arXiv subjects

Yi Hu

Publications and source records attributed to Yi Hu.

At least 37 records · Page 2Linked to original sources

arXiv2026

Design, Testing, and Commissioning of the Sun Yat-sen University (SYSU) 80 cm Infrared Telescope

The Sun Yat-sen University (SYSU) 80 cm telescope is a new generation near-infrared (NIR) facility in China dedicated to time-domain astronomy, while also serving as a testbed for emerging NIR cameras. Commissioned in October 2024 at the 4100 m Lenghu site on the Tibetan Plateau in China, the telescope adopts a reflective Cassegrain design with two Nasmyth foci for J and K bands. The J band imaging system, initially equipped with a 640 x 512 off-the-shelf InGaAs camera (INS Mars640) and upgraded in June 2025 to a 1280 x 1024 science-grade, deeply cooled camera (YNAOIR), achieves background-limited performance with a dark current of ~ 14 e-/s/pix and a readout noise of ~ 11 e-. The system reaches a limiting magnitude of J ~ 17 mag (Vega system) in single 20 s exposures and depths of J ~ 19.4 mag with stacked 30 minute exposures. For a variable with J ~ 14 mag during on-sky tests, the system delivers millimagnitude-level photometric precision. Since commissioning, the telescope observed transients such as gamma-ray bursts (GRBs), supernovae and comets, variables including active galactic nuclei (AGNs), high-redshift quasars (z > 6), and brown dwarfs, as well as deep-field imaging reaching J ~ 20.5 mag. This validates the feasibility of using InGaAs cameras for astronomical observations, encouraging other institutions to develop dedicated infrared telescopes or integrate infrared cameras into existing optical telescopes.

astro-ph.IM↗

arXiv2026

Position: How can Graphs Help Large Language Models?

With the rapid advancement of large language models (LLMs), classic graph learning tasks have greatly benefited from LLMs, including improved encoding of textual features, more efficient construction of graphs from text, and enhanced reasoning over knowledge graphs. In this paper, we ask a complementary question: How can graphs help LLMs? We address this question from three perspectives: 1) graphs provide an up-to-date knowledge source that helps reduce LLM hallucinations, 2) graph-based prompting techniques-such as Chain-of-Thought (CoT), Tree-of-Thought (ToT), and Graph-of-Thought (GoT)-enhance LLM reasoning capabilities, and 3) integrating graphs into LLMs improves their understanding of structured data, expanding their applicability to domains such as e-commerce, code, and relational databases (RDBs). We further outlook some future directions including designing sparse LLM architectures based on graphs and brain-inspired memory systems.

cs.AI↗

arXiv2026

SoccerRef-Agents: Multi-Agent System for Automated Soccer Refereeing

Refereeing is vital in sports, where fair, accurate, and explainable decisions are fundamental. While intelligent assistant technologies are being widely adopted in soccer refereeing, current AI-assisted approaches remain preliminary. Existing research mostly focuses on isolated video perception tasks and lacks the ability to understand and reason about foul scenarios. To fill this gap, we propose SoccerRef-Agents, a holistic and explainable multi-agent decision-making framework for soccer refereeing. The main contributions are: (i) constructing the multimodal benchmark SoccerRefBench with over 1,200 referee theory questions and 600 foul video clips; (ii) building a vector-based knowledge base RefKnowledgeDB using the latest "Laws of the Game" and a classic case database for precise, knowledge-driven reasoning; (iii) designing a novel multi-agent architecture that collaborates via cross-modal RAG to bridge the semantic gap between visual content and regulatory texts. This work explores the technical capability of integrating MLLMs with refereeing expertise, and evaluations show our system significantly outperforms general-purpose MLLMs in decision accuracy and explanation quality. All databases, benchmarks, and code will be made available.

cs.AI↗

arXiv2026

Editing on the Generative Manifold: A Theoretical and Empirical Study of General Diffusion-Based Image Editing Trade-offs

