arXiv subjects

Yuan Liu

Publications and source records attributed to Yuan Liu.

At least 217 records · Page 12Linked to original sources

arXiv2025

EquiBench: Benchmarking Large Language Models' Reasoning about Program Semantics via Equivalence Checking

As large language models (LLMs) become integral to code-related tasks, a central question emerges: Do LLMs truly understand program semantics? We introduce EquiBench, a new benchmark for evaluating LLMs through equivalence checking, i.e., determining whether two programs produce identical outputs for all possible inputs. Unlike prior code generation benchmarks, this task directly tests a model's ability to reason about program semantics. EquiBench consists of 2400 program pairs across four languages and six categories. These pairs are generated through program analysis, compiler scheduling, and superoptimization, ensuring high-confidence labels, nontrivial difficulty, and full automation. We evaluate 19 state-of-the-art LLMs and find that in the most challenging categories, the best accuracies are 63.8% and 76.2%, only modestly above the 50% random baseline. Further analysis reveals that models often rely on syntactic similarity rather than exhibiting robust reasoning about program semantics, highlighting current limitations. Our code and dataset are publicly available at https://github.com/Anjiang-Wei/equibench

cs.LG↗

arXiv2025

Conceptual Design Report of Super Tau-Charm Facility: The Accelerator

Electron-positron colliders operating in the GeV region of center-of-mass energies or the Tau-Charm energy region, have been proven to enable competitive frontier research, due to its several unique features. With the progress of high energy physics in the last two decades, a new-generation Tau-Charm factory, Super Tau Charm Facility (STCF) has been actively promoting by the particle physics community in China. STCF holds great potential to address fundamental questions such as the essence of color confinement and the matter-antimatter asymmetry in the universe in the next decades. The main design goals of STCF are with a center-of-mass energy ranging from 2 to 7 GeV and a peak luminosity surpassing 5*10^34 cm^-2s^-1 that is optimized at a center-of-mass energy of 4 GeV, which is about 50 times that of the currently operating Tau-Charm factory - BEPCII. The STCF accelerator is composed of two main parts: a double-ring collider with the crab-waist collision scheme and an injector that provides top-up injections for both electron and positron beams. As a typical third-generation electron-positron circular collider, the STCF accelerator faces many challenges in both accelerator physics and technology. In this paper, the conceptual design of the STCF accelerator complex is presented, including the ongoing efforts and plans for technological R&D, as well as the required infrastructure. The STCF project aims to secure support from the Chinese central government for its construction during the 15th Five-Year Plan (2026-2030) in China.

physics.acc-ph↗

arXiv2025

Scatterer Localization Using Multi-Bounce Paths

Indoor sensing is challenging because of the multi-bounce effect, spherical wavefront, and spatial nonstationarity (SNS) of the near-field effect. This paper addresses radio-based environment sensing considering these issues. Specifically, graph theory (GT) is used to model the multi-bounce propagation of the near field. In this manner, indoor reflectors/scatterers are modeled as vertices in a propagation graph, the multi-bounce paths are modeled by the edges linking the vertices. Besides, the coupled multipath parameters in the near field, i.e., range and angles, are denoted directly by the coordinates of vertices. Then, the space-alternating generalized expectation-maximization (SAGE) algorithm is adapted to the proposed Graph theory-based dictionary-aided Multi-bounce SAGE (GM-SAGE), where the searching parameters including range and angle of departure/arrival (AoD/AoA) are transformed to the coordinates of scatterers in the graph. The proposed algorithm is validated through measurement-calibrated ray tracing (RT) in a complex indoor office. The results demonstrate that the proposed GM-SAGE can deal with multi-bounce channels.

eess.SP↗

arXiv2025

Environment Reconstruction in Multi-Bounce Channels with Array Partial Blockage

Extremely-large antenna arrays (ELAA) are important in applications requiring high angular resolution. However, a prominent issue is the spatial non-stationary (SNS) channels due to partial blockage to the ELAA. In this paper, we address the scatterer localization and subsequent environment reconstruction considering partially blocked SNS channels. Specifically, the SNS effects are parametrically modeled through spatial-varying amplitudes with sparsity. Based on the established signal model, the graph-based dictionary-aided multi-bounce space-alternating generalized expectation-maximization (GM-SAGE) algorithm is applied to estimate the channel parameters and the channel sparsity is empirically detected along with amplitude estimation. To validate the proposed approach, we generate multi-bounce paths through ray tracing (RT) simulations, where the SNS channels caused by partial blockage could be configured flexibly. The simulation results demonstrate the robustness of the proposed approach in dealing with the SNS channels.

