arXiv subjects

Yukun Huang

Publications and source records attributed to Yukun Huang.

At least 19 recordsLinked to original sources

arXiv2026

SearchAtlas: Analyzing Agentic Search Strategies via Evidential Query Graphs

LLM search agents are often evaluated on final-answer accuracy, overlooking the process. Analyzing a search strategy requires understanding how credible evidence is retrieved to address question constraints. This valuable information is buried in raw search trajectories that are long and difficult to parse. We introduce SearchAtlas, a framework that converts search trajectories into structured graphs whose edges represent how evidence is propagated across the reasoning trace, from the query that retrieves it to the final answer. Our automated parsing pipeline achieves a mean edge F1 of 86.0% against human-annotated graphs and remains consistent across repeated runs. We analyze five search agents on three benchmarks, revealing systematic differences in search scale and evidence aggregation. SearchAtlas exposes fragmented answer support, question constraints that do not reach the answer, and unverified parametric knowledge entering the response. These process failures are strongly associated with incorrect answers, even more so than an LLM judge given either the raw trajectory or the ordered query list, suggesting that the constructed graphs provide useful interpretability. Moreover, an audit of cases in which process-diagnostic scores disagree with final-answer correctness shows that they capture information not reducible to answer accuracy.

cs.CL↗

arXiv2026

Lazy Grounding: Attacking Search Agents with Factual Evidence

Search agents mitigate hallucination by grounding their answers in retrieved web results. However, retrieval-based approaches also introduce an attack surface: agents may cite misinformation from poisoned search corpora containing false or malicious documents. We demonstrate that, in some cases, search agents' reasoning and responses may be steered by completely factual but distracting information. We refer to this failure as lazy grounding. We expose lazy grounding by injecting nearby evidence from answer-changing rewrites of benchmark questions into the search corpora. Each document contains factual evidence that supports a neighboring rewritten question but is retrieved for the original question. Across 12 model-benchmark pairs, the attack causes the accuracy of search agents' responses to drop by 5.9 points on average and by up to 17.3 points, while inducing nearby-answer adoption in every setting. The effect is even stronger when nearby evidence appears later or is more answer-shaped. Our results show that robust search agents must defend against not only misinformation but also the misapplication of factual evidence. The code is publicly available at https://github.com/frankyzha/lazy-grounding.

cs.CL↗

arXiv2026

Cite Pretrain: Retrieval-Free Knowledge Attribution for Large Language Models

Trustworthy language models should provide both correct and verifiable answers. However, citations generated directly by standalone LLMs are often unreliable. As a result, current systems insert citations by querying an external retriever at inference time, introducing latency, infrastructure dependence, and vulnerability to retrieval noise. We explore whether LLMs can be made to reliably attribute to the documents seen during continual pretraining without test-time retrieval, by revising the training process. To study this, we construct CitePretrainBench, a benchmark that mixes real-world corpora (Wikipedia, Common Crawl, arXiv) with novel documents and probes both short-form (single-fact) and long-form (multi-fact) citation tasks. Our approach follows a two-stage process: (1) continual pretraining to index factual knowledge by binding it to persistent document identifiers; and (2) instruction tuning to elicit citation behavior. We introduce Active Indexing for the first stage, which creates generalizable, source-anchored bindings by augmenting training with synthetic data that (i) restate each fact in diverse, compositional forms and (ii) enforce bidirectional training (source-to-fact and fact-to-source). This equips the model to both generate content from a cited source and attribute its own answers, improving robustness to paraphrase and composition. Experiments with Qwen-2.5-7B&3B show that Active Indexing consistently outperforms a Passive Indexing baseline, which simply appends an identifier to each document, achieving citation precision gains of up to 30.2% across all tasks and models. Our ablation studies reveal that performance continues to improve as we scale the amount of augmented data, showing a clear upward trend even at 16x the original token count. Finally, we show that internal citations complement external ones by making the model more robust to retrieval noise.

cs.AI↗

arXiv2026

DeepFact: Co-Evolving Benchmarks and Agents for Deep Research Factuality

Search-augmented LLM agents can produce deep research reports (DRRs), but verifying claim-level factuality remains challenging. Existing fact-checkers are primarily designed for general-domain, factoid-style atomic claims, and there is no benchmark to test whether such verifiers transfer to DRRs. Yet building such a benchmark is itself difficult. We first show that static expert-labeled benchmarks are brittle in this setting: in a controlled study with PhD-level specialists, unassisted experts achieve only 60.8% accuracy on a hidden micro-gold set of verifiable claims. We propose Evolving Benchmarking via Audit-then-Score (AtS), where benchmark labels and rationales are explicitly revisable: when a verifier disagrees with the current benchmark, it must submit evidence; an auditor adjudicates the dispute; and accepted revisions update the benchmark before models are scored. Across four AtS rounds, expert micro-gold accuracy rises to 90.9%, indicating experts are substantially more reliable as auditors than as one-shot labelers. We instantiate AtS as DeepFact-Bench, a versioned DRR factuality benchmark with auditable rationales, and DeepFact-Eval, a document-level verification agent (with a grouped lite variant) that outperforms existing verifiers on DeepFact-Bench and transfers well to external factuality datasets.