Diffusion-based editing has rapidly evolved from curated inpainting tools into general-purpose editors spanning text-guided instruction following, mask-localized edits, drag-based geometric manipulation, exemplar transfer, and training-free composition systems. Despite strong empirical progress, the field lacks a unified treatment of core desiderata that govern practical usability: controllability (how precisely and continuously the user can specify an edit), faithfulness to user intent (semantic alignment to instructions), semantic consistency (preservation of identity and non-target content), locality (containment of changes), and perceptual quality (artifact suppression and detail retention). This paper provides a theoretical and empirical analysis of general diffusion-based image editing, connecting diverse paradigms through a common view of editing as guided transport on a learned image manifold. We first formalize editing as an operator induced by a conditional reverse-time generative process and define task-agnostic metrics capturing instruction adherence, region preservation, semantic consistency, and stability under repeated edits. We then develop theory describing edit dynamics under (i) noise-injection and denoising transport, (ii) inversion-and-edit pipelines and the propagation of inversion errors, and (iii) locality constraints implemented via masked guidance or hard constraints. Under mild Lipschitz assumptions on the learned score or flow field, we derive bounds connecting guidance strength and inversion error to measurable deviations in non-target regions, and we characterize accumulation effects under iterative multi-turn editing. Empirically, we benchmark representative paradigms.

cs.MM↗

arXiv2026

FODMP: Fast One-Step Diffusion of Movement Primitives Generation for Time-Dependent Robot Actions

Diffusion models are increasingly used for robot learning, but current designs face a clear trade-off. Action-chunking diffusion policies like ManiCM are fast to run, yet they only predict short segments of motion. This makes them reactive, but unable to capture time-dependent motion primitives, such as following a spring-damper-like behavior with built-in dynamic profiles of acceleration and deceleration. Recently, Movement Primitive Diffusion (MPD) partially addresses this limitation by parameterizing full trajectories using Probabilistic Dynamic Movement Primitives (ProDMPs), thereby enabling the generation of temporally structured motions. Nevertheless, MPD integrates the motion decoder directly into a multi-step diffusion process, resulting in prohibitively high inference latency that limits its applicability in real-time control settings. We propose FODMP (Fast One-step Diffusion of Movement Primitives), a new framework that distills diffusion models into the ProDMPs trajectory parameter space and generates motion using a single-step decoder. FODMP retains the temporal structure of movement primitives while eliminating the inference bottleneck through single-step consistency distillation. This enables robots to execute time-dependent primitives at high inference speed, suitable for closed-loop vision-based control. On standard manipulation benchmarks (MetaWorld, ManiSkill), FODMP runs up to 10 times faster than MPD and 7 times faster than action-chunking diffusion policies, while matching or exceeding their success rates. Beyond speed, by generating fast acceleration-deceleration motion primitives, FODMP allows the robot to intercept and securely catch a fast-flying ball, whereas action-chunking diffusion policy and MPD respond too slowly for real-time interception.

cs.RO↗

arXiv2026

A Robust Geometric Distortion Solution for Main Survey Camera of CSST

The advancement in sensitivity and field of view of next-generation wide-field survey telescopes requires astrometric measurements with high precision, even in the presence of significant geometric distortions. To address this challenge, we develop a Weighted Polynomial Distortion Correction in 2-Phase (WPDC-2P) method. This approach enhances stellar cross-matching, incorporates distance-based weighting into the traditional polynomial fitting, and employs a look-up table to absorb the remaining distortion residuals. Validated on simulated data from the Main Survey Camera of the \emph{Chinese Space Station Survey Telescope} (CSST), incorporating geometric distortions up to approximately $200$ pixels, the method achieves astrometric standard deviation ranging from 0.013 to 0.107 pixels (0.03 pixels for the $g$-1 detector) across all 18 detectors. Under extreme crowding conditions (e.g., globular cluster NGC 2298), the astrometric precision for the $g$-1 detector reaches 0.05-pixel level within the central region ($r_d < 4000$), despite a centroiding precision of $\sim$0.04 pixels. When applied to the Beijing-Arizona Sky Survey data, for which the standard pipeline delivers an astrometric uncertainty of $\sim$20 mas, our method reduces the positional scatter to $ σ_{Δα}=5.494$ mas (0.01 pixels) and $ σ_{Δδ}=9.981$ mas (0.02 pixels) using only a weighted 3rd-order polynomial correction. The method has been integrated into the CSST data processing pipeline and is prepared for further refinement using on-orbit calibration data.

