arXiv subjects

Yixi Zhou

Publications and source records attributed to Yixi Zhou.

9 recordsLinked to original sources

arXiv2026

MSUE: Multi-Modal Soccer Understanding Expert

This paper presents our solution to the 2026 SoccerNet VQA Challenge. We first develop a cost-effective data synthesis pipeline driven by a Vision-Language Model (VLM), which systematically restructures raw domain data into diverse VQA samples, including concise answers and long-form responses. Second, we propose MSUE, a multi-expert question answering architecture that employs a Large Language Model (LLM) to dynamically dispatch questions to text, image, and video experts. These experts are instantiated as a strong text baseline Gemini3-Flash, a fine-tuned Qwen3-VL, and an external knowledge base, respectively, working collaboratively to enhance VQA performance. MSUE achieves an accuracy of \textbf{0.95} on the challenge benchmark, securing third place in the leaderboard.

cs.CV↗

arXiv2026

SQLStructEval: Structural Evaluation of LLM Text-to-SQL Generation

Despite strong performance on Text-to-SQL benchmarks, it remains unclear whether LLM-generated SQL programs are structurally reliable. In this work, we investigate the structural behavior of LLM-generated SQL queries and introduce SQLStructEval, a framework for analyzing program structures through canonical abstract syntax tree (AST) representations. Our experiments on the Spider benchmark show that modern LLMs often produce structurally diverse queries for the same input, even when execution results are correct, and that such variance is frequently triggered by surface-level input changes such as paraphrases or schema presentation. We further show that generating queries in a structured space via a compile-style pipeline can improve both execution accuracy and structural consistency. These findings suggest that structural reliability is a critical yet overlooked dimension for evaluating LLM-based program generation systems. Our code is available at https://anonymous.4open.science/r/StructEval-2435.

cs.CL↗

arXiv2026

FinReporting: An Agentic Workflow for Localized Reporting of Cross-Jurisdiction Financial Disclosures

Financial reporting systems increasingly leverage Large Language Models (LLMs) to extract and summarize corporate disclosures. However, most existing approaches assume a single-market setting and overlook structural differences across jurisdictions. Variations in accounting taxonomies, tagging infrastructures (e.g., XBRL vs.\ PDF), and aggregation conventions introduce substantial challenges for semantic alignment and reliable verification. Here, we aim to bridge this gap. We present FinReporting, an agentic workflow for localized cross-jurisdiction financial reporting. The system constructs a unified canonical ontology spanning the income statement, balance sheet, and cash flow statement, and decomposes reporting into auditable stages, including filing acquisition, extraction, canonical mapping, and anomaly logging. Rather than treating LLMs as free-form generators, FinReporting employs them as constrained verifiers operating under explicit decision rules with evidence grounding. Evaluated on annual filings from the USA, Japan, and China, FinReporting improves consistency and reliability under heterogeneous reporting regimes. We further release an interactive demo that enables cross-market inspection and supports structured export of localized financial statements. Our demo is available at url{https://huggingface.co/spaces/BoomQ/FinReporting-Demo. A video describing our system is available at https://www.youtube.com/watch?v=f65jdEL31Kk.

cs.CL↗

arXiv2026

Fine-grained Classification of A Million Life Trajectories from Wikipedia

Life trajectories of notable people convey essential messages for human dynamics research. These trajectories consist of (\textit{person, time, location, activity type}) tuples recording when and where a person was born, went to school, started a job, or fought in a war. However, current studies only cover limited activity types such as births and deaths, lacking large-scale fine-grained trajectories. Using a tool that extracts (\textit{person, time, location}) triples from Wikipedia, we formulate the problem of classifying these triples into 24 carefully-defined types using textual context as complementary information. The challenge is that triple entities are often scattered in noisy contexts. We use syntactic graphs to bring triple entities and relevant information closer, fusing them with text embeddings to classify life trajectory activities. Since Wikipedia text quality varies, we use LLMs to refine the text for more standardized syntactic graphs. Our framework achieves 84.5\% accuracy, surpassing baselines. We construct the largest fine-grained life trajectory dataset with 3.8 million labeled activities for 589,193 individuals spanning 3 centuries. In the end, we showcase how these trajectories can support grand narratives of human dynamics across time and space. Code/data are publicly available.

cs.CY↗

arXiv2026

FinCARDS: Card-Based Analyst Reranking for Financial Document Question Answering

Financial question answering (QA) over long corporate filings requires evidence to satisfy strict constraints on entities, financial metrics, fiscal periods, and numeric values. However, existing LLM-based rerankers primarily optimize semantic relevance, leading to unstable rankings and opaque decisions on long documents. We propose FinCards, a structured reranking framework that reframes financial evidence selection as constraint satisfaction under a finance-aware schema. FinCards represents filing chunks and questions using aligned schema fields (entities, metrics, periods, and numeric spans), enabling deterministic field-level matching. Evidence is selected via a multi-stage tournament reranking with stability-aware aggregation, producing auditable decision traces. Across two corporate filing QA benchmarks, FinCards substantially improves early-rank retrieval over both lexical and LLM-based reranking baselines, while reducing ranking variance, without requiring model fine-tuning or unpredictable inference budgets. Our code is available at https://github.com/XanderZhou2022/FINCARDS.

