arXiv subjects

Yuxuan Zhang

Publications and source records attributed to Yuxuan Zhang.

At least 19 recordsLinked to original sources

arXiv2026

Where, What, Why, and Importance: Structured Defect Grounding for Text-to-Image Feedback

Despite generating increasingly photorealistic images, text-to-image (T2I) models still exhibit localized, subtle, and structurally complex failures. Diagnosing these failures requires instance-level feedback that answers where a defect occurs, what type it is, why it is defective, and its importance to overall image quality. While recent dense-feedback methods move beyond scalar supervision, their heatmap-centric representations still formulate diagnosis as pixel-field regression, making it difficult to localize variable-cardinality defects and bind semantic reasons to individual failures. To address this representation bottleneck, we propose Structured Defect Grounding (SDG), which casts T2I diagnosis as structured set prediction by modeling each defect as a (location, type, reason, importance) tuple. To make this formulation trainable and measurable, we introduce SDG-30K, a 30K-image dataset with box-grounded annotations across four modern T2I generators, together with a dedicated evaluation protocol, SDG-Eval. Building on this structured representation, we further present a diagnosis-to-alignment framework in which a Vision-Language Model (VLM) serves as the SDG detector, and BoxFlow-GRPO converts predicted defect sets into box-derived, importance-weighted spatial rewards for diffusion model alignment. Extensive experiments show that our SDG detector outperforms leading proprietary VLMs on structured defect grounding, while SDG-guided rewards consistently improve T2I alignment and support localized image refinement. These results establish SDG as a unified, instance-level interface for diagnosing, evaluating, and enhancing modern generative models.

cs.CV↗

arXiv2026

Community Search under Member-Specific Tolerances

We study community search when each vertex has its own tolerance for differences in position. For positions such as opinions or homes, a latitude-consistent community (LCC) contains the query vertex, is connected, has minimum degree at least k, and bounds its diameter by the smallest member tolerance. Motivated by social judgment theory, the bound exactly characterizes persistent acceptance of every interaction in a closed group under bounded-confidence dynamics with arbitrary positive weights. It also prevents rejection under social-judgment dynamics, whereas natural relaxations preserve neither guarantee. Finding a maximum LCC is NP-hard even in the plane, and with unbounded dimension, deciding whether an LCC exists is NP-complete for every fixed k >= 2. LatSweep turns the coupled bound into a sweep over the smallest tolerance, yielding two feasible heuristics, a (1,2)-bicriteria approximation and an exact branch-and-bound algorithm. Across eight graphs with up to 1.8 billion edges, the exact algorithm proves optimality on all but 9 of 3,700 efficiency queries and usually certifies absence within microseconds when no LCC exists. Its branch-and-bound solver proves more queries than a reimplemented maximum-(k,r)-core algorithm, a mixed-integer program and CP-SAT. In simulations, uneven influence disrupts initially accepted interactions in pairwise groups. Outside ties invalidate the guarantee for many real groups. On two location-based networks, tolerance-agnostic groups often place another member's home beyond a member's estimated travel radius; comparisons with spatial communities of the same size detect no co-visit loss from enforcing every member's tolerance.

cs.SI↗

arXiv2026

LongBanana: An Expert-Verified Benchmark for Long-Context Multi-Reference Image Synthesis

We introduce LongBanana, to our knowledge the first expert-designed and expert-verified benchmark for long-context multi-reference image synthesis in the age of agent-driven image generation. The task requires retrieving dispersed visual evidence, rejecting distractors, binding entities and relations, and composing one coherent final image from multiple references. LongBanana contains 90 carefully designed tasks organized by a 3-by-3 taxonomy; each task includes about 32 reference images, 20 constraints, and 50 evaluation items, yielding 4,670 expert-specified atomic evaluation items in total. Its graph-grounded evaluation design instantiates each task with reference images and an entity-relation graph, then converts node, edge, and whole-graph obligations into checklist items for Content Fidelity, Relational Binding, and Global Constraint Satisfaction. Reviewed expert labels include a 15% double-labeled subset with 89.4% exact agreement and 0.83 quadratic weighted kappa. We further introduce LongBanana-Gen, an agent-harness framework for long-horizon image generation; experiments show that minimal agent harnesses improve over direct generation, while a ReAct-Plan-Exec-Reflect (RPR) workflow yields harness-dependent gains by strengthening evidence grounding and repair. Together, LongBanana provides an auditable testbed for measuring dense visual evidence use and a shared target for developing stronger agent-driven image-generation systems. Code and data are available at the project page: https://long-banana.com/.