astro-ph.IM↗

arXiv2026

Output Prediction of Quantum Circuits based on Graph Neural Networks

The output prediction of quantum circuits is a formidably challenging task imperative in developing quantum devices. Motivated by the natural graph representation of quantum circuits, this paper proposes a Graph Neural Networks (GNNs)-based framework to predict the output expectation values of quantum circuits under noisy and noiseless conditions and compare the performance of different parameterized quantum circuits (PQCs). We construct datasets under noisy and noiseless conditions using a non-parameterized quantum gate set to predict circuit expectation values. The node feature vectors for GNNs are specifically designed to include noise information. In our simulations, we compare the prediction performance of GNNs in both noisy and noiseless conditions against Convolutional Neural Networks (CNNs) on the same dataset and their qubit scalability. GNNs demonstrate superior prediction accuracy across diverse conditions. Subsequently, we utilize the parameterized quantum gate set to construct noisy PQCs and compute the ground state energy of hydrogen molecules using the Variational Quantum Eigensolver (VQE). We propose two schemes: the Indirect Comparison scheme, which involves directly predicting the ground state energy and subsequently comparing circuit performances, and the Direct Comparison scheme, which directly predicts the relative performance of the two circuits. Simulation results indicate that the Direct Comparison scheme significantly outperforms the Indirect Comparison scheme by an average of 36.2% on the same dataset, providing a new and effective perspective for using GNNs to predict the overall properties of PQCs, specifically by focusing on their performance differences.

quant-ph↗

arXiv2026

How should AI knowledge be governed? Epistemic authority, structural transparency, and the case for open cognitive graphs

Through widespread use in formative assessment and self-directed learning, educational AI systems exercise de facto epistemic authority. Unlike human educators, however, these systems are not embedded in institutional mechanisms of accountability, review, and correction, creating a structural governance challenge that cannot be resolved through application-level regulation or model transparency alone. This paper reconceptualizes educational AI as public educational cognitive infrastructure and argues that its governance must address the epistemic authority such systems exert. We propose the Open Cognitive Graph (OCG) as a technical interface that externalizes pedagogical structure in forms aligned with human educational reasoning. By explicitly representing concepts, prerequisite relations, misconceptions, and scaffolding, OCGs make the cognitive logic governing AI behaviour inspectable and revisable. Building on this foundation, we introduce the trunk-branch governance model, which organizes epistemic authority across layers of consensus and pluralism. A case study of a community-governed educational foundation model demonstrates how distributed expertise can be integrated through institutionalized processes of validation, correction, and propagation. The paper concludes by discussing implications for educational equity, AI policy, and sustainability. By shifting attention from access to governance conditions, the proposed framework offers a structural approach to aligning educational AI with democratic accountability and public responsibility.

cs.CY↗

arXiv2026

$\mathfrak{G}$-Quotients of Grassmannians and Equations

Laurent Lafforgue's presentation of a Grassmannian Gr$^{d, E}$ naturally comes equipped with the induced action of a subtorus $\mathbb{T}_\bullet$ of PGL$(E)$. By investigating the defining ideals of $\mathbb{T}_\bullet$-orbit closures through general points of Gr$^{d,E}$ and studying their degenerations, we obtain a morphsim $\mathfrak{q}: \mathbb{F}^{d, E_\bullet} \to \mathbb{H}^{d, E_{\bullet}}$ such that $\mathbb{H}^{d, E_\bullet}$, termed the $\mathfrak{G}$-quotient of Gr$^{d,E}$ by $\mathbb{T}_\bullet$, is birational to $[{\rm Gr}^{d, E}/\mathbb{T}_\bullet]$, and $\mathfrak{q}$, termed $\mathfrak{G}$-family of Gr$^{d,E}$ by $\mathbb{T}_\bullet$, is a family of general $\mathbb{T}_\bullet$-orbit closures and their degenerations. We obtain a series of new results on $\mathbb{H}^{d, E_{\bullet}}$ and $\mathbb{F}^{d, E_\bullet}$.

