Searcharxiv⌕ Search

arXiv subjects

Ziyang Wang

Publications and source records attributed to Ziyang Wang.

At least 19 recordsLinked to original sources

Protected Unbiased Sparsification: A Sharp Loewner Curve

Random sparsification replaces a probability vector by a random vector with few nonzero entries. We ask for unbiasedness, exact preservation of prescribed linear statistics, and a covariance guarantee for every downstream linear response. Let $p$ have $k$ positive coordinates, let $A$ specify the protected statistics, and let $d\le k-2$ be their effective rank. For every $d+1\le m\le k$, we construct an unbiased random probability vector $Q$ with at most $m$ nonzero coordinates, $AQ=Ap$ on every realization, and $\operatorname{Cov}(Q)\preceq h C_{p\mid A}$, where $h=(k-m)/((m-d)(k-d-1))$ and $C_{p\mid A}$ is the categorical covariance after optimal linear adjustment for the protected statistics. This coefficient is sharp uniformly over instances, even among biased estimators when error is measured by its second moment. A residual covariance bridge extends the concurrent unprotected curve to arbitrary pathwise linear protection. We also characterize instance-optimal constants for partitions with uniform weights within classes and derive bounds for peaked weights. For rational inputs, an exact Las Vegas sampler attains the curve within a factor $1+ε$ in expected polynomial time. A separate fast block-pivotal sampler preserves block masses and is no worse in covariance order than independent sampling within blocks. Experiments illustrate the theory on synthetic weights and language-model attention. In Qwen2.5-7B-Instruct decoding, the fast sampler preserves block masses, while comparisons with VarOpt remain statistically inconclusive and top-$m$ truncation scores highest on RULER.

cs.LG↗

Searching Videos as Trees: Self-Correcting Agents for Grounded Long Video QA

Grounded long-video question answering (Grounded LVQA) requires answering a question about a long video while localizing the short evidence interval that supports the answer. Recent agentic methods frame this task as multi-turn exploration with a single crop_video(start, end) action, which supports coarse-to-fine narrowing but provides no primitive for fine-to-coarse backtracking. As a result, these agents typically converge prematurely and cannot recover from an early mistake. We propose VideoTreeSearch (VTS), a framework that casts grounded LVQA as iterative self-correcting search over an adaptive temporal tree. VTS constructs a non-uniform tree through adaptive temporal segmentation so that each node corresponds to a semantically coherent segment, and trains an agent to navigate the tree through four discrete operations: zoom_in, zoom_out, shift, and answer. These operations expose backtracking and recovery as explicit, learnable primitives rather than implicit behaviors. To train this navigation, we introduce a trajectory synthesis pipeline that produces multi-step paths through the tree, including deliberate detours into incorrect branches followed by recovery. We use these trajectories for supervised fine-tuning, followed by reinforcement learning with grounding and answer-accuracy rewards. On three Grounded LVQA benchmarks (CG-Bench, Haystack-LVBench, Haystack-Ego4D), VTS outperforms the strongest prior multi-turn cropping agents by +12.5 mIoU on CG-Bench and +7.4 T-F1 on Haystack-Ego4D. The learned policy also transfers to general long-video QA, surpassing all prior multi-turn cropping agents on Video-MME, MLVU, and LVBench by up to +7.1 accuracy points. Ablations confirm that self-correcting hierarchical search is the central mechanism behind these gains: removing either adaptive descent or explicit backtracking substantially degrades performance.