cs.CV↗

arXiv2026

Single-shot ground-state preparation via strongly coupled baths

Preparing ground and thermal states of quantum many-body systems is a central task in quantum simulation. Dissipative protocols have emerged as a promising solution for both near-term and fault-tolerant hardware, but most constructions rely on weak system-bath coupling, leading to slow cooling. Here, we construct cooling protocols at strong system-bath coupling, and demonstrate accurate state preparation within even a single cooling cycle. Our protocols combine dynamically ramped ancilla energy levels with time-dependent couplings to optimize the trade-off between cooling error and physical time. For a general class of free-fermion systems, we reduce the dynamics exactly to a two-level scattering problem: ground-state cooling becomes a control problem of maximizing resonant excitation transfer while suppressing non-adiabatic heating. We then construct a family of ``Möbius-Ingham'' protocols, which reach optimal error scaling for gapless spectra and almost-exponential suppression of the leading heating error for gapped spectra, as a function of the cycle period. In our free-fermion benchmarks, Möbius-Ingham cooling lies on the Pareto frontier between steady-state error and cooling time among the controls tested. We further demonstrate its effectiveness in interacting Ising and $t$-$V$ chains and the two-dimensional Fermi-Hubbard model. Finally, we extend our construction to Gibbs state preparation, showing how free-fermion thermal states can be prepared in a single cycle.

quant-ph↗

arXiv2026

FastAFD: When Does Attention-FFN Disaggregation Pay on Rack-Scale NVLink?

Attention-FFN disaggregation (AFD) separates attention from the experts of a mixture-of-experts (MoE) model and has been reported to raise decode throughput over colocated serving. Whether this advantage holds on rack-scale NVLink systems such as GB200 NVL72 remains unclear: prior results were obtained on other hardware, against narrower colocated baselines, or with analytical models, whereas production engines serve large MoE models on these systems with wide expert parallelism (wide EP). We build and tune FastAFD, an AFD engine for GB200 NVL72 that overlaps the MoE of one microbatch with the attention of the other, and compare it against the best vLLM wide EP setting. Experiments show that AFD has a throughput advantage of up to 1.17x over wide EP on Qwen3 and 1.16x on Qwen3.5, and matches it on MiniMax-M2.5. We derive an analytical model that predicts AFD throughput within 3% using only profiles of colocated vLLM, telling whether switching to AFD would pay off before deploying it. The model yields a concise insight: AFD's speedup factors into a batch gain (larger batches, since attention GPUs hold no expert weights), an overlap gain (shorter steps, since the MoE runs on the expert GPUs behind attention), and a disaggregation cost (the GPUs given to experts). AFD has the advantage when the two gains outweigh the cost. This view explains which models benefit from AFD, and shows that speculative decoding helps light-expert models most, where the enlarged overlap gain outweighs the larger disaggregation cost and adds a further 8% to AFD's advantage over wide EP. Looking ahead, serving the experts on accelerators that keep their weights in on-chip SRAM, such as the Groq LPX accelerators paired with Vera Rubin NVL72, would remove most of the disaggregation cost, raising the ideal speedup to 2.48x with speculative decoding. FastAFD is open source at https://github.com/hao-ai-lab/FastAFD.

cs.DC↗

arXiv2026

Discovering Semantic Structures in Psychological Scales: A Response-Free Front-End to Efficient Simplification