math.AG↗

arXiv2026

Integrated laser heating stage with active geometry modulation for simultaneous in-situ X-ray transmission and evolved gas analysis of molten liquids

We report the design and development of a compact, integrated laser heating stage tailored for in situ high-temperature X ray transmission studies of molten oxides. In horizontal beam geometries, widely used in both laboratory and synchrotron facilities, the natural spreading (wetting) of molten samples on substrates significantly reduces the effective vertical optical path length, detrimental to signal quality in transmission-mode measurements. To overcome this limitation, we introduced a thermocouple assisted active geometry modulation technique. This method mechanically lifts the spreading melt into a liquid bridge via surface tension, optimizing the transmission path length while simultaneously enabling in situ temperature monitoring. The device features a triple fiber coupled laser head with high power density, a precision closed loop Proportional Integral Derivative temperature control system, and an atmosphere controlled vacuum chamber coupled with a mass spectrometer. This integration allows for simultaneous evolved gas analysis, enabling the correlation of structural phase transitions with chemical volatilization or reaction dynamics. Validated by tracking the melting kinetics of a multicomponent glass precursor, this versatile setup provides a comprehensive solution for high quality data acquisition in X ray transmission experiments across various sources.

physics.app-ph↗

arXiv2026

Deep-learning-based prediction of Precipitable Water Vapor in the Chajnantor area

Astronomical observations at millimeter and submillimeter wavelengths heavily depend on the amount of Precipitable Water Vapor (PWV) in the atmosphere, directly affecting the sky transparency and degrading the quality of the signals received by radio telescopes. Predictions of PWV at different forecasting horizons is crucial to support telescope operations, engineering planning, and observational scheduling and efficiency of radio observatories installed in the Chajnantor area in northern Chile. We developed and validated a Long Short-Term Memory (LSTM) deep learning-based model to predict PWV at forecasting horizons of 12, 24, 36, and 48 hours using historical data from two 183 GHz radiometers and a weather station in the Chajnantor area. We find the LSTM method is able to predict PWV in the 12 and 24 hours forecasting horizons with Mean Absolute Percentage Error (MAPE) of 22% compared to 36% of the traditional Global Forecast System (GFS) method used by Atacama Pathfinder EXperiment (APEX) and the Root Mean Square Error (RMSE) in mm are reduced by 50%. We present a first application of deep learning techniques for preliminary predictions of PWV in the Chajnantor area. The prediction performance shows significant improvements to traditional methods in 12 and 24 hours time windows. We also propose upgrades to improve our method in short (< 1 hour) and long (> 36 hours) forecasting timescales for future work.

astro-ph.IM↗

arXiv2026

Antarctic Infrared Binocular Telescope: Early Data Release of observations in the 1.4 μm water-vapor-absorption band

Ground-based observations around 1.4 $μ$m are normally limited by strong absorption of telluric water-vapor. However, Dome A, Antarctica has exceptionally dry conditions that offer a unique opportunity for observations in this band. We designed a new filter covering 1.34--1.48 $μ$m, namely $W'$, and installed it on the Antarctic Infrared Binocular Telescope (AIRBT) at Dome A in 2025. AIRBT comprises two identical 15 cm optical tube assemblies and two InGaAs cameras equipped with $J$ and $W'$ filters, respectively. With this Early Data Release (EDR), we aim to evaluate the performance of the $W'$ band at Dome A to observe objects with water-vapor features. This EDR covers $\thicksim 20 \ \mathrm{deg^2}$ in the Galactic plane using $\thicksim 20,000$ images in three nights. For 2 s exposures, the 5 $σ$ limiting magnitude histogram peaks at $J \thicksim 11.5$ mag (Vega) and $W' \thicksim 9.9$ mag, respectively. The $J-W'$ vs $J-H$ color-color diagram distinguishes ultracool candidates with water-vapor-absorption features from reddened early type stars. Furthermore, later-type stars tend to exhibit stronger water-vapor absorption. Some sources show larger $ΔW'$ than $ΔJ$ across the three nights, which we attribute to variations of their water-vapor-absorption depth. We conclude that it will be efficient to search for ultracool stars and estimate their spectral subtypes using $W'$ band imaging at Dome A, where the atmospheric transmission is high and stable.