cs.AI↗

arXiv2026

CineScene: Implicit 3D as Effective Scene Representation for Cinematic Video Generation

Cinematic video production requires control over scene-subject composition and camera movement, but live-action shooting remains costly due to the need for constructing physical sets. To address this, we introduce the task of cinematic video generation with decoupled scene context: given multiple images of a static environment, the goal is to synthesize high-quality videos featuring dynamic subject while preserving the underlying scene consistency and following a user-specified camera trajectory. We present CineScene, a framework that leverages implicit 3D-aware scene representation for cinematic video generation. Our key innovation is a novel context conditioning mechanism that injects 3D-aware features in an implicit way: By encoding scene images into visual representations through VGGT, CineScene injects spatial priors into a pretrained text-to-video generation model by additional context concatenation, enabling camera-controlled video synthesis with consistent scenes and dynamic subjects. To further enhance the model's robustness, we introduce a simple yet effective random-shuffling strategy for the input scene images during training. To address the lack of training data, we construct a scene-decoupled dataset with Unreal Engine 5, containing paired videos of scenes with and without dynamic subjects, panoramic images representing the underlying static scene, along with their camera trajectories. Experiments show that CineScene achieves state-of-the-art performance in scene-consistent cinematic video generation, handling large camera movements and demonstrating generalization across diverse environments.

cs.CV↗

arXiv2026

Analytical Solutions for Planet-Scattering Small Bodies

Gravitational scattering of small bodies (planetesimals) by a planet remains a fundamental problem in celestial mechanics. It is traditionally modeled within the circular restricted three-body problem (CR3BP), where individual particle trajectories are obtained via numerical integrations. Here, we use {Ö}pik's close-encounter framework to study the random walk of the orbital energy $x$ for an ensemble of test particles on planet-crossing orbits. We show that the evolution of each particle's orbital elements $(a, e, i)$ is fully encapsulated by the 3D rotation of the relative velocity vector $\bm{U}_\infty$, whose magnitude remains constant. Consequently, the system can be reduced to two degrees of freedom. By averaging over all possible flyby geometries, we derive explicit expressions for the drift and diffusion coefficients of $x$. We then solve the resulting Fokker--Planck equation to obtain a closed-form solution for the time evolution of the particle distribution. A characteristic scattering timescale naturally emerges, scaling as $(P_{p}/M_{p}^{2})/500$, where $P_{p}$ is the planet's orbital period and $M_{p}$ its mass ratio to the central star. The typical ejection speed of small bodies by a planet is estimated to be $3 v_p M_{p}^{1/3}$, where $v_p$ is the planet's orbital speed. Our analytical solution constitutes a universal law applicable to both the Solar System and exoplanetary systems, providing a computationally efficient alternative to costly $N$-body simulations for studying the orbital distributions and ejection of planetesimals and planets (e.g., Kuiper Belt, Oort Cloud, debris disks, interstellar objects, and free-floating planets).

astro-ph.EP↗

arXiv2025

Soliton-Assisted Massive Signal Broadcasting via Exceptional Points

Chip-scale all-optical signal broadcasting enables data replication from an optical signal to a large number of wavelength channels, playing a critical role in enabling massive-throughput optical communication and computing systems. The underlying process is four-wave mixing between an optical signal and a multi-wavelength pump source via optical Kerr nonlinearity. To enhance the generally weak nonlinearity, high-quality (Q) microcavities are commonly used to achieve practical efficiency. However, the ultra-narrow linewidths of high Q cavities prohibit achieving massive throughput broadcasting due to Fourier reciprocity. Here, we overcome this challenge by harnessing a parity-time symmetric coupled-cavity system that supports equally spaced exceptional points in the frequency domain. This design seamlessly integrates generation of dissipative Kerr soliton comb source and all-optical signal broadcasting into a unified nonlinear process. As a result, we realize soliton-assisted intracavity massive signal broadcasting with a channel count exceeding 100 over 200 nm wavelength range, resulting in Terabit-per-second aggregated rates. This throughput surpasses the intrinsic microcavity linewidth constraint (~200 MHz) by over three orders of magnitude. We further demonstrate the utility of this approach through an optical convolutional accelerator, highlighting its potential to enable transformative capabilities in photonic computing. Our work establishes a new paradigm for chip-scale photonic processing devices based on non-Hermitian optical design.

