arXiv subjects

Biao Wu

Publications and source records attributed to Biao Wu.

At least 19 recordsLinked to original sources

arXiv2026

Quantum Path Finder: Navigating Constraint Space via Non-Abelian Gauge Dynamics

Quantum algorithms are commonly designed within the circuit paradigm or through adiabatic evolution. Here we introduce \emph{Quantum Path Finder} (QPF), a distinct algorithm-design paradigm based on non-Abelian gauge dynamics in a constraint-protected degenerate manifold. Each Boolean variable is promoted to a qutrit whose additional state represents an unassigned variable. The problem Hamiltonian then defines a connected manifold of feasible partial assignments, and time-dependent local qutrit rotations generate coherent transport on this manifold. Unlike conventional adiabatic quantum computation, QPF does not require the state to follow an instantaneous eigenstate along a prescribed interpolation: adiabaticity with respect to the constraint gap protects the feasible manifold, while adiabatic following within that manifold is optional. The algorithmic task is therefore to design paths and interference in constraint space rather than only to engineer a sequence of gates or a ground-state interpolation. We illustrate the framework with three applications. For unstructured search, a two-level analysis and numerical evolution recover Grover-type scaling under explicitly stated assumptions. For complete $3$-XORSAT, the accessible gauge sector has a size-independent gap, in contrast to the exponentially closing gap of a conventional transverse-field adiabatic path. For satisfiable $3$-SAT instances, finite-size benchmarks demonstrate high satisfied-clause fractions at short classical comparison budgets, without claiming an equal-cost quantum advantage. We also outline a superconducting-qutrit implementation and discuss the conceptual relationship between QPF and nondeterministic Turing machine.

quant-ph↗

arXiv2026

On the product of nontrivial cross-intersecting families

Let $n\geq2\ell\geq2k\geq6$ and $n>k+\ell$. We determine the maximum of $|\mathcal A||\mathcal B|$ over all nonempty, nontrivial cross-intersecting families $\mathcal A\subseteq\binom{[n]}k$ and $\mathcal B\subseteq\binom{[n]}\ell$. The exact bound is $|\mathcal A||\mathcal B|\leq\max\{M_1,M_2\}$, where \[ \begin{aligned} M_1 &=\left(\binom{n-1}{k-1}-\binom{n-\ell-1}{k-1}+1\right) \left(\binom{n-1}{\ell-1}-\binom{n-k-1}{\ell-1}+1\right),\\ M_2 &=\left(\binom{n-2}{\ell-2}+1\right) \sum_{i=1}^{2} \left(\binom{n-i}{k-1}-\binom{n-\ell-i}{k-1}\right). \end{aligned} \] Both terms are attained by explicit constructions: a Hilton--Milner-type pair and a two-point construction, respectively. We classify all equality cases, including the additional cross-triangle pairs when $k=\ell=3$. Two structural results underlie the proof. First, we establish a restricted-lex comparison theorem showing that restricted lexicographic initial segments minimize disjointness neighborhoods, and characterize equality at the first nontrivial threshold. Second, a decomposition of critically shifted families into two parts with disjoint cores leads to a sharp product theorem, with equality characterized, for all cross-intersecting pairs admitting this decomposition, without additional nontriviality assumptions. Endpoint comparisons and reverse shifting complete the proof for unequal as well as equal uniformities.

math.CO↗

arXiv2026

Automotive-ENV: Benchmarking Multimodal Models in Automotive Cockpit Environments

Multimodal models have demonstrated strong capabilities in web, desktop, and mobile GUI environments, but their performance in automotive cockpit systems remains largely unexplored. In-vehicle GUIs introduce unique challenges beyond traditional GUI interaction, including vehicle-state reasoning, implicit intent understanding, and safety-critical decision making. To address this gap, we introduce Automotive-ENV, an interactive benchmark for evaluating multimodal models in automotive cockpit environments. Built on Android Automotive OS, Automotive-ENV combines cockpit-screen interaction with vehicle-state signals and deterministic state-based evaluation. The benchmark contains 420 carefully constructed tasks covering routine cockpit operation and safety-critical scenarios. Unlike prior GUI benchmarks that rely on action matching or LLM-based judging, Automotive-ENV evaluates model behavior through reproducible validators over UI states, vehicle signals, and safety-related conditions. We benchmark representative general-purpose multimodal models and GUI-specialized vision-language models on Automotive-ENV. Experimental results show that current models remain far from reliable in automotive environments, with substantial performance degradation in implicit and safety-critical tasks. Further analysis reveals consistent failure modes in safety-aware reasoning, vehicle-control understanding, and implicit hazard recognition. We release Automotive-ENV, together with tasks, validators, and evaluation tools, to support future research on automotive multimodal systems.