cs.CV↗

Structured Channel Network Pruning with Bi-Sparse Matrix Factorization

Structured channel pruning for convolutional neural networks (CNNs) plays an important role in pattern recognition and deep learning. Nevertheless, existing methods usually calculate channel importance scores based on the magnitudes of individual filters, making it difficult to distinguish shared redundancy from individual channel contributions. To tackle this limitation, we propose a novel structured channel pruning method termed bi-sparse matrix factorization (BSMF). Specifically, BSMF characterizes the convolutional weight matrix as output-channel coefficients over a shared basis with row-wise $\ell_{2,p}$-norm and column-wise $\ell_{2,q}$-norm regularization terms, respectively. Note that $p,q\in[0,1)$, which, to our best knowledge, has not been explored in the literature. To solve the resulting non-convex optimization problem, we develop an efficient update scheme built upon the alternating direction method of multipliers (ADMM), in which all subproblems admit closed-form solutions. After that, a compact CNN is obtained by reconstructing the convolutional filters from the learned factors and discarding the redundant channels. Extensive experiments on the CIFAR-10 and ImageNet-1K datasets demonstrate that our proposed BSMF achieves significant network compression while maintaining high reconstruction accuracy. The source code is publicly available at https://github.com/xianchaoxiu/BSMF

cs.CV↗

Consistent but Biased: Chair Personas Trade Noise for Bias in Multi-Agent Clinical Decisions

In clinical systems built from several large language model (LLM) agents, one agent often chairs the discussion and makes the final decision. Whether the chair's persona prompt changes that decision, or the agreement between repeated runs that some systems treat as reliability, has not been tested with the case held constant. We ask whether the persona changes the decision or only its timing, how it affects agreement between runs, and which behaviors appear when the prompt describes only personality traits. A simulated department of seven agents discussed 20 published cases of degenerative cervical myelopathy, with the case reports replayed unchanged in every run. Only the prompts of the chair and its deputy changed (no persona, or a consultative, autocratic or avoidant one), written either as rules of conduct or as personality traits. With personas written as rules, the chair chose to operate now in 93% of runs when autocratic and in 2% when avoidant, against 57% without a persona. Surgery was still planned in nearly all decisions, so the personas mainly changed whether to operate now or after more tests. These effects held with a second model and reworded prompts. The autocratic and avoidant personas made the three runs of a case agree more often, but only about as often as chance predicts, and Fleiss $κ$ did not rise. Prompts that described only personality traits kept more than half of the autocratic and avoidant effects on operating now. They reproduced quick decisions, delays and extra tests, but not ignoring red flags, dropping backup plans or passing the decision to the family. Studies should report and vary the chair's persona as they do the model, and correct agreement between runs for chance before calling it reliability.

cs.MA↗

UGOD: Uncertainty-Guided Opacity and Dropout for Sparse-View 3D Gaussian Splatting

Sparse-view 3D Gaussian Splatting is prone to overfitting because limited observations leave many Gaussian primitives weakly constrained, yet their contributions are still accumulated through alpha blending. Without uncertainty estimation, the renderer cannot distinguish unreliable primitives from well-constrained ones, allowing their erroneous contributions to corrupt novel-view synthesis. We introduce UGOD, an uncertainty-guided framework that estimates a view-dependent uncertainty score for each Gaussian and uses it to regulate its rendering contribution. A lightweight uncertainty head conditioned on Gaussian attributes and viewing direction predicts this score, which then drives a differentiable opacity-modulation mechanism that attenuates high-uncertainty primitives before compositing. During training, a detached soft-dropout branch applies an uncertainty-controlled continuous keep mask to discourage the model from relying on poorly constrained Gaussians and thereby reduce overfitting. Crucially, detaching the uncertainty score prevents gradients from this stochastic regulariser from biasing or collapsing the uncertainty prediction. Experiments on Mip-NeRF~360 and LLFF show that UGOD improves sparse-view novel-view synthesis while producing more compact Gaussian representations than the compared methods. These results demonstrate that Gaussian uncertainty provides an effective rendering-time control for sparse-view reconstruction.

cs.CV↗

Query, Align, and Distill: Navigation-Aware Cross-Modal Interaction for Efficient Vision-and-Language Navigation