physics.optics↗

arXiv2025

OmniX: From Unified Panoramic Generation and Perception to Graphics-Ready 3D Scenes

There are two prevalent ways for automatic 3D scene construction: procedural generation and 2D lifting. Among these, panorama-based 2D lifting has emerged as a promising technique, leveraging powerful 2D generative priors to produce immersive, realistic, and diverse 3D environments. In this work, we advance this technique to generate graphics-ready 3D scenes suitable for physically based rendering (PBR), relighting, and simulation. Our key insight is to repurpose 2D generative models for panorama perception of geometry, textures, and PBR materials. Unlike existing 2D lifting approaches that emphasize appearance generation and neglect the perception of intrinsic properties, we present OmniX, a versatile and unified framework for panorama generation, perception, and completion. Built upon cross-modal adapter structure and cyclic spatial operators, OmniX effectively repurposes pre-trained 2D flow matching priors for joint modeling of multimodal, seamless equirectangular representations. Furthermore, we construct a large-scale synthetic panorama dataset comprising high-quality multimodal panoramas from diverse indoor and outdoor scenes. Extensive experiments demonstrate the effectiveness and generality of OmniX as a unified framework for panorama generation and perception across geometry, lighting, and semantics, enabling graphics-ready 3D scene generation and opening new possibilities for immersive and physically realistic virtual world creation.

cs.CV↗

arXiv2025

The Detectability of Lunar-Origin Asteroids in the LSST Era

While most near-Earth asteroids (NEAs) are thought to originate from the main belt, recent discoveries have suggested the existence of a lunar-derived NEA population, such as the asteroids Kamo'oalewa and 2024 PT5. These objects may hold key clues to the dynamical evolution of NEAs and the recent impact history of the Earth-Moon system. However, the population, distribution, and dynamical characteristics of these Lunar-Origin Asteroids (LOAs) remain poorly constrained. By combining the lunar ejecta production with N-body orbital simulations of the ejecta, we investigate their orbital evolution in the past millions of years and the current LOA population, revealing their significant potential for detection by future surveys. Specifically for the Vera C. Rubin Observatory's upcoming Legacy Survey of Space and Time (LSST), we predict an average detection rate of about 6 LOAs (with D > 5 m) per year. Additionally, we find that the LOAs tend to approach from sunward and anti-sunward directions, with encounter velocities significantly lower than those of typical NEAs. These findings offer valuable insights in guiding targeted ground-based surveys and planetary defense efforts for LOAs in the future.

astro-ph.EP↗

arXiv2025

When Greedy Wins: Emergent Exploitation Bias in Meta-Bandit LLM Training

While Large Language Models (LLMs) hold promise to become autonomous agents, they often explore suboptimally in sequential decision-making. Recent work has sought to enhance this capability via supervised fine-tuning (SFT) or reinforcement learning (RL), improving regret on the classic multi-armed bandit task. However, it remains unclear how these learning methods shape exploration strategies and how well they generalize. We investigate both paradigms by training LLMs with SFT on expert trajectories and RL with a range of tailored reward signals including a strategic, regret-shaped reward to reduce variance, and an algorithmic reward that enables oracle imitation. The resulting agents outperform pre-trained models and achieve performance comparable to Upper Confidence Bound (UCB) and Thompson Sampling, with robust generalization to 6x longer horizons and across bandit families. Behavioral analysis reveals that gains often stem from more sophisticated but greedier exploitation: RL/SFT agents are more prone to early catastrophic failure than pre-trained models, prematurely abandoning exploration. Furthermore, agents trained to imitate UCB learn to outperform their teacher by adopting more exploitative variants. Our findings clarify when each training paradigm is preferable and advocate tailored reward design and evaluation beyond average regret to promote robust exploratory behavior.

cs.LG↗

arXiv2025

Early Stellar Flybys are Unlikely: Improved Constraints from Sednoids and Large-$q$ TNOs