cs.RO↗

arXiv2026

Non-trivial intersecting families for direct products

A family of sets is intersecting if any two of its sets intersect, and non-trivial intersecting if there is not any element which appears in every set of the family. The study of intersection problems (or Erdős-Ko-Rado problems) on families of sets is one of the most important topics in extremal combinatorics. A natural extension of the Erdős-Ko-Rado problem in direct products setting was introduced by Frankl, and developed by Alon and Katona et al. In 2018, Kwan, Sudakov and Vieira [J. Combin. Theory, Ser. A 156 (2018), 44--60] determine the maximum size of a non-trivial intersecting direct product family under the assumption that the sizes of all parts are sufficiently large relative to the set sizes in all parts. Meanwhile, they raised the question of whether this result holds for the size of each part which depends only on its corresponding set size, and the question of characterizing the extremal families. In the present paper, we answere these two questions. Furthermore, the bounds on the size of each part in our result are best possible.1

math.CO↗

arXiv2026

Cracking Mobile Manipulation: From Failure Diagnosis to Agentic Robot System

Recent vision-language-action (VLA) models have shown strong performance on local robot manipulation, yet their success does not readily translate to long-horizon mobile manipulation. In our experiments, a pretrained subtask VLA achieves 80% success rate on atomic manipulation tasks but only 19% when these skills are composed into mobile manipulation. We argue that this gap stems from failures across the execution process rather than from local manipulation capability alone. To uncover these bottlenecks, we systematically diagnose failures in long-horizon mobile manipulation and identify three dominant regimes: task planning, navigation and base positioning, and local manipulation. Based on this diagnosis, we develop failure-guided agentic intervention that augments a fixed subtask VLA with demonstration-based planning guidance, adaptive visual memory, accessibility-aware base adjustment, and local end-effector adjustment. On RoboCasa365, these interventions progressively improve task success from 19% to 47%, with the finalized system reaching 49% on held-out episode seeds, compared with 18% for the baseline. The same design improves performance from 14% to 37% on BEHAVIOR-1K and from 5/20 to 13/20 on real-robot tasks. These results suggest that reliable mobile manipulation can be advanced by diagnosing and intervening on the current bottleneck, rather than uniformly scaling the underlying policy. Our code will be released at https://github.com/PKU-VaLuE-Lab/Cracking-Mobile-Manipulation.git.

cs.RO↗

arXiv2026

MARS: Systematic Multi-Level Acceleration for World Action Models

World Action Models (WAMs) have emerged as a promising alternative for embodied intelligence by jointly predicting future world states and robot actions. However, their heavy visual encoding and iterative generative inference incur substantial latency, limiting their applicability to high-frequency closed-loop control. Existing training-free acceleration methods typically optimize a single stage, yielding limited speedups or degraded control under aggressive reuse. In this work, we systematically profile the inference overhead of WAMs and observe abundant redundancy across Transformer computation, diffusion denoising, and consecutive policy queries. However, the safe removal of such redundancy depends strongly on its impact on action prediction and control sensitivity. Based on these observations, we introduce \textbf{MARS}, a training-free and plug-and-play framework for multi-level WAM acceleration. MARS selectively removes low-impact attention computation, adaptively accelerates diffusion inference according to motion granularity, and speculatively reuses predicted future visual representations across consecutive queries with online verification. Together, these complementary mechanisms provide remarkable acceleration while limiting approximation error. Extensive experiments across LIBERO, LIBERO-PRO, and RoboCasa show that MARS achieves up to $4.00\times$, $3.94\times$, and $2.10\times$ end-to-end speedups, respectively, while maintaining task success rate close to the original policy. Moreover, MARS generalizes across different model backbones and across simulation and real-robot experiments, substantially improving the efficiency--performance trade-off of WAM inference. Our code is available at https://github.com/PKU-VaLuE-Lab/MARS.