eess.SP↗

arXiv2025

SPGen: Spherical Projection as Consistent and Flexible Representation for Single Image 3D Shape Generation

Existing single-view 3D generative models typically adopt multiview diffusion priors to reconstruct object surfaces, yet they remain prone to inter-view inconsistencies and are unable to faithfully represent complex internal structure or nontrivial topologies. In particular, we encode geometry information by projecting it onto a bounding sphere and unwrapping it into a compact and structural multi-layer 2D Spherical Projection (SP) representation. Operating solely in the image domain, SPGen offers three key advantages simultaneously: (1) Consistency. The injective SP mapping encodes surface geometry with a single viewpoint which naturally eliminates view inconsistency and ambiguity; (2) Flexibility. Multi-layer SP maps represent nested internal structures and support direct lifting to watertight or open 3D surfaces; (3) Efficiency. The image-domain formulation allows the direct inheritance of powerful 2D diffusion priors and enables efficient finetuning with limited computational resources. Extensive experiments demonstrate that SPGen significantly outperforms existing baselines in geometric quality and computational efficiency.

cs.CV↗

arXiv2025

No Bound Randomness in Quantum Nonlocality

Recently it has been found that there exist maximally nonlocal quantum correlations that fail to certify randomness for any fixed input pair, rendering them useless for device-independent spot-checking randomness expansion schemes. Here we show that conversely, in DI randomness amplification protocols where all input pairs are used for randomness generation, any amount of quantum nonlocality is sufficient to certify randomness. This shows that no bound randomness exists in quantum nonlocality - any quantum nonlocal behavior is useful in a DI randomness generation task with appropriate modification of protocol structure. Secondly, we show that in contrast to the hitherto considered fixed-input guessing probability, the average guessing probability over all inputs is a faithful and monotonic measure of nonlocality. We use the average guessing probability to show that in contrast to findings in PRL 134, 090201, the detection efficiency threshold for randomness generation is never lower than that for nonlocality detection. Finally, we analytically compute the average guessing probability by a quantum adversary of a single player's measurement outputs in a standard CHSH Bell test, and use it to demonstrate an improvement in the generation rate in state-of-art amplification protocols.

quant-ph↗

arXiv2025

POINTS-Reader: Distillation-Free Adaptation of Vision-Language Models for Document Conversion

High-quality labeled data is essential for training accurate document conversion models, particularly in domains with complex formats such as tables, formulas, and multi-column text. However, manual annotation is both costly and time-consuming, while automatic labeling using existing models often lacks accuracy in handling such challenging scenarios. Consequently, training student models by distilling outputs from teacher models can significantly limit their performance in real-world applications. In this paper, we propose a fully automated, distillation-free framework comprising two stages for constructing high-quality document extraction datasets and models capable of handling diverse document formats and layouts. In the first stage, we introduce a method for generating large-scale, diverse synthetic data, which enables a model to extract key elements in a unified format with strong initial performance. In the second stage, we present a self-improvement approach that further adapts the model, initially trained on synthetic data, to real-world documents. Specifically, we first use the fine-tuned model to annotate real documents, then apply a suite of filtering strategies to verify annotation quality, and finally retrain the model on the verified dataset. By iteratively repeating this process, we progressively enhance both the model's conversion capabilities and the quality of the generated data. We train a public POINTS-1.5 model to obtain POINTS-Reader, which surpasses many existing public and proprietary models of comparable or larger size. Our model is available at https://github.com/Tencent/POINTS-Reader.

cs.CV↗

arXiv2025

NeuralGS: Bridging Neural Fields and 3D Gaussian Splatting for Compact 3D Representations

3D Gaussian Splatting (3DGS) achieves impressive quality and rendering speed, but with millions of 3D Gaussians and significant storage and transmission costs. In this paper, we aim to develop a simple yet effective method called NeuralGS that compresses the original 3DGS into a compact representation. Our observation is that neural fields like NeRF can represent complex 3D scenes with Multi-Layer Perceptron (MLP) neural networks using only a few megabytes. Thus, NeuralGS effectively adopts the neural field representation to encode the attributes of 3D Gaussians with MLPs, only requiring a small storage size even for a large-scale scene. To achieve this, we adopt a clustering strategy and fit the Gaussians within each cluster using different tiny MLPs, based on importance scores of Gaussians as fitting weights. We experiment on multiple datasets, achieving a 91-times average model size reduction without harming the visual quality.