Sedna-like objects (a.k.a. sednoids) are transneptunian objects (TNOs) characterized by large semimajor axes and exceptionally high perihelia. Their high-$q$ orbits are detached from the influence of the four giant planets and need extra perturbation to form. One hypothesis posits that close stellar flybys could have perturbed objects from the primordial scattering disk, generating the sednoid population. In this study, we run N-body simulations with different stellar encounter configurations to explore whether such a close stellar flyby can satisfy new constraints identified from sednoid (and detached extreme TNO) observation, including the low-inclination ($i<30^\circ$) profile and primordial orbital alignment. Our results suggest that flybys with field stars are unable to generate a sufficient population, whereas flybys within the birth cluster fail to produce the primordial orbital alignment. To meet the inclination constraint of detached extreme TNOs, flybys have to be either coplanar ($i_\star \sim 0^\circ$) or symmetric about the ecliptic plane ($ω_\star \sim 0^\circ, i_\star \sim 90^\circ$). After taking into account their occurrence rate at the early stage of the Solar System, we conclude that close-in stellar flybys ($q_\star \le 1000$~au) that satisfy all constraints are unlikely to happen ($\lesssim$5\%). Future discoveries of additional sednoids with precise orbital determinations are crucial to confirm the existence of the low-inclination tendency and the primordial alignment, and to further constrain the early dynamical evolution of the Solar System.

astro-ph.EP↗

arXiv2025

Discovery and dynamics of a Sedna-like object with a perihelion of 66 au

Trans-Neptunian objects (TNOs) with large perihelion distances ($q > 60$ au) and semi-major axes ($a > 200$ au) provide insights into the early evolution of the solar system and the existence of a hypothetical distant planet. These objects are still rare and their detection is challenging, yet they play a crucial role in constraining models of solar system formation. Here we report the discovery of a Sedna-like TNO, 2023\,KQ$_{14}$, nicknamed `Ammonite', with $q = 66$ au, $a = 252$ au, and inclination $i=11^\circ$. Ammonite's orbit does not align with those of the other Sedna-like objects and fills the previously unexplained `$q$-gap' in the observed distribution of distant solar system objects. Simulations demonstrate that Ammonite is dynamically stable over 4.5 billion years. % with less than 1\% variation in its semi-major axis. Our analysis suggests that Ammonite and the other Sedna-like objects may have shared a primordial orbital clustering around 4.2 billion years ago. Furthermore, Ammonite's stable orbit favors larger orbits ($\sim$ 500 au) rather than closer ones for a large hypothetical planet in present-day trans-Neptunian space.

astro-ph.EP↗

arXiv2025

Real-time Factuality Assessment from Adversarial Feedback

We show that existing evaluations for assessing the factuality of news from conventional sources, such as claims on fact-checking websites, result in high accuracies over time for LLM-based detectors-even after their knowledge cutoffs. This suggests that recent popular false information from such sources can be easily identified due to its likely presence in pre-training/retrieval corpora or the emergence of salient, yet shallow, patterns in these datasets. Instead, we argue that a proper factuality evaluation dataset should test a model's ability to reason about current events by retrieving and reading related evidence. To this end, we develop a novel pipeline that leverages natural language feedback from a RAG-based detector to iteratively modify real-time news into deceptive variants that challenge LLMs. Our iterative rewrite decreases the binary classification ROC-AUC by an absolute 17.5 percent for a strong RAG-based GPT-4o detector. Our experiments reveal the important role of RAG in both evaluating and generating challenging news examples, as retrieval-free LLM detectors are vulnerable to unseen events and adversarial attacks, while feedback from RAG-based evaluation helps discover more deceitful patterns.

cs.CL↗

arXiv2025

OmniPart: Part-Aware 3D Generation with Semantic Decoupling and Structural Cohesion

The creation of 3D assets with explicit, editable part structures is crucial for advancing interactive applications, yet most generative methods produce only monolithic shapes, limiting their utility. We introduce OmniPart, a novel framework for part-aware 3D object generation designed to achieve high semantic decoupling among components while maintaining robust structural cohesion. OmniPart uniquely decouples this complex task into two synergistic stages: (1) an autoregressive structure planning module generates a controllable, variable-length sequence of 3D part bounding boxes, critically guided by flexible 2D part masks that allow for intuitive control over part decomposition without requiring direct correspondences or semantic labels; and (2) a spatially-conditioned rectified flow model, efficiently adapted from a pre-trained holistic 3D generator, synthesizes all 3D parts simultaneously and consistently within the planned layout. Our approach supports user-defined part granularity, precise localization, and enables diverse downstream applications. Extensive experiments demonstrate that OmniPart achieves state-of-the-art performance, paving the way for more interpretable, editable, and versatile 3D content.

cs.CV↗

arXiv2025

DreamComposer++: Empowering Diffusion Models with Multi-View Conditions for 3D Content Generation