Recent large-scale Vision-and-Language Navigation (VLN) models deliver strong accuracy but remain costly to deploy due to their large parameter counts and computational requirements. We tackle efficient VLN in two steps. First, we build a high-performing teacher that makes navigation evidence selection explicit and compressible. The teacher introduces a small set of learnable query slots to extract global and local action-sufficient navigable evidence from panoramic observations via a Navigable Query Generator, then progressively grounds these evidence tokens to the instruction with an Instruction-Query Aligner for policy prediction. Second, using this explicit query bottleneck as a distillation interface, we train a compact student by transferring both where to attend and what to do. We distill the teacher's global and local navigable queries with a navigation-aware token-adaptive objective, then further match action distributions during fine-tuning. Experiments on standard VLN benchmarks demonstrate that our student nearly matches the teacher's navigation performance while reducing the number of parameters by 93.65% compared to the teacher.

cs.CV↗

OneTrans-V2: Unifying Retrieval, Pre-rank, and Fine-rank with One Transformer in Industrial Recommender

Industrial recommendation systems typically operate as a \emph{cascade} of retrieval, pre-rank, and fine-rank, but these stages are usually trained and served as separate models, causing repeated user-sequence encoding, isolated optimization, and duplicated engineering effort. Building on OneTrans' model-level unification, we present OneTrans-V2, one Transformer that unifies the entire cascade. It encodes the user behavior sequence once as a shared context while preserving stage-specific candidate features and computation. Joint training lets the three stages reinforce one another and enables in-model knowledge distillation from fine-rank to pre-rank. We scale the shared backbone with sparse mixture-of-experts (MoE), which increases capacity with bounded activated computation, and stabilize scaling with $μ$P-style parameterization. To consolidate objective-specific retrieval channels, we introduce Decision-Conditioned Generative Retrieval (DCGR). DCGR predicts a decision prefix describing the upcoming interaction and generates items conditioned on it, allowing business objectives to steer a single generative process. Finally, Sequence-Native Training (SNT) organizes training around each user's lifelong behavior sequence and amortizes its encoding across exposures. Deployed across all three stages of a large-scale industrial recommendation system, OneTrans-V2 improves gross merchandise value (GMV) by 9.74\% and, with a co-designed serving stack, delivers $3.2\times$ the throughput of the cascade it replaces under the same hardware budget.

cs.IR↗

Study on Thickness and Temperature Dependence of Thermoelectric Properties in SnS Nanofilms

SnS as an environmentally friendly, cost-effective, and earth-abundant narrow-bandgap semiconductor material, has demonstrated significant application potential in the field of medium-temperature thermoelectric conversion. However, the thermoelectric performance of its bulk counterpart is inherently constrained by intrinsic point defects (e.g., vacancies) and the material's specific band structure. Low-dimensional engineering has emerged as a pivotal strategy for overcoming these limitations and enhancing thermoelectric performance. In this work, we systematically investigate the thermoelectric properties of SnS nanofilms with distinct thicknesses (82 nm, 199 nm, 616 nm, and 813 nm) across a temperature range of 300-600 K. Measurements were conducted using time-domain thermoreflectance (TDTR) and a dedicated thin-film thermoelectric parameter test system (ZEM-3). Our results confirm that low-dimensionalization effectively boosts the thermoelectric performance of SnS, with the thermoelectric figure of merit (ZT) displaying a pronounced dependence on both film thickness and temperature. All four SnS thin films exhibit thermoelectric performance that is markedly superior to that of bulk SnS. This enhancement is primarily attributed to the quantum confinement effect, energy filtering effect, and intensified phonon scattering, all of which are induced by the low-dimensional structural characteristics. This work provides not only experimental evidence and theoretical insights for the performance optimization of SnS nanofilms but also establishes a foundational framework for the development of high-efficiency, eco-friendly medium-temperature thermoelectric materials, thereby holding significant scientific value and practical implications.