cs.CV↗

arXiv2026

Unitary designs from perturbed time evolutions of a chaotic Hamiltonian

Unitary designs provide efficient substitutes for Haar-random unitaries in quantum information processing, randomized measurements, and many-body quantum dynamics. We propose a protocol based on time evolutions of a single chaotic Hamiltonian interleaved with unitary perturbations to generate unitary designs. We derive the frame potential of the resulting ensemble in terms of those of the intermediate ensemble. The intermediate ensemble need not itself be Haar random or exhibit Haar-level frame potentials, provided that its trace moments and frame-potential growth are sufficiently suppressed. The nontrivial Pauli set and the Clifford ensemble are simple examples, whereas ensembles whose cardinality does not grow with the Hilbert-space dimension generally fail. We further establish a frame-potential correspondence between multi-Hamiltonian temporal protocols and one-Hamiltonian protocols interleaved with repeated fixed trace-suppressed perturbations, showing that two perturbations suffice to recover Haar-level frame potentials. For local Hamiltonians, we derive a lower bound on the space--depth control resource, while more generally we establish a universal finite-window bound that constrains the required temporal sampling range. Together, these results identify complementary spatial and temporal resource requirements for perturbation-assisted randomness generation.

quant-ph↗

arXiv2026

DramaAgent: Agentic Storytelling Video Generation

Recent diffusion and autoregressive models have substantially improved text-to-video generation, yet producing coherent long-form story videos with consistent characters and aligned audio remains challenging. Existing methods often suffer from narrative drift, unstable character identity, weak cross-scene continuity, and audio-visual mismatch over extended sequences. We propose DramaAgent, a hierarchical, agentic, and model-agnostic framework for long-form text-to-video-and-audio generation. Rather than improving the underlying video backbone itself, DramaAgent introduces an upper-level control layer that decomposes generation into story planning, persistent character conditioning, scene-wise synthesis, and reflection-guided targeted repair. The framework maintains reusable story and character states across scenes, diagnoses failures such as identity drift, missing scene semantics, temporal discontinuity, and cross-modal mismatch, and repairs problematic clips in a stage-specific manner. Experiments across multiple video generation backbones show that DramaAgent improves long-horizon coherence, character consistency, narrative fidelity, and scene-level audio-visual consistency over direct generation and strong baselines. These results suggest that hierarchical agentic control is a practical direction for controllable long-form audiovisual generation. Code: https://github.com/AIGeeksGroup/DramaAgent. Website: https://aigeeksgroup.github.io/DramaAgent.

cs.CV↗

arXiv2026

Area-Law Entanglement in Quantum Chaotic System

Entanglement entropy is a fundamental diagnostic for quantum chaos, typically exhibiting volume-law scaling in highly excited eigenstates of chaotic many-body systems. In this work, we present a striking counterexample: a Floquet-driven quantum many-body system with Rydberg-like blockade that, despite being fully chaotic as indicated by its Wigner-Dyson level statistics and local thermalization, exhibits a strict area-law entanglement entropy. Specifically, the entanglement entropy of every Floquet eigenstate is bounded by $\ln2$, independent of system size. We trace this anomaly to the specific Hilbert space structure imposed by the blockades, which restricts the Schmidt rank across a bipartition. Furthermore, we generalize this discovery by establishing a duality between constrained many-body Hamiltonians and single-particle quantum walks on median graphs, and we outline a general procedure for constructing systems with an entanglement entropy bounded by a predetermined constant. Our results demonstrate that entanglement entropy alone is an insufficient diagnostic of many-body quantum chaos and highlight the profound impact of Hilbert space geometry on quantum dynamics and thermalization.