astro-ph.IM↗

arXiv2026

SubTokenTest: A Practical Benchmark for Real-World Sub-token Understanding

Recent advancements in large language models (LLMs) have significantly enhanced their reasoning capabilities. However, they continue to struggle with basic character-level tasks, such as counting letters in words, a problem rooted in their tokenization process. While existing benchmarks have highlighted this weakness through basic character operations, such failures are often dismissed due to lacking practical relevance. Yet, many real-world applications, such as navigating text-based maps or interpreting structured tables, rely heavily on precise sub-token understanding. In this regard, we introduce SubTokenTest, a comprehensive benchmark that assesses sub-token understanding through practical, utility-driven tasks. Our benchmark includes ten tasks across four domains and isolates tokenization-related failures by decoupling performance from complex reasoning. We provide a comprehensive evaluation of nine advanced LLMs. Additionally, we investigate the impact of test-time scaling on sub-token reasoning and explore how character-level information is encoded within the hidden states.

cs.CL↗

arXiv2026

Towards a Mechanistic Understanding of Large Reasoning Models: A Survey of Training, Inference, and Failures

Reinforcement learning (RL) has catalyzed the emergence of Large Reasoning Models (LRMs) that have pushed reasoning capabilities to new heights. While their performance has garnered significant excitement, exploring the internal mechanisms driving these behaviors has become an equally critical research frontier. This paper provides a comprehensive survey of the mechanistic understanding of LRMs, organizing recent findings into three core dimensions: 1) training dynamics, 2) reasoning mechanisms, and 3) unintended behaviors. By synthesizing these insights, we aim to bridge the gap between black-box performance and mechanistic transparency. Finally, we discuss under-explored challenges to outline a roadmap for future mechanistic studies, including the need for applied interpretability, improved methodologies, and a unified theoretical framework.

cs.CL↗

arXiv2026

MindWatcher: Toward Smarter Multimodal Tool-Integrated Reasoning

Traditional workflow-based agents exhibit limited intelligence when addressing real-world problems requiring tool invocation. Tool-integrated reasoning (TIR) agents capable of autonomous reasoning and tool invocation are rapidly emerging as a powerful approach for complex decision-making tasks involving multi-step interactions with external environments. In this work, we introduce MindWatcher, a TIR agent integrating interleaved thinking and multimodal chain-of-thought (CoT) reasoning. MindWatcher can autonomously decide whether and how to invoke diverse tools and coordinate their use, without relying on human prompts or workflows. The interleaved thinking paradigm enables the model to switch between thinking and tool calling at any intermediate stage, while its multimodal CoT capability allows manipulation of images during reasoning to yield more precise search results. We implement automated data auditing and evaluation pipelines, complemented by manually curated high-quality datasets for training, and we construct a benchmark, called MindWatcher-Evaluate Bench (MWE-Bench), to evaluate its performance. MindWatcher is equipped with a comprehensive suite of auxiliary reasoning tools, enabling it to address broad-domain multimodal problems. A large-scale, high-quality local image retrieval database, covering eight categories including cars, animals, and plants, endows model with robust object recognition despite its small size. Finally, we design a more efficient training infrastructure for MindWatcher, enhancing training speed and hardware utilization. Experiments not only demonstrate that MindWatcher matches or exceeds the performance of larger or more recent models through superior tool invocation, but also uncover critical insights for agent training, such as the genetic inheritance phenomenon in agentic RL.