cond-mat.mtrl-sci↗

APPSim-Bench: Bridging Real-world Apps and Reproducible Evaluation for Mobile GUI Agents

Mobile GUI agents can execute tasks from natural-language instructions, but their evaluation remains difficult to make both realistic and reproducible. Existing benchmarks typically trade off these goals: simplified apps lack real-world mobile complexity, whereas live commercial apps introduce uncontrolled variation from recommendations, advertisements, accounts, and changing content. We propose AppSim-Bench, which addresses this trade-off through controllable simulated apps that preserve task-relevant interaction logic while supporting deterministic evaluation. Built through a coding-agent-assisted and human-verified workflow, it contains 557 tasks across 17 high-frequency Chinese and English apps. Its controllable backend data and outcome-based verification remove major sources of environmental stochasticity, enabling reproducible cross-model comparison. Evaluating 19 GUI agents, spanning general-purpose and GUI-specialized systems, we find that autonomous mobile execution remains far from solved. The best model completes only 50.27% of tasks, and 28.55% of tasks are not solved by any agent. Further analysis shows that failures concentrate in longer workflows, numerical reasoning tasks, and inefficient trajectories marked by high action overhead and budget exhaustion. Our project is available at https://github.com/Acrab-Agentic-Labs/AppSim.

cs.AI↗

Direct Detection of Type II-P Supernova Progenitors with the $\textit{Euclid}$ and CSST Surveys

Identifying and characterizing supernova (SN) progenitor stars remains a central yet difficult goal in SN research, limited by archival images lacking sufficient depth or spatial resolution and circumstellar dust biasing intrinsic parameter estimates. This field will be revolutionized by $\textit{Euclid}$ and the upcoming Chinese Space-station Survey Telescope (CSST), which conduct deep, wide-field, high-resolution and multi-band imaging surveys. We evaluate their detection capability by comparing model magnitudes of RSG progenitors with detection limits, finding their optical and near-infrared filters highly effective. Monte-Carlo simulations predict that completed $\textit{Euclid}$ and CSST surveys will enable $\lesssim$13 (or 24) progenitor detections per year within the mass range of 8--16 (or 8--25)\,$M_\odot$, an order of magnitude higher than the current detection rate of $\sim$1 per year (primarily based on HST). With the circumstellar dust, the emerging spectral energy distribution (SED) of the SN progenitor is mainly affected by the optical depth and is almost independent of dust temperature in their survey filters. Mock tests demonstrate that the progenitor mass and dust optical depth can be derived simultaneously by fitting the observed SED over 11 survey filters while fixing dust temperature to a typical value. $\textit{Euclid}$ and CSST will significantly enlarge the sample of direct progenitor detections with accurate mass measurements, crucial for resolving the long-standing RSG problem.

astro-ph.SR↗

A statistical study of the environmental age of core-collapse supernovae based on VLT/MUSE integral-field-unit spectroscopy

We aim to understand the progenitor channels of CCSNe via a statistical study of the ages of their environments. We compiled a large and minimally biased sample of 128 CCSNe discovered by untargeted wide-field transient surveys and with archival VLT/MUSE integral-field-unit spectroscopy. We measured the local Hα luminosity within a 300-pc aperture centered on the SN explosion site as an empirical proxy for the environmental age. We find that the mean local H$α$ luminosities are ordered as II(P) $\approx$ IIb $\lesssim$ Ib $<$ Ic. The differences among Types~II(P), IIb and Ib are very small, if any. Type~Ic SNe are located in clearly younger environments than the other types. Our result suggests that Type Ic SNe have much younger and more massive progenitors than the other CCSN types and they likely originate from a distinct progenitor channel. The distinction between Types II(P), IIb and Ib SNe is insensitive to progenitor mass and mainly due to the different binary separation; in contrast, Type Ic SNe predominantly require much higher-mass progenitors accompanied by close companions with large mass ratios and/or much stronger stellar wind that depends sensitively on progenitor mass.