quant-ph↗

arXiv2026

Von Neumann's Two Quantum Entropies

Von Neumann proposed two quantum entropies, one in 1927 and the other in 1929. He intended to define thermodynamic entropy for quantum states in both cases. The 1927 entropy is the well-known von Neumann entropy, which is zero for all pure states. The 1929 entropy, which is nonzero for almost all pure states, has largely been ignored. After reviewing the history and properties of both entropies, we argue that the 1927 entropy is not thermodynamic entropy and should be used simply as a measure of entanglement, whereas the 1929 entropy is the true thermodynamic entropy for quantum states.

cond-mat.stat-mech↗

arXiv2026

Spark-to-Paper: End-to-End Research Paper Generation as a Composable Skill

Turning a research idea into a complete paper requires more than text generation: the system must retrieve literature, design and execute experiments, revise claims according to evidence, produce publication-ready figures, and maintain consistency across a long generation process. We present Spark-to-Paper, an end-to-end research paper generation system implemented as thirteen composable skills inside an existing coding assistant, without requiring a separate agent platform or orchestration service. Spark-to-Paper separates model-based judgment from deterministic operations that can be directly executed and checked. It further separates experiment planning from reporting, so that required evidence is specified before results are observed and manuscript claims are revised according to measured outcomes. To improve reliability over long research trajectories, the system combines deterministic integrity checks with self-critique and bounds a failure mode we call the Self-Refutation Loop, in which repeated experiments continue to reject the original research objective. Spark-to-Paper also produces editable vector figures through programmatic plotting for experimental results and code-based reconstruction for generated method diagrams. Across eight controlled research topics, Spark-to-Paper achieves 99.5% citation validity and 96.4% figure editability. A controlled ablation increases fabrication detection from 14% for a single-pass draft to 92% with the full integrity and review stack, while adversarial review achieves 74% precision. The full system uses 11.9M tokens, costs $8.1 per manuscript, and requires 3.2 hours on average. These results show that end-to-end research paper generation can be implemented as a lightweight, composable workflow inside existing coding assistants while keeping experimental evidence central to how claims are accepted, revised, or abandoned.

cs.CL↗

arXiv2026

Scaling GUI Agents with Visual State Transitions

We introduce State Transition Pretraining (STP) as a new scaling axis for GUI agents. During the STP stage, we continually pretrain a unified multimodal model on visual state transitions by jointly optimizing inverse dynamics (predicting actions from state changes) and forward dynamics (predicting next states from current states and actions). This optimization equips the model with better action-grounded visual representations and an internal world model of GUI dynamics. When subsequently fine-tuned on trajectories with task instructions, our STP-trained models consistently outperform baselines trained solely via direct trajectory fine-tuning across agent benchmarks in both desktop and mobile GUI scenarios (AgentNetBench, AndroidControl, and GUIOdyssey). Further empirical studies show that joint dynamics optimization yields stable improvements over single-objective training, and downstream performance scales steadily with the volume of transition data.

cs.AI↗

arXiv2026

PaperJury: Due-Process Review for Bounded LaTeX Revision

Pre-submission hardening of human-authored LaTeX computer science papers differs from drafting assistance because it requires adversarial whole-paper review, explicit no-fix outcomes, and bounded artifact-safe revision. Existing writing assistants, critique generators, and judge-centered loops lack durable issue identity across rounds, deterministic routing from critique to adjudication, and manuscript control that can reject invalid concerns or defer author-dependent ones. We present PaperJury, a closed-loop review-verdict-revise-verify system built on a deterministic-versus-semantic split: deterministic orchestration manages decomposition, a frozen claim spine, a durable ledger, routing, stopping, and exact-once patch application, while semantic agents are limited to bounded review, judgment, and repair. PaperJury combines bounded holistic review, contestability-based routing, a due-process trial, and risk-proportional guard chains for anchor-bounded edits, yielding terminal outcomes of invalid-drop, valid-fixable, and author-required. In a two-arm expert-review evaluation on held-out Vision, natural language processing, and machine learning papers against four baselines, we assess issue quality, verdict and routing quality, edit safety, convergence behavior, and cost, supporting the thesis that load-bearing safety and completion logic should reside in deterministic orchestration rather than model discretion. PaperJury is available at https://github.com/u7079256/paperjury.