Psychological scale refinement traditionally relies on response-based methods such as factor analysis, item response theory, and network psychometrics to uncover latent constructs and optimize item composition. Although rigorous, these approaches require large samples and may be constrained by data availability and cross-cultural comparability. Recent advances in natural language processing suggest that the wording of scale items contains construct-relevant semantic organization, offering a complementary response-free perspective. We introduce a topic-modeling framework that operationalizes item-level semantic structure for systematic scale simplification. Items are encoded using contextual sentence embeddings and grouped through density-based clustering to identify semantic topics without predefining their number. Within each cluster, class-based term weighting identifies representative keywords that summarize each topic and support an interpretable semantic pre-structure of the item pool. Semantically related clusters are hierarchically merged, and representative items are selected using structure-aware membership criteria within an integrated reduction pipeline. We evaluated the framework on three widely used instruments: DASS, IPIP, and EPOCH, covering negative affect, personality traits, and well-being. Results showed that the identified semantic topics often formed factor-like groupings aligned with established constructs. Representative items selected through semantic membership probabilities reduced item counts by 60.5% on average while maintaining acceptable psychometric properties and preserving factor correlation patterns. Our framework formalizes semantic topic structure as an inspectable front-end for scale construction and simplification. To facilitate adoption, we provide a visualization-supported tool enabling one-click semantic analysis and structured simplification.

cs.CL↗

arXiv2026

DoAtlas-2: A Foundation for Self-Evolving Causal Biomedical Discovery

We introduce DoAtlas-2, a foundation for self-evolving causal biomedical discovery that organizes knowledge around causal mechanisms and advances through external evidence from human populations. DoAtlas-2 integrates 771 research resources covering more than 720,000 participants in 48 countries, from longitudinal clinical phenotypes, medical imaging, and continuous physiological signals to eight molecular layers, together with an evidence network of approximately 4.7 million literature-derived records over 93,566 concepts and 149,383 candidate causal relations. DoAtlas-2 autonomously formulates research questions from evidence gaps and unresolved mechanisms, prespecifies their causal designs, and generates validated analyses. Supporting, challenging, and unresolved results continuously revise mechanistic interpretations, the causal evidence state, and the discovery frontier, so that DoAtlas-2 self-evolves within a closed loop of hypothesis generation, empirical testing, and renewed discovery. DoAtlas-2 has systematically evaluated 2,031 research questions. In the Human Phenotype Project (HPP), it formulated 4,014 candidate pathway questions across vascular, early-glycemic, and hepatic-metabolic systems, and screening of the first 1,079 yielded statistical support for 756. Representative studies identify blood pressure as a convergence node linking adiposity, hepatic, and lipid phenotypes to vascular outcomes, and show that an adiposity-inflammation-blood-pressure pathway is largely attenuated by joint adjustment for body mass index (BMI) and smoking. The discovered vascular network constitutes a completely interpretable predictive foundation, admitting exact attribution of every prediction and closed-form mediation effects. DoAtlas-2 thereby unifies causal mechanism discovery, population-evidence testing, and interpretable prediction within one continuously evolving foundation.

cs.AI↗

arXiv2026

WebHarbor: Docking Real Websites for Evolving GUI Agent Environments

GUI web agents require realistic and interactive environments that remain stable during both evaluation and training. Live website evaluation suffers from changing content, access restrictions, and limited coverage of authenticated workflows, while existing synthetic environments for training often lack functional depth and visual fidelity. We propose WebHarbor, a suite of 153 executable and dockerized website environments that faithfully reconstruct the visual interfaces, content, and functional workflows of real-world web services under deterministic control. We build WebHarbor through task-driven human-agent collaboration, where environments evolve as new tasks expand their functional scope, coding agents implement the required changes, and human contributors audit visual fidelity, functional correctness, and task validity. The resulting suite comprises 3,518 tasks spanning multi-step and post-authentication workflows, where existing works often fail to cover. Beyond reliable evaluation, WebHarbor provides high-fidelity synthetic web environments for cost-efficient reinforcement learning (RL). We demonstrate that a Qwen 4B-level GUI agent trained with RL in WebHarbor generalizes to unseen tasks and performs competitively with leading open-source GUI web agents across four benchmarks. The code and data are publicly available at https://github.com/aiming-lab/WebHarbor.