astro-ph.SR↗

Doc-V*:Coarse-to-Fine Interactive Visual Reasoning for Multi-Page Document VQA

Multi-page Document Visual Question Answering requires reasoning over semantics, layouts, and visual elements in long, visually dense documents. Existing OCR-free methods face a trade-off between capacity and precision: end-to-end models scale poorly with document length, while visual retrieval-based pipelines are brittle and passive. We propose Doc-$V^*$, an \textbf{OCR-free agentic} framework that casts multi-page DocVQA as sequential evidence aggregation. Doc-$V^*$ begins with a thumbnail overview, then actively navigates via semantic retrieval and targeted page fetching, and aggregates evidence in a structured working memory for grounded reasoning. Trained by imitation learning from expert trajectories and further optimized with Group Relative Policy Optimization, Doc-$V^*$ balances answer accuracy with evidence-seeking efficiency. Across five benchmarks, Doc-$V^*$ outperforms open-source baselines and approaches proprietary models, improving out-of-domain performance by up to \textbf{47.9\%} over RAG baseline. Other results reveal effective evidence aggregation with selective attention, not increased input pages.

cs.CL↗

KAN We Flow? Advancing Robotic Manipulation with 3D Flow Matching via KAN & RWKV

Diffusion-based visuomotor policies excel at modeling action distributions but are inference-inefficient, since recursively denoising from noise to policy requires many steps and heavy UNet backbones, which hinders deployment on resource-constrained robots. Flow matching alleviates the sampling burden by learning a one-step vector field, yet prior implementations still inherit large UNet-style architectures. In this work, we present KAN-We-Flow, a flow-matching policy that draws on recent advances in Receptance Weighted Key Value (RWKV) and Kolmogorov-Arnold Networks (KAN) from vision to build a lightweight and highly expressive backbone for 3D manipulation. Concretely, we introduce an RWKV-KAN block: an RWKV first performs efficient time/channel mixing to propagate task context, and a subsequent GroupKAN layer applies learnable spline-based, groupwise functional mappings to perform feature-wise nonlinear calibration of the action mapping on RWKV outputs. Moreover, we introduce an Action Consistency Regularization (ACR), a lightweight auxiliary loss that enforces alignment between predicted action trajectories and expert demonstrations via Euler extrapolation, providing additional supervision to stabilize training and improve policy precision. Without resorting to large UNets, our design reduces parameters by 86.8\%, maintains fast runtime, and achieves state-of-the-art success rates on Adroit, Meta-World, and DexArt benchmarks. Our project page can be viewed in \href{https://zhihaochen-2003.github.io/KAN-We-Flow.github.io/}{\textcolor{red}{link}}

cs.RO↗

EgoMemReason: A Memory-Driven Reasoning Benchmark for Long-Horizon Egocentric Video Understanding

Next-generation visual assistants, such as smart glasses, embodied agents, and always-on life-logging systems, must reason over an entire day or more of continuous visual experience. In ultra-long videos, relevant information is sparsely distributed across hours or days, making memory a fundamental challenge: models must accumulate information over time, recall prior states, track temporal order, and abstract recurring patterns. However, existing week-long video benchmarks are primarily designed for perception and recognition, such as moment localization or global summarization, rather than reasoning that requires integrating evidence across multiple days. To address this gap, we introduce EgoMemReason, a comprehensive benchmark for week-long egocentric video understanding through memory-driven reasoning. EgoMemReason evaluates three complementary memory types: entity memory, tracking how object states evolve and change across days; event memory, recalling and ordering activities separated by hours or days; and behavior memory, abstracting recurring patterns from sparse, repeated observations over the whole week period. EgoMemReason comprises 500 questions across three memory types and six core challenges, with an average of 5.1 video segments of evidence per question and 25.9 hours of memory backtracking. We evaluate EgoMemReason on 17 methods across MLLMs and agentic frameworks, revealing that even the best model achieves only 39.6% overall accuracy. Further analysis shows that the three memory types fail for distinct reasons and that performance degrades as evidence spans longer temporal horizons, revealing that long-horizon memory remains far from solved. We believe EgoMemReason establishes a strong foundation for evaluating and advancing long-context, memory-aware multimodal systems.

