reveal the physics of the exchanges of the OAM and spin angular momentum. The strong impact of the laser intensity and the opening angle of the initial vortex electron is demonstrated. Our results also suggest possible scenarios for the separation of the emitted $γ$ photons with different OAMs, as well as a possible way which could help to distinguish the OAM and spin of the vortex particle by the multi-peak structure in the spectrum of the emitted photon.

hep-ph↗