cs.AI↗

arXiv2026

EvoMemState: Evolving Memory State for Personal Agents

Long-term memory enables large language model (LLM) agents to provide continuous personalized assistance by retaining user information across interactions. Existing systems support this process through fact extraction, organization, updating, and retrieval, but largely maintain memories as independent facts. This fact-centric paradigm may overlook how new facts update, constrain, or conflict with earlier ones, while leaving their applicability, role, and priority for the current request implicit. In this paper, we introduce EvoMemState, a state-centric framework that augments persistent facts with two complementary states: Relational States that capture how facts remain valid and relate to one another as memory evolves, and a Reasoning State that determines how these facts should be applied to the current request. Specifically, EvoMemState consists of three key components: (i) Ontology-Guided Memory State Induction extracts Persistent Facts and constructs Relational States that maintain their validity, scope, and evolving relations. (ii) RL-Based State Evolution internalizes memory-use experience into a policy through reinforcement learning to generate a request-specific Reasoning State. (iii) State-Aware Reasoning then combines the relevant facts with their Relational States and the generated Reasoning State to guide the personal agent. Extensive experiments on long-term dialogue and memory-use benchmarks demonstrate consistent improvements over existing memory systems in retrieval completeness and memory-grounded response accuracy. Our code and implementation details are available at https://github.com/XMUDeepLIT/EvoMemState.git.

cs.IR↗

arXiv2026

ReGE: Stronger Agents through Experience Synthesis

Repeated attempts at interactive tasks uncover facts, useful operations, and failures that can inform subsequent execution. We introduce ReGE, a framework for turning this recorded experience into guidance for a fresh attempt. Its main implementation, ReGE-Syn, uses a single coach call to jointly synthesize bounded execution records and available artifacts into self-contained guidance, connecting task-specific findings with proposed actions and relevant pitfalls. A student starts from the task's initial environment state and adapts this guidance to current observations. Source generation, synthesis, and student execution use the same base model, without parameter updates or a stronger teacher. Across four benchmarks, ReGE improves accuracy by an average of 9.9 percentage points for GPT and 8.8 percentage points for Claude compared with unguided execution. On matched BrowseComp-Plus subsets, ReGE-Syn reduces guidance-construction tokens by 98.2\% on GPT and 95.5\% on Claude relative to adapted AggAgent, with observed success. Combined construction and execution tokens decrease by 15.7\% and 35.4\%, respectively, counting input plus output and excluding source generation. BrowseComp-Plus ablations further associate access to process information with fewer student steps and show that joint synthesis produces substantially shorter guidance than Direct history or Independent summaries, with observed success. These results demonstrate how agents can use their own recorded experience to improve subsequent task completion, and suggest a low-cost way to consolidate the execution records of parallel agent systems such as agent swarms.

cs.AI↗

arXiv2026

OPIS: An Input-Grounded Benchmark for Multi-Object Memory in Video World Models

Video world models must preserve the visual state of the world over time, but existing evaluation protocols often rely on generated histories, video reference, or selected revisit viewpoints that can confound the assessment of a model's true memory capability. To address this, we introduce OPIS, an input-grounded benchmark that strictly anchors the assessment to a fixed set of object instances from the initial observation for evaluating multi-object memory in video world models. The OPIS dataset comprises 500 cases across real-world, embodied-robotic, and game-world domains, providing dense object-level annotations for 12,672 rigid, articulated, and deformable instances. Our object-centric evaluator combines association and explicit visibility reasoning to hierarchically measure Object (O) Presence (P), Identity (I), and Structure (S), utilizing static or dynamic evaluation tracks based on object kinematics. Across eight image-to-video or camera-conditioned world models, our proposed OPIS scores range from 48.65 to 56.01. As the reference inventory grows from less than 20 to more than 40 objects, the Presence, Identity, and Structure scores show an overall decline, with the average Identity score falling from 40.22 to 23.11. The results demonstrate that preserving the particular object instances in the input is considerably harder than generating plausible visual elements.