cs.IR↗

arXiv2024

Unveiling the Mechanism of Phonon-Polariton Damping in α-MoO_3

Phonon polaritons (PhPs) (light coupled to lattice vibrations) in the highly anisotropic polar layered material molybdenum trioxide (α-MoO_3) are currently the focus of intense research efforts due to their extreme subwavelength field confinement, directional propagation and unprecedented low losses. Nevertheless, prior research has primarily concentrated on exploiting the squeezing and steering capabilities of α-MoO_3 PhPs, without inquiring much into the dominant microscopic mechanism that determines their long lifetimes, key for their implementation in nanophotonic applications. This study delves into the fundamental processes that govern PhP damping in α-MoO_3 by combining ab initio calculations with scattering-type scanning near-field optical microscopy (s-SNOM) and Fourier-transform infrared (FTIR) spectroscopy measurements across a broad temperature range (from 8 to 300 K). The remarkable agreement between our theoretical predictions and experimental observations allows us to identify third-order anharmonic phonon-phonon scattering as the main damping mechanism of α-MoO_3 PhPs. These findings shed light on the fundamental limits of low-loss PhPs, a crucial factor for assessing their implementation into nanophotonic devices.

cond-mat.mtrl-sci↗

arXiv2024

Highly confined epsilon-near-zero- and surface-phonon polaritons in SrTiO3 membranes

Recent theoretical studies have suggested that transition metal perovskite oxide membranes can enable surface phonon polaritons in the infrared range with low loss and much stronger subwavelength confinement than bulk crystals. Such modes, however, have not been experimentally observed so far. Here, using a combination of far-field Fourier-transform infrared (FTIR) spectroscopy and near-field synchrotron infrared nanospectroscopy (SINS) imaging, we study the phonon-polaritons in a 100 nm thick freestanding crystalline membrane of SrTiO3 transferred on metallic and dielectric substrates. We observe a symmetric-antisymmetric mode splitting giving rise to epsilon-near-zero and Berreman modes as well as highly confined (by a factor of 10) propagating phonon polaritons, both of which result from the deep-subwavelength thickness of the membranes. Theoretical modeling based on the analytical finite-dipole model and numerical finite-difference methods fully corroborate the experimental results. Our work reveals the potential of oxide membranes as a promising platform for infrared photonics and polaritonics.

cond-mat.mtrl-sci↗

arXiv2023

Thermal and electrostatic tuning of surface phonon-polaritons in LaAlO3/SrTiO3 heterostructures

Phonon polaritons are promising for infrared applications due to a strong light-matter coupling and subwavelength energy confinement they offer. Yet, the spectral narrowness of the phonon bands and difficulty to tune the phonon polariton properties hinder further progress in this field. SrTiO3 - a prototype perovskite oxide - has recently attracted attention due to two prominent far-infrared phonon polaritons bands, albeit without any tuning reported so far. Here we show, using cryogenic infrared near-field microscopy, that long-propagating surface phonon polaritons are present both in bare SrTiO3 and in LaAlO3/SrTiO3 heterostructures hosting a two-dimensional electron gas. The presence of the two-dimensional electron gas increases dramatically the thermal variation of the upper limit of the surface phonon polariton band due to temperature dependent polaronic screening of the surface charge carriers. Furthermore, we demonstrate a tunability of the upper surface phonon polariton frequency in LaAlO3/SrTiO3 via electrostatic gating. Our results suggest that oxide interfaces are a new platform bridging unconventional electronics and long-wavelength nanophotonics.

cond-mat.mtrl-sci↗

arXiv2023

Characteristic lengthscales of the electrically-induced insulator-to-metal transition

Some correlated materials display an insulator-to-metal transition as the temperature is increased. In most cases this transition can also be induced electrically, resulting in volatile resistive switching due to the formation of a conducting filament. While this phenomenon has attracted much attention due to potential applications, many fundamental questions remain unaddressed. One of them is its characteristic lengths: what sets the size of these filaments, and how does this impact resistive switching properties. Here we use a combination of wide-field and scattering-type scanning near-field optical microscopies to characterize filament formation in NdNiO3 and SmNiO3 thin films. We find a clear trend: smaller filaments increase the current density, yielding sharper switching and a larger resistive drop. With the aid of numerical simulations, we discuss the parameters controlling the filament width and, hence, the switching properties.

cond-mat.mtrl-sci↗