cs.CL↗

arXiv2026

Explicit Representation Alignment for Multimodal Sentiment Analysis

Multimodal affective analysis aims to understand human sentiment and emotion by jointly modeling heterogeneous modalities such as text and images. However, multimodal models often fail to consistently outperform strong text-only baselines, with performance varying significantly across fusion strategies. In this work, we identify representation misalignment between independently pretrained modality encoders as a key bottleneck for effective multimodal learning, and show through controlled experiments that alignment prior to fusion is often more important than fusion complexity. To address this issue, we propose a unified multimodal affective analysis framework that leverages vision-language models (VLMs) to convert visual content into structured textual descriptions, projecting heterogeneous modalities into a shared linguistic space and enabling interpretable text-centric reasoning. To further improve robustness, we introduce a hybrid learning strategy that combines semantic token selection with a batch-level uniformity regularization objective, encouraging a more dispersed and stable global feature space while mitigating noise introduced by VLM-generated descriptions. Experiments on multiple multimodal sentiment and emotion benchmarks show that our method consistently outperforms strong unimodal and multimodal baselines, achieving state-of-the-art performance. Our analysis further highlights the critical role of representation alignment in multimodal affective learning.

cs.CL↗

arXiv2026

Engineered dissipation for faster adiabatic state preparation

Adiabatic state preparation is often slowed by nonadiabatic leakage near small spectral gaps. We propose an engineered dissipative protocol that uses a filtered reservoir to induce predominantly downward transitions in the instantaneous energy eigenbasis while leaving the instantaneous ground state dark. The leaked population generated by nonadiabatic driving is therefore continuously relaxed back toward the low-energy sector. An effective avoided-crossing analysis shows that in the regime where the engineered relaxation strength is much larger than the minimum gap, the runtime scaling can improve from the closed-system behavior $\mathcal{O}(Δ^{-2})$ to $\mathcal{O}(Δ^{-1})$ Finite-temperature upward transitions introduce a thermal error floor, but the enhancement survives when this heating rate remains below the target error tolerance. Numerical results show improved ground-state preparation over closed-system annealing. We also discuss a possible superconducting-circuit implementation using structured bosonic reservoirs.

quant-ph↗

arXiv2026

Theory of Quantum Phase Space: Foundations and Applications

This article provides a concise review of quantum phase space theory, beginning with its foundational principles and the properties of standard quantum quasi-probability distributions, specifically the Wigner, Husimi Q, and Glauber--Sudarshan P functions. We discuss the intrinsic limitations of these distributions, such as the appearance of negative values and phase-space blurring. A significant portion of this review highlights recent theoretical developments, particularly the quantum Wannier basis. This approach establishes a unitary mapping between the Hilbert space and a discretized phase space, yielding a genuine probability distribution in phase space and thereby providing a basis-dependent entropy for pure quantum states. Furthermore, we examine Bourgain's nonperiodic basis as a theoretical framework to circumvent the constraints imposed by the Balian--Low theorem. These developments provide practical tools for numerical studies based on the quantum Wannier basis, as well as conceptual benchmarks for understanding the localization limits of orthonormal phase-space representations.

quant-ph↗

arXiv2026

PaperVoyager : Building Interactive Web with Visual Language Models

Recent advances in visual language models have enabled autonomous agents for complex reasoning, tool use, and document understanding. However, existing document agents mainly transform papers into static artifacts such as summaries, webpages, or slides, which are insufficient for technical papers involving dynamic mechanisms and state transitions. In this work, we propose a Paper-to-Interactive-System Agent that converts research papers into executable interactive web systems. Given a PDF paper, the agent performs end-to-end processing without human intervention, including paper understanding, system modeling, and interactive webpage synthesis, enabling users to manipulate inputs and observe dynamic behaviors. To evaluate this task, we introduce a benchmark of 19 research papers paired with expert-built interactive systems as ground truth. We further propose PaperVoyager, a structured generation framework that explicitly models mechanisms and interaction logic during synthesis. Experiments show that PaperVoyager significantly improves the quality of generated interactive systems, offering a new paradigm for interactive scientific paper understanding.

cs.CL↗

arXiv2026

I-WebGenBench : Evaluating Interactivity in LLM-Generated Scientific Web Applications

cs.CL↗