cs.AI↗

arXiv2025

Blowup rate for rotational NLS with a repulsive potential

In this paper we give an analytical proof of the ``$\log$-$\log$'' blowup rate for mass-critical nonlinear Schrödinger equation (NLS) with a rotation ($Ω\neq 0$) and a repulsive harmonic potential $V_γ(x) = \textrm{sgn}(γ) γ^2 |x|^2$, $γ< 0$ when the initial data has a mass slightly above that of $Q$, the ground state solution to the free NLS. The proof is based on a virial identity and an $\mathcal{R}_γ$-transform, a pseudo-conformal transform in this setting. Further, we obtain a limiting behavior description concerning the mass concentration near blowup time. A remarkable finding is that increasing the value $|γ|$ for the repulsive potential $V_γ$ can give rise to global in time solution for the focusing RNLS, which is in contrast to the case where $γ$ is positive. This kind of phenomenon was earlier observed in the non-rotational case $Ω= 0$ in Carles' work. In addition, we provide numerical simulations to partially illustrate the blowup profile along with the blowup rate using dynamic rescaling and adaptive mesh refinement method.

math.AP↗

arXiv2025

What Affects the Effective Depth of Large Language Models?

The scaling of large language models (LLMs) emphasizes increasing depth, yet performance gains diminish with added layers. Prior work introduces the concept of "effective depth", arguing that deeper models fail to fully utilize their layers for meaningful computation. Building on this, we systematically study how effective depth varies with model scale, training type, and task difficulty. First, we analyze the model behavior of Qwen-2.5 family (1.5B-32B) and find that while the number of effective layers grows with model size, the effective depth ratio remains stable. Besides, comparisons between base and corresponding long-CoT models show no increase in effective depth, suggesting that improved reasoning stems from longer context rather than deeper per-token computation. Furthermore, evaluations across tasks of varying difficulty indicate that models do not dynamically use more layers for harder problems. Our results suggest that current LLMs underuse available depth across scales, training paradigms and tasks of varying difficulties, pointing out research opportunities on increasing the layer utilization rate of LLMs, model pruning, and early exiting. Our code is released at https://github.com/AheadOFpotato/what_affects_effective_depth.

cs.CL↗

arXiv2025

RecGPT-V2 Technical Report

Large language models (LLMs) have demonstrated remarkable potential in transforming recommender systems from implicit behavioral pattern matching to explicit intent reasoning. While RecGPT-V1 successfully pioneered this paradigm by integrating LLM-based reasoning into user interest mining and item tag prediction, it suffers from four fundamental limitations: (1) computational inefficiency and cognitive redundancy across multiple reasoning routes; (2) insufficient explanation diversity in fixed-template generation; (3) limited generalization under supervised learning paradigms; and (4) simplistic outcome-focused evaluation that fails to match human standards. To address these challenges, we present RecGPT-V2 with four key innovations. First, a Hierarchical Multi-Agent System restructures intent reasoning through coordinated collaboration, eliminating cognitive duplication while enabling diverse intent coverage. Combined with Hybrid Representation Inference that compresses user-behavior contexts, our framework reduces GPU consumption by 60% and improves exclusive recall from 9.39% to 10.99%. Second, a Meta-Prompting framework dynamically generates contextually adaptive prompts, improving explanation diversity by +7.3%. Third, constrained reinforcement learning mitigates multi-reward conflicts, achieving +24.1% improvement in tag prediction and +13.0% in explanation acceptance. Fourth, an Agent-as-a-Judge framework decomposes assessment into multi-step reasoning, improving human preference alignment. Online A/B tests on Taobao demonstrate significant improvements: +2.98% CTR, +3.71% IPV, +2.19% TV, and +11.46% NER. RecGPT-V2 establishes both the technical feasibility and commercial viability of deploying LLM-powered intent reasoning at scale, bridging the gap between cognitive exploration and industrial utility.

cs.IR↗