cs.CV↗

arXiv2026

StopLoss: Teaching Agents to Predict Failure and Stop Early

Coding agents keep working on runs that will fail verification, and asking a frontier model to watch for this is expensive in its own right. We introduce StopLoss, a small external monitor that learns failure recognition from completed trajectories and their verifier outcomes. The name is inspired by the stop-loss order in finance: StopLoss ends a run early when the run is likely to fail, which saves resources. Final outcomes label earlier local windows, so a fine-tuned language model can score the most recent tool steps in a single forward pass, and a threshold calibrated on successful runs turns that score into an abort decision without changing the executor. Trained on GPT-family SWE-rebench trajectories and evaluated on SWE-bench Verified and a SWE-bench Pro subset with GPT and Claude executors, StopLoss is compared against six general-purpose judges reading identical windows. Outcome supervision lifts failure prediction from near chance to a useful signal at the first check, at one to two orders of magnitude lower cost per check than any tested judge. This translates into savings: relative to completing every task, StopLoss lowers cost per retained solve by 4.5\% on Verified while preserving 99\% of successful runs, and by up to 31\% on the costlier Pro subset, reaching 48\% with a Qwen3-8B backbone, and in all four settings its cost per retained solve is below that of every judge policy. Outcome-trained small monitors can make failure recognition cheap enough for early stopping to pay for itself.

cs.AI↗

arXiv2026

Quantum computational advantage in random-circuit sampling on IBM superconducting quantum computers

We report forward random-circuit sampling (RCS) on the 120-qubit Nighthawk r2 superconducting processor (ibm_phoenix) with square-lattice connectivity, using 61 qubits, native CZ gates, and the standard cloud execution stack with no benchmark-specific calibration. Two independent fidelity estimators---mirror benchmarking and three- and four-patch cross-entropy benchmarking (XEB)---agree with each other at every measured depth, the mirror from 4 to 40 cycles and the patched estimators from 20 to 40 cycles, across more than two orders of magnitude of fidelity decay, and exceed the first-generation Nighthawk r1 device by more than an order of magnitude at fixed depth. The 36-cycle circuits sit at the depth where tensor-network contraction cost saturates at system size: a contraction-cost estimator validated against the published Sycamore and Zuchongzhi networks places the single-amplitude cost at $\sim$$10^{22}$ complex operations. At $F_{\mathrm{XEB}}(36)=2.3\times10^{-3}$ under favorable memory assumptions this implies $1.2\times10^{27}$ machine operations within the bounded-fidelity rejection-sampling model --- more than a century of runtime on the Frontier supercomputer --- to collect a $10^{6}$-sample ensemble, which takes only 19\,s on Nighthawk r2. To our knowledge, this is the first demonstration of quantum advantage for a vanilla random-circuit sampling on a commercially and broadly accessible quantum processor that most non-expert quantum computer users can easily replicate.

quant-ph↗

arXiv2026

PremiseCheck: Efficient Fact-Checking of Human and AI Peer Reviews

Peer review is under strain from the growth of submissions, and review comments written by humans or AI can misstate what a paper says. We introduce ReviewFact, a benchmark that labels such factual errors in review comments under one definition across NLP, machine learning, and the natural sciences, and that measures cost and latency alongside accuracy. A frontier LLM checks comments well but must read the whole paper for each check, while cheap decision models, which score whether a statement holds in a text, fall far short on whole comments that mix several implicit assertions. PremiseCheck is a framework that combines the two. An LLM rewrites each comment as explicit premises without reading the paper, and a decision model checks each premise against the paper. When each comment is checked on its own, PremiseCheck reaches accuracy comparable to GPT-5.5 across all three fields at about half its cost, and on NLP reviews at a quarter of its cost and less than half its latency.

cs.CL↗

arXiv2026

Speculative Sampling in Interactive Video Generation

Interactive video generation commonly relies on small bidirectional chunks to reduce response latency. Yet our preliminary experiments show that larger chunks improve video quality and better utilize GPU parallelism, allowing more latents to be sampled with only a modest increase in sampling time. Motivated by these observations, we propose Speculative Sampling, a strategy that combines the modeling power of large bidirectional windows with the low-latency interaction of small ones. It expands the modeling window by appending noise latents as padding and caches the resulting predictions for reuse when the condition remains unchanged. Predictions are delivered incrementally, while the uncommitted future remains replaceable. Our latency analysis shows that this removes the waiting-time lower bound imposed by chunk boundaries, shifting the optimization focus from reducing chunk size to accelerating per-chunk sampling. Bidirectional window size thus becomes a decision variable for jointly optimizing visual quality and sampling time, rather than a necessary constraint on low-latency interaction. This opens a path to serving large bidirectional models in interactive streaming systems without complex streaming post-training.