cs.CV↗

arXiv2025

Hybrid Oscillator-Qubit Quantum Processors: Instruction Set Architectures, Abstract Machine Models, and Applications

Quantum computing with discrete variable (DV, qubit) hardware is approaching the large scales necessary for computations beyond the reach of classical computers. However, important use cases such as quantum simulations of physical models containing bosonic modes, and quantum error correction are challenging for DV-only systems. Separately, hardware containing native continuous-variable (CV, oscillator) systems has received attention as an alternative approach, yet the universal control of such systems is non-trivial. In this work, we show that hybrid CV-DV hardware offers a great advantage in meeting these challenges, offering a powerful computational paradigm that inherits the strengths of both DV and CV processors. We provide a pedagogical introduction to CV-DV systems and the multiple abstraction layers needed to produce a full software stack connecting applications to hardware. We present a variety of new hybrid CV-DV compilation techniques, algorithms, and applications, including the extension of quantum signal processing concepts to CV-DV systems and strategies to simulate systems of interacting spins, fermions, and bosons. To facilitate the development of hybrid CV-DV processor systems, we introduce formal Abstract Machine Models and Instruction Set Architectures -- essential abstractions that enable developers to formulate applications, compile algorithms, and explore the potential of current and future hardware for realizing fault-tolerant circuits, modules, and processors. Hybrid CV-DV quantum computations are beginning to be performed in superconducting, trapped ion, and neutral atom platforms, and large-scale experiments are set to be demonstrated in the near future. We present a timely and comprehensive guide to this relatively unexplored yet promising approach to quantum computation and providing an architectural backbone to guide future development.

quant-ph↗

arXiv2025

Parallel Quantum Signal Processing Via Polynomial Factorization

Quantum signal processing (QSP) is a methodology for constructing polynomial transformations of a linear operator encoded in a unitary. Applied to an encoding of a state $ρ$, QSP enables the evaluation of nonlinear functions of the form $\text{tr}(P(ρ))$ for a polynomial $P(x)$, which encompasses relevant properties like entropies and fidelity. However, QSP is a sequential algorithm: implementing a degree-$d$ polynomial necessitates $d$ queries to the encoding, equating to a query depth $d$. Here, we reduce the depth of these property estimation algorithms by developing Parallel Quantum Signal Processing. Our algorithm parallelizes the computation of $\text{tr} (P(ρ))$ over $k$ systems and reduces the query depth to $d/k$, thus enabling a family of time-space tradeoffs for QSP. This furnishes a property estimation algorithm suitable for distributed quantum computers, and is realized at the expense of increasing the number of measurements by a factor $O( \text{poly}(d) 2^{O(k)} )$. We achieve this result by factorizing $P(x)$ into a product of $k$ smaller polynomials of degree $O(d/k)$, which are each implemented in parallel with QSP, and subsequently multiplied together with a swap test to reconstruct $P(x)$. We characterize the achievable class of polynomials by appealing to the fundamental theorem of algebra, and demonstrate application to canonical problems including entropy estimation and partition function evaluation.

quant-ph↗

arXiv2025

Spectral Hardening Reveals Afterglow Emergence in Long-Duration Fast X-ray Transients: A Case Study of GRB 250404A/EP250404a

The prompt emission and afterglow phases of gamma-ray bursts (GRBs) have been extensively studied, yet the transition between these two phases remains inadequately characterized due to limited multiwavelength observational coverage. Among the recent growing samples of fast X-ray transients observed by Einstein Probe (EP), a subgroup of GRBs are captured with long-duration X-ray emission, potentially containing featured evolution from prompt emission to the afterglow phase. In this Letter, we present a detailed analysis of GRB 250404A/EP250404a, a bright fast X-ray transient detected simultaneously by EP and the Fermi Gamma-ray Burst Monitor in X-rays and gamma rays. Its continuous X-ray emission reveals a long-duration tail, accompanied by distinct spectral evolution manifested by the spectral index $α_{\rm X}$ with an initial softening, followed by an evident hardening, eventually reaching a plateau at the value of $\sim$ -2. Early optical and near-infrared observations enable broadband modeling with forward- and reverse-shock components, confirming that the X-ray hardening signals the emergence of the external-shock afterglow. From this spectral hardening we infer that the prompt phase in soft X-rays lasted $\sim300\;\mathrm{s}$, which is more than 3 times longer than the gamma-ray $T_{90}$. This well-tracked soft-hard-flat spectral pattern provides a clear indication of afterglow emergence from the fading prompt emission and offers a practical criterion for identifying a distinct population of GRBs among fast X-ray transients, even when the detection of the gamma-ray counterpart or obvious temporal break is absent.

astro-ph.HE↗

arXiv2025

Laser Resonance Ionization Spectroscopy of Thorium

High-lying Rydberg and autoionizing (AI) states of thorium (Th) have been studied via resonance laser ionization spectroscopy at both TRIUMF Canada's particle accelerator centre and Oak Ridge National Lab (ORNL). Multiple Rydberg series converging to the ionization potential (IP) were observed via different stepwise laser excitation schemes and were assigned to be the $6d^27s (^4F_{3/2})$ $np$, $nd$, and $nf$ series. Analysis of these series enabled the determination of the IP to be 50868.735(54) cm$^{-1}$, which improved the precision by two orders of magnitude over the current adopted NIST value of 50867(2) cm$^{-1}$. Additionally, four AI Rydberg series were identified and assigned to $nf$ and $nd$ series converging to the $6d^27s$ $^4F_{5/2}$ and $6d^27s$ $^2D_{3/2}$ metastable states of Th$^+$. The measured energies of the Rydberg and AI Rydberg states are reported, and observed perturbations within the series are discussed.

physics.atom-ph↗

arXiv2025

Optimal and Feasible Contextuality-based Randomness Generation

Semi-device-independent (SDI) randomness generation protocols based on Kochen-Specker contextuality offer the attractive features of compact devices, high rates, and ease of experimental implementation over fully device-independent (DI) protocols. Here, we investigate this paradigm and derive four results to improve the state-of-art. Firstly, we introduce a family of simple, experimentally feasible orthogonality graphs (measurement compatibility structures) for which the maximum violation of the corresponding non-contextuality inequalities allows to certify the maximum amount of $\log_2 d$ bits of randomness from a qu$d$it system with projective measurements for $d \geq 3$. We analytically derive the Lovász theta and fractional packing number for this graph family, and thereby prove their utility for optimal randomness generation in both randomness expansion and amplification tasks. Secondly, a central additional assumption in contextuality-based protocols over fully DI ones, is that the measurements are repeatable and satisfy an intended compatibility structure. We frame a relaxation of this condition in terms of $ε$-orthogonality graphs for a parameter $ε> 0$, and derive quantum correlations that allow to certify randomness for arbitrary relaxation $ε\in [0,1)$. Thirdly, it is well known that a single qubit is non-contextual, i.e., the qubit correlations can be explained by a non-contextual hidden variable (NCHV) model. We show however that a single qubit is \textit{almost} contextual, in that there exist qubit correlations that cannot be explained by $ε$-faithful NCHV models for small $ε> 0$. Finally, we point out possible attacks by quantum and general consistent (non-signalling) adversaries for certain classes of contextuality tests over and above those considered in DI scenarios.

quant-ph↗

arXiv2025

Reconstructing 4D Spatial Intelligence: A Survey

Reconstructing 4D spatial intelligence from visual observations has long been a central yet challenging task in computer vision, with broad real-world applications. These range from entertainment domains like movies, where the focus is often on reconstructing fundamental visual elements, to embodied AI, which emphasizes interaction modeling and physical realism. Fueled by rapid advances in 3D representations and deep learning architectures, the field has evolved quickly, outpacing the scope of previous surveys. Additionally, existing surveys rarely offer a comprehensive analysis of the hierarchical structure of 4D scene reconstruction. To address this gap, we present a new perspective that organizes existing methods into five progressive levels of 4D spatial intelligence: (1) Level 1 -- reconstruction of low-level 3D attributes (e.g., depth, pose, and point maps); (2) Level 2 -- reconstruction of 3D scene components (e.g., objects, humans, structures); (3) Level 3 -- reconstruction of 4D dynamic scenes; (4) Level 4 -- modeling of interactions among scene components; and (5) Level 5 -- incorporation of physical laws and constraints. We conclude the survey by discussing the key challenges at each level and highlighting promising directions for advancing toward even richer levels of 4D spatial intelligence. To track ongoing developments, we maintain an up-to-date project page: https://github.com/yukangcao/Awesome-4D-Spatial-Intelligence.

cs.CV↗

arXiv2025

The imaginary case of the nonabelian Cohen--Lenstra heuristics

For a finite group $Γ$, we study the distribution of the Galois group $G_{\emptyset}^{\#}(K)$ of the maximal unramified extension of $K$ that is split completely at $\infty$ and has degree prime to $|Γ|$ and $\textit{Char}(K)$, as $K$ varies over imaginary $Γ$-extensions of $\mathbb{Q}$ or $\mathbb{F}_q(t)$. In the function field case, we compute the moments of the distribution of $G_{\emptyset}^{\#}(K)$ by counting points on Hurwitz stacks. In order to understand the probability of the distribution, we prove that $G_{\emptyset}^{\#}(K)$ admits presentations of a specific form, then use this presentation to build random groups to simulate the behavior of $G_{\emptyset}^{\#}(K)$, and make the conjecture to predict the distribution using the probability measures of these random groups. Our results provide the imaginary analog of the work of Wood, Zureick-Brown, and the first author on the nonabelian Cohen--Lenstra heuristics.

math.NT↗

arXiv2025

PDT: Point Distribution Transformation with Diffusion Models

Point-based representations have consistently played a vital role in geometric data structures. Most point cloud learning and processing methods typically leverage the unordered and unconstrained nature to represent the underlying geometry of 3D shapes. However, how to extract meaningful structural information from unstructured point cloud distributions and transform them into semantically meaningful point distributions remains an under-explored problem. We present PDT, a novel framework for point distribution transformation with diffusion models. Given a set of input points, PDT learns to transform the point set from its original geometric distribution into a target distribution that is semantically meaningful. Our method utilizes diffusion models with novel architecture and learning strategy, which effectively correlates the source and the target distribution through a denoising process. Through extensive experiments, we show that our method successfully transforms input point clouds into various forms of structured outputs - ranging from surface-aligned keypoints, and inner sparse joints to continuous feature lines. The results showcase our framework's ability to capture both geometric and semantic features, offering a powerful tool for various 3D geometry processing tasks where structured point distributions are desired. Code will be available at this link: https://github.com/shanemankiw/PDT.

cs.CV↗

arXiv2025

MEGA-Bench: Scaling Multimodal Evaluation to over 500 Real-World Tasks

We present MEGA-Bench, an evaluation suite that scales multimodal evaluation to over 500 real-world tasks, to address the highly heterogeneous daily use cases of end users. Our objective is to optimize for a set of high-quality data samples that cover a highly diverse and rich set of multimodal tasks, while enabling cost-effective and accurate model evaluation. In particular, we collected 505 realistic tasks encompassing over 8,000 samples from 16 expert annotators to extensively cover the multimodal task space. Instead of unifying these problems into standard multi-choice questions (like MMMU, MMBench, and MMT-Bench), we embrace a wide range of output formats like numbers, phrases, code, \LaTeX, coordinates, JSON, free-form, etc. To accommodate these formats, we developed over 40 metrics to evaluate these tasks. Unlike existing benchmarks, MEGA-Bench offers a fine-grained capability report across multiple dimensions (e.g., application, input type, output format, skill), allowing users to interact with and visualize model capabilities in depth. We evaluate a wide variety of frontier vision-language models on MEGA-Bench to understand their capabilities across these dimensions.

cs.CV↗

arXiv2025

The stability of bi-polarization on dynamical directed graphs: an emergent game perspective

This paper proposes a co-evolutionary model of directed graphs and three opinions, i.e., conservative$(+)$, neutral$(\odot)$ and liberal$(-)$. Agents update both opinions and social relationships with bias. We find that an emergent game suffices to predict the stability of bi-polarization under a rare opinion updating limit and a large system size limit. The bi-polarization is stable if and only if the emergent game has an internal Nash equilibrium. The necessary and sufficient condition is explained by both risk dominance and evolutionary stability. This game approach facilitates us to reveal the stability of bi-polarization in empirical systems. Our work fosters the understanding of opinion formation for controversial topics, and shows a deep connection between opinion dynamics and evolutionary game theory.

nlin.AO↗