Recent advancements in leveraging pre-trained 2D diffusion models achieve the generation of high-quality novel views from a single in-the-wild image. However, existing works face challenges in producing controllable novel views due to the lack of information from multiple views. In this paper, we present DreamComposer++, a flexible and scalable framework designed to improve current view-aware diffusion models by incorporating multi-view conditions. Specifically, DreamComposer++ utilizes a view-aware 3D lifting module to extract 3D representations of an object from various views. These representations are then aggregated and rendered into the latent features of target view through the multi-view feature fusion module. Finally, the obtained features of target view are integrated into pre-trained image or video diffusion models for novel view synthesis. Experimental results demonstrate that DreamComposer++ seamlessly integrates with cutting-edge view-aware diffusion models and enhances their abilities to generate controllable novel views from multi-view conditions. This advancement facilitates controllable 3D object reconstruction and enables a wide range of applications.

cs.CV↗

arXiv2025

FilMaster: Bridging Cinematic Principles and Generative AI for Automated Film Generation

AI-driven content creation has shown potential in film production. However, existing film generation systems struggle to implement cinematic principles and thus fail to generate professional-quality films, particularly lacking diverse camera language and cinematic rhythm. This results in templated visuals and unengaging narratives. To address this, we introduce FilMaster, an end-to-end AI system that integrates real-world cinematic principles for professional-grade film generation, yielding editable, industry-standard outputs. FilMaster is built on two key principles: (1) learning cinematography from extensive real-world film data and (2) emulating professional, audience-centric post-production workflows. Inspired by these principles, FilMaster incorporates two stages: a Reference-Guided Generation Stage which transforms user input to video clips, and a Generative Post-Production Stage which transforms raw footage into audiovisual outputs by orchestrating visual and auditory elements for cinematic rhythm. Our generation stage highlights a Multi-shot Synergized RAG Camera Language Design module to guide the AI in generating professional camera language by retrieving reference clips from a vast corpus of 440,000 film clips. Our post-production stage emulates professional workflows by designing an Audience-Centric Cinematic Rhythm Control module, including Rough Cut and Fine Cut processes informed by simulated audience feedback, for effective integration of audiovisual elements to achieve engaging content. The system is empowered by generative AI models like (M)LLMs and video generation models. Furthermore, we introduce FilmEval, a comprehensive benchmark for evaluating AI-generated films. Extensive experiments show FilMaster's superior performance in camera language design and cinematic rhythm control, advancing generative AI in professional filmmaking.

cs.CV↗

arXiv2025

DreamCube: 3D Panorama Generation via Multi-plane Synchronization

3D panorama synthesis is a promising yet challenging task that demands high-quality and diverse visual appearance and geometry of the generated omnidirectional content. Existing methods leverage rich image priors from pre-trained 2D foundation models to circumvent the scarcity of 3D panoramic data, but the incompatibility between 3D panoramas and 2D single views limits their effectiveness. In this work, we demonstrate that by applying multi-plane synchronization to the operators from 2D foundation models, their capabilities can be seamlessly extended to the omnidirectional domain. Based on this design, we further introduce DreamCube, a multi-plane RGB-D diffusion model for 3D panorama generation, which maximizes the reuse of 2D foundation model priors to achieve diverse appearances and accurate geometry while maintaining multi-view consistency. Extensive experiments demonstrate the effectiveness of our approach in panoramic image generation, panoramic depth estimation, and 3D scene generation.

cs.GR↗

arXiv2025

Fuzzy Speculative Decoding for a Tunable Accuracy-Runtime Tradeoff

Speculative Decoding (SD) enforces strict distributional equivalence to the target model when accepting candidate tokens. While it maintains the target model's generation quality, this strict equivalence limits the speedup achievable by SD and prevents users from trading deviations from the target distribution in exchange for further inference speed gains. To address these limitations, we introduce Fuzzy Speculative Decoding (FSD) - a decoding algorithm that generalizes SD by accepting candidate tokens based on the divergences between the target and draft model distributions. By allowing for controlled divergence from the target model, FSD enables users to flexibly trade generation quality for inference speed. Across several benchmarks, our method is able to achieve significant runtime improvements of over 5 tokens per second faster than SD at only an approximate 2% absolute reduction in benchmark accuracy. In many cases, FSD is even able to match SD benchmark accuracy at over 2 tokens per second faster, demonstrating that distributional equivalence is not necessary to maintain target model performance. Furthermore, FSD can be seamlessly integrated into existing SD extensions; we demonstrate this by applying FSD to EAGLE-2, greatly enhancing this existing extension's efficiency while allowing it to leverage FSD's tunable quality-speed trade-off.

cs.AI↗