cs.CV↗

D2-ScaleAgent: Dual-Dimensional Scaling for Long Document Understanding

Multi-modal retrieval-augmented generation (RAG) is a key technique for visually rich long document understanding. Existing multi-modal RAG methods are progressively advancing toward multi-agent systems: they first retrieve relevant pages based on a query, and then iteratively understand information within those pages. However, these methods typically rely on fixed workflows and lack the ability to dynamically scale computation at test time, often leading to insufficient evidence. To address this, we propose D2-ScaleAgent, an agentic framework that introduces a dual-dimensional scaling paradigm for retrieval and reasoning. The core of D2-ScaleAgent is a Verifier agent-driven dynamic routing loop based on the intrinsic difficulty of the query, centered around a continuously updated evidence bank that serves as the agent's dynamic working memory: when retrieval needs to be expanded, the agent routes outward (retrieval scaling), decomposing the query into attributes and performing parallel page retrieval, followed by adaptive pruning to ensure comprehensive evidence coverage. When fine-grained reasoning is required, the agent routes inward (reasoning scaling), dynamically selecting sub-agents with varying granularity and count to extract evidence from pages. Finally, D2-ScaleAgent achieves logical closure over the evidence chain. Extensive experiments demonstrate that D2-ScaleAgent is effective on long and visually rich document benchmarks like MMLongBench-Doc, LongDocURL, etc.

cs.CL↗

WitCert: Sound Runtime Risk Observability and Gating for KV-Cache Quantization

KV-cache quantization is validated today by offline benchmark averages; a deployed system cannot tell whether compression is damaging the request it is serving right now. We give it a provably sound runtime meter -- a "DTrace for KV quantization": a per-(layer, head, step) upper bound on the total variation between exact and compressed attention. The meter has two tiers: a deterministic band-norm-witness bound, sound for any cache-preserving black-box quantizer and for any query (adaptive-safe, worst-case Cauchy--Schwarz plus RoPE band-unitarity), and a tighter probabilistic certificate for a controlled subtractively-dithered INT8 quantizer under an explicit request-level failure budget (stated for non-adaptive queries; core theorems machine-checked in Lean 4). Three results. Observability: the meter enters SGLang through an env-guarded patch, and any scheme registered as one tensor function is measured in live serving. Repair: meter-driven gating -- risk-ranked where the witness is saturated, certified where it is informative -- empirically restores the quality floor at benchmark scale, e.g. raw-cast fp8 from 22.8 back to 79.7 on hard RULER tasks with the difference from uncompressed bounded at $[+0.0,+0.8]$ by a paired test. Analysis: aggressive schemes survive on cross-layer error cancellation, not per-step fidelity -- in a 28-layer sweep, no single layer's pollution alone loses anything (0/28) -- and the certified int8 cache serves $1.88\times$ more KV tokens at the same memory in SGLang. All artifacts, guards, and the Lean development are released at https://github.com/metask-ai/witcert-kv-certificates; every number regenerates from the shipped artifacts by one command.

cs.AR↗

Runtime Observability for Heterogeneous Attention Memory

Modern models no longer keep a plain KV cache: latent caches, learned sparse selectors and recurrent states each carry the model's memory in a different form, and each fails differently under compression. We give a runtime observability contract that covers all four memory classes with three operators, instantiate it on six model configurations across five architecture families, and compose the per-stage bounds into an executable request-level risk ledger. Contracts carry their error metric as a type -- composition is only defined when metrics match, and this check rejected our own first composed chain; the repaired chain crosses metrics through two proