arXiv subjects

Qianyi Wang

Publications and source records attributed to Qianyi Wang.

15 recordsLinked to original sources

arXiv2026

AgentPanel: Toward a New Paradigm for Human--AI Collaboration in Exploring Scientific Questions

Identifying promising scientific ideas remains an important challenge in research practice. Researchers commonly rely on small-group discussions or one-to-one interactions with a single large language model, yet these approaches often expose them to only a limited range of perspectives and directions. We present AgentPanel, a multi-agent forum for human--AI collaboration in scientific exploration. Heterogeneous agents asynchronously discuss scientific questions in a forum-style environment, while researchers can submit questions, browse and organize candidate ideas, engage agents in follow-up interactions, and optionally generate post-hoc summary reports. We evaluate AgentPanel in terms of idea quality, exploration breadth, interaction effectiveness, candidate-selection efficiency, and practical utility. Offline experiments show that AgentPanel outperforms a centralized multi-agent debate baseline. A human study with 20 participants further shows that users value AgentPanel for perspective diversity and exploration support. In experience-based comparisons with commonly used LLM tools, 65\% of participants favored AgentPanel for both breadth of research directions and overall suitability for early-stage exploration. The platform is publicly available at https://agentpanel.cc/.

cs.AI↗

arXiv2026

Scaling the Horizon, Not the Parameters: Reaching Trillion-Parameter Performance with a 35B Agent

We introduce Agents-A1, a 35B Mixture-of-Experts Agentic Model that reaches trillion-parameter-level performance by scaling the agent horizon. We investigate agent-horizon scaling from two perspectives: scaling long-horizon trajectories and scaling heterogeneous agent abilities. To support this goal, we build a long-horizon knowledge-action infrastructure that connects external knowledge, actions, observations, and verifier outcomes, producing agentic trajectories with an average length of 45K tokens. Based on this, we train Agents-A1 with a three-stage recipe. First, we perform full-domain supervised fine-tuning to align the base model with broad agentic behaviors. Second, we train domain-level teacher models to capture specialized expertise in each domain. Third, we propose a multi-teacher domain-routed on-policy distillation with salient vocabulary alignment to improve knowledge transfer efficiency across different domains, unifying six heterogeneous domains into one deployable student model. Agents-A1 achieves strong and broad performance for long-horizon agent benchmarks. Compared with 1T-parameter model such as Kimi-K2.6 and DeepSeek-V4-pro, Agents-A1 achieves leading results on SEAL-0 (56.4), IFBench (80.6), HiPhO (46.4), FrontierScience-Olympiad (79.0), and MolBench-Bind (56.8), and remains highly competitive on SciCode (44.3), HLE (47.6) and BrowseComp (75.5). We hope this work provides the community with a practical path for scaling the horizon using a 35B agent that can reach or match the performance of 1T models on long-horizon tasks.

cs.CL↗

arXiv2026

Self-Ensembling Vision-Language Models for Chart Data Extraction

Charts effectively convey quantitative information, but the underlying data are often locked in image form, hindering reuse and analysis. Manually digitizing charts is time-consuming and error-prone, motivating automatic chart-to-table extraction. Recent approaches use specialized vision-language models (VLMs), yet performance still lags on charts with many datapoints or substantial stylistic variation. We propose a VLM self-ensembling method that repeatedly samples multiple tabular outputs from the same VLM for a fixed chart image and aggregates them at the level of individual table cells. We align candidate tables and take per-cell medians over numerical values to produce a more accurate consensus table. Our method also includes convergence detection to stop sampling once the aggregated table stabilizes, and uncertainty estimation based on dispersion across samples to help users assess extraction reliability. Because existing chart extraction benchmarks contain relatively simple plots with limited room for improvement, we introduce WB-ChartExtract, a new benchmark built from World Bank data with more complex and stylistically diverse charts; on average, its charts contain 7 times more datapoints than those in the ChartQA benchmark. Across both ChartQA and WB-ChartExtract, our approach improves extraction accuracy over single-pass VLM outputs, yielding up to 23% relative improvement on WB-ChartExtract after ensembling. More broadly, our method helps unlock tabular data previously siloed in chart images, enabling downstream analysis and reuse.

cs.CL↗

arXiv2026

Picturing Perceptions: An Open-Source Toolkit to Uncover Bias in Humans and Machines

Bias in human judgment and artificial intelligence systems poses critical challenges across consequential domains like hiring, loans, and criminal justice. However, traditional bias measurement tools face fundamental limitations: they struggle to capture intersectional identities, cannot evaluate AI systems, lack grounding in demographic reality, and remain vulnerable to social desirability effects. We introduce PictoPercept, an open-source toolkit that measures bias through visual forced-choice comparisons grounded in population level benchmarks. Participants view pairs of normed facial photographs and assess who is more likely to have higher earnings, with selections compared against actual U.S. Bureau of Labor Statistics data. We validate PictoPercept with a nationally representative sample of 283 American adults and assess GPT-5, a mainstream generative model, using identical stimuli. Our study reveals three key findings: First, participants dramatically underestimate Asian American earnings despite this group having the highest actual earnings, while overestimating Latino male and White male earnings. Second, ingroup favoritism is not universal as White males show clear ingroup bias, but Asian participants actually underestimate their own group's earnings. Third, GPT-5 exhibits substantially stronger biases than humans, with stark systematic underestimation of all female groups. These findings suggest that PictoPercept enables unified bias assessment across human and AI systems while revealing systematic misperceptions that diverge from demographic reality.

cs.CY↗

arXiv2026

Experimental Efficient Influence Sampling of Quantum Processes

Characterizing quantum processes is essential for unlocking the potential of quantum devices. However, standard quantum process tomography is resource-intensive and becomes infeasible on large-scale systems. Despite alternative approaches have been successfully developed for specific scenarios, they typically rely on multi-qubit gates or extensive prior knowledge, limiting their practicability and scalability. To address these challenges and complement existing approaches, we introduce $\textit{influence sampling}$, an efficient and scalable protocol that quantifies the $\textit{influence}$ of a quantum process on all qubit subsets using only single-qubit test gates, with sample complexity independent of system size. Using a photonic platform, we demonstrate influence sampling to identify high-influence qubits, reduce the full process to a smaller effective process, i.e., a junta approximation, and then learn it. We further confirm scalability by applying the protocol to a 24-qubit system and validate the junta approximation on a two-qubit process. These results establish influence sampling as a critical characterization technique, facilitating process learning and device assessment.

quant-ph↗

arXiv2026

LOONG: Online Time-Optimal Autonomous Flight for MAVs in Cluttered Environments

Autonomous flight of micro air vehicles (MAVs) in unknown, cluttered environments remains challenging for time-critical missions due to conservative maneuvering strategies. This article presents an integrated planning and control framework for high-speed, time-optimal autonomous flight of MAVs in cluttered environments. In each replanning cycle (100 Hz), a time-optimal trajectory under polynomial presentation is generated as a reference, with the time-allocation process accelerated by imitation learning. Subsequently, a time-optimal model predictive contouring control (MPCC) incorporates safe flight corridor (SFC) constraints at variable horizon steps to enable aggressive yet safe maneuvering, while fully exploiting the MAV's dynamics. We validate the proposed framework extensively on a custom-built LiDAR-based MAV platform. Simulation results demonstrate superior aggressiveness compared to the state of the art, while real-world experiments achieve a peak speed of 18 m/s in a cluttered environment and succeed in 10 consecutive trials from diverse start points. The video is available at the following link: https://youtu.be/vexXXhv99oQ.

cs.RO↗

arXiv2025

A lightweight detector for real-time detection of remote sensing images

Remote sensing imagery is widely used across various fields, yet real-time detection remains challenging due to the prevalence of small objects and the need to balance accuracy with efficiency. To address this, we propose DMG-YOLO, a lightweight real-time detector tailored for small object detection in remote sensing images. Specifically, we design a Dual-branch Feature Extraction (DFE) module in the backbone, which partitions feature maps into two parallel branches: one extracts local features via depthwise separable convolutions, and the other captures global context using a vision transformer with a gating mechanism. Additionally, a Multi-scale Feature Fusion (MFF) module with dilated convolutions enhances multi-scale integration while preserving fine details. In the neck, we introduce the Global and Local Aggregate Feature Pyramid Network (GLAFPN) to further boost small object detection through global-local feature fusion. Extensive experiments on the VisDrone2019 and NWPU VHR-10 datasets show that DMG-YOLO achieves competitive performance in terms of mAP, model size, and other key metrics.

cs.CV↗

arXiv2025

Photonic implementation of quantum hidden subgroup database compression

We experimentally demonstrate quantum data compression exploiting hidden subgroup symmetries using a photonic quantum processor. Classical databases containing generalized periodicities-symmetries that are in the worst cases inefficient for known classical algorithms to be detect-can efficiently compressed by quantum hidden subgroup algorithms. We implement a variational quantum autoencoder that autonomously learns both the symmetry type (e.g., $\mathbb{Z}_2 \times \mathbb{Z}_2$ vs. $\mathbb{Z}_4$) and the generalized period from structured data. The system uses single photons encoded in path, polarization, and time-bin degrees of freedom, with electronically controlled waveplates enabling tunable quantum gates. Training via gradient descent successfully identifies the hidden symmetry structure, achieving compression by eliminating redundant database entries. We demonstrate two circuit ansatzes: a parametrized generalized Fourier transform and a less-restricted architecture for Simon's symmetry. Both converge successfully, with the cost function approaching zero as training proceeds. These results provide experimental proof-of-principle that photonic quantum computers can compress classical databases by learning symmetries inaccessible to known efficient classical methods, opening pathways for quantum-enhanced information processing.

quant-ph↗

arXiv2025

PEMUTA: Pedagogically-Enriched Multi-Granular Undergraduate Thesis Assessment

The undergraduate thesis (UGTE) plays an indispensable role in assessing a student's cumulative academic development throughout their college years. Although large language models (LLMs) have advanced education intelligence, they typically focus on holistic assessment with only one single evaluation score, but ignore the intricate nuances across multifaceted criteria, limiting their ability to reflect structural criteria, pedagogical objectives, and diverse academic competencies. Meanwhile, pedagogical theories have long informed manual UGTE evaluation through multi-dimensional assessment of cognitive development, disciplinary thinking, and academic performance, yet remain underutilized in automated settings. Motivated by the research gap, we pioneer PEMUTA, a pedagogically-enriched framework that effectively activates domain-specific knowledge from LLMs for multi-granular UGTE assessment. Guided by Vygotsky's theory and Bloom's Taxonomy, PEMUTA incorporates a hierarchical prompting scheme that evaluates UGTEs across six fine-grained dimensions: Structure, Logic, Originality, Writing, Proficiency, and Rigor (SLOWPR), followed by holistic synthesis. Two in-context learning techniques, \ie, few-shot prompting and role-play prompting, are also incorporated to further enhance alignment with expert judgments without fine-tuning. We curate a dataset of authentic UGTEs with expert-provided SLOWPR-aligned annotations to support multi-granular UGTE assessment. Extensive experiments demonstrate that PEMUTA achieves strong alignment with expert evaluations, and exhibits strong potential for fine-grained, pedagogically-informed UGTE evaluations.

cs.CY↗

arXiv2025

A Geometric Quantum Speed Limit: Theoretical Insights and Photonic Implementation

Quantum mechanics imposes a lower bound on the time required for a quantum system to reach certain given targets. In this paper, from a geometric perspective, we introduce a new quantum speed limit (QSL) based on the Bloch angle and derive the condition for it to saturate. Experimentally, we demonstrate the feasibility of measuring this QSL using a photonic system through direct Bloch angle measurements via a swap test, bypassing the need for comprehensive quantum state tomography. Compared to the existing Bloch-angle-based QSL mentioned in prior work, our QSL requires fewer computational and experimental resources and provides tighter constraints for specific dynamics. Our work underscores the Bloch angle's effectiveness in providing tighter and experimentally accessible QSLs and advances the understanding of quantum dynamics.

quant-ph↗

arXiv2025

Quantum Geometric Tensor for Mixed States Based on the Covariant Derivative

The quantum geometric tensor (QGT) is a fundamental quantity for characterizing the geometric properties of quantum states and plays an essential role in elucidating various physical phenomena. The traditional QGT, defined only for pure states, has limited applicability in realistic scenarios where mixed states are common. To address this limitation, we generalize the definition of the QGT to mixed states using the purification bundle and the covariant derivative. Notably, our proposed definition reduces to the traditional QGT when mixed states approach pure states. In our framework, the real and imaginary parts of this generalized QGT correspond to the Bures metric and the mean gauge curvature, respectively, endowing it with a broad range of potential applications. Additionally, using our proposed mixed-state QGT (MSQGT), we derive the geodesic equation applicable to mixed states. This work establishes a unified framework for the geometric analysis of both pure and mixed states, thereby deepening our understanding of the geometric properties of quantum states.

quant-ph↗

arXiv2024

Resource-efficient Direct Characterization of General Density Matrix

Sequential weak measurements allow the direct extraction of individual density-matrix elements instead of globally reconstructing the whole density matrix, opening a new avenue for the characterization of quantum systems. Nevertheless, the requirement of multiple coupling for each qudit of quantum systems and the lack of appropriate precision evaluation constraint its applicability extension, especially for multi-qudit quantum systems. Here, we propose a resource-efficient scheme (RES) to directly characterize the density matrix of general multi-qudit systems, which not only optimizes the measurements but also establishes a feasible estimation analysis. In this scheme, an efficient observable of quantum system is constructed such that a single meter state coupled to each qudit is sufficient to extract the corresponding density-matrix element. An appropriate model based on the statistical distribution of errors are used to evaluate the precision and feasibility of the scheme. We experimentally apply the RES to the direct characterization of general single-photon qutrit states and two-photon entangled states. The results show that the RES outperforms the sequential schemes in terms of efficiency and precision in both weak- and strong- coupling scenarios. This work sheds new light on the practical characterization of large-scale quantum systems and investigation of their non-classical properties.

quant-ph↗

arXiv2020

On the stability of two-flavor and three-flavor quark stars

Following our recenty proposed self-consistent mean-field approximation approach, we have done some researches on the chiral phase transition of strong interaction matter within the framework of Nambu-Jona-Lasinio (NJL) model. The chiral susceptibility and equation of state (EOS) are computed in this work for both two-flavor and three-flavor quark matter for contrast. The Pauli-Villars scheme, which can preserve gauge invariance, is used in this paper. Moreover, whether the three-flavor quark matter is more stable than the two-flavor quark matter or not in quark stars is discussed in this work. In our model, when the bag constant are the same, the two-flavor quark matter has a higher pressure than the three-flavor quark matter, which is different from what Witten proposed in his pioneering work.

hep-ph↗

arXiv2016

Analysis of Sparse Cutting-planes for Sparse MILPs with Applications to Stochastic MILPs

In this paper, we present an analysis of the strength of sparse cutting-planes for mixed integer linear programs (MILP) with sparse formulations. We examine three kinds of problems: packing problems, covering problems, and more general MILPs with the only assumption that the objective function is non-negative. Given a MILP instance of one of these three types, assume that we decide on the support of cutting-planes to be used and the strongest inequalities on these supports are added to the linear programming relaxation. Call the optimal objective function value of the linear programming relaxation together with these cuts as $z^{cut}$. We present bounds on the ratio of $z^{cut}$ and the optimal objective function value of the MILP that depends only on the sparsity structure of the constraint matrix and the support of sparse cuts selected, that is, these bounds are completely data independent. These results also shed light on the strength of scenario-specific cuts for two stage stochastic MILPs.

math.OC↗

arXiv2014

How Good Are Sparse Cutting-Planes?

Sparse cutting-planes are often the ones used in mixed-integer programing (MIP) solvers, since they help in solving the linear programs encountered during branch-&-bound more efficiently. However, how well can we approximate the integer hull by just using sparse cutting-planes? In order to understand this question better, given a polyope $P$ (e.g. the integer hull of a MIP), let $P^k$ be its best approximation using cuts with at most $k$ non-zero coefficients. We consider $d(P, P^k) = \max_{x \in P^k} \left(min_{y \in P} \| x - y\|\right)$ as a measure of the quality of sparse cuts. In our first result, we present general upper bounds on $d(P, P^k)$ which depend on the number of vertices in the polytope and exhibits three phases as $k$ increases. Our bounds imply that if $P$ has polynomially many vertices, using half sparsity already approximates it very well. Second, we present a lower bound on $d(P, P^k)$ for random polytopes that show that the upper bounds are quite tight. Third, we show that for a class of hard packing IPs, sparse cutting-planes do not approximate the integer hull well, that is $d(P, P^k)$ is large for such instances unless $k$ is very close to $n$. Finally, we show that using sparse cutting-planes in extended formulations is at least as good as using them in the original polyhedron, and give an example where the former is actually much better.

math.OC↗