cs.CV↗

arXiv2026

LEV: Thinking Fast and Deep with Looped Decision Models

Decision models answer typed questions about an input with one probability per option, computed in a single forward pass without generating text. Like the fast System 1 of dual-process theory, they answer at once, but their depth is fixed at training time, so they spend the same computation on easy and hard questions. Models that reason in generated text can think longer, but they pay for every extra step in decoded tokens. We present LEV, a decision model that thinks deeper by rerunning its layers in a loop instead of generating tokens. LEV builds on a language model pretrained as a loop, reads the options after every loop step and trains all steps jointly, so one checkpoint can answer at any depth. On three looped backbones, accuracy rises with depth on the same weights, mostly on hard questions, and a 2.6B LEV goes from 63.2% at depth one to 76.2% at depth four on JevBench. On-demand depth, set per question by a small stopper with a threshold fitted on unlabelled requests, saves 15% to 24% of the computation at a cost of under one accuracy point, level with or slightly above a fixed mix of depths. On a locked test set, the 1.4B and 3B models are 5.4 to 8.8 points more accurate than dense (non-looped) models and cascades of the same latency below 50 ms.

cs.AI↗

arXiv2026

RewardHarness: Self-Evolving Agentic Post-Training

Aligning generative models with human preferences remains a central challenge. While RLHF is widely adopted, it relies on large-scale preference data. This limitation is particularly evident in evaluating instruction-guided image edits, even though humans can often learn such preferences from only a few examples. We present RewardHarness, a training-free, self-evolving agentic reward framework that reframes reward modeling as context evolution rather than weight optimization. Instead of learning from large-scale annotations, RewardHarness aligns with human preferences by iteratively evolving a library of tools and skills from as few as 100 preference demonstrations. Given a source image, an edited image, and an editing instruction, an orchestrator agent selects the most relevant subset of tools and skills from the maintained library, and sub-agents invoke them to construct a reasoning chain that produces a reward score. By comparing predicted scores with ground-truth preferences and analyzing successes and failures in the reasoning process, the orchestrator automatically refines its library of tools and skills without additional human annotation. Using only 0.05% of the EditReward preference data, RewardHarness achieves 47.4% average accuracy on EditReward-Bench K=2/3/4 and GenAI-Bench, surpassing GPT-5 by 5.3 points. When used as a reward signal in FlowGRPO, RL-tuned models achieve 3.52 on ImgEdit-Bench. Code and model are available at the project page: https://rewardharness.com/

cs.AI↗

arXiv2026

ModularRSI: Modular and Generalizable Recursive Harness Self-Improvement

Recent work extends recursive self-improvement (RSI) to agent harnesses for long-horizon coding and terminal tasks, enabling agents to improve execution mechanisms from experience. However, generalizable harness RSI remains challenging. First, evolving harnesses on evaluation benchmarks or their subsets makes it difficult to distinguish reusable improvements from benchmark-specific adaptation. Second, single-trajectory updates can conflate systematic harness deficiencies with instance-specific reasoning and solution details, producing modifications that transfer poorly to unseen tasks. Third, localizing recurring behavioral deficiencies within monolithic harnesses is difficult, while whole-harness optimization can entangle unrelated mechanisms and complicate attribution and validation. We propose ModularRSI, a benchmark-disjoint, contrastive, and modular framework for generalizable harness evolution. ModularRSI contrasts successful and failed trajectories for the same task and aggregates evidence across tasks to identify recurring behavioral deficiencies. It decomposes the evolvable harness into five functional modules: Agent Loop, Tool Use, Observation Management, Context Management, and Task Completion Detection. Each module evolves independently within a restricted modification scope, followed by an integration stage that combines the evolved modules into a unified harness and resolves potential conflicts. To support benchmark-disjoint evolution, we curate 2,000 executable evolution tasks from external sources that are disjoint from downstream evaluation benchmarks. Experiments on TB2.0 and SWE-Bench Verified show consistent improvements on unseen in-domain and cross-domain tasks, with the evolved harness also transferring across different foundation models.

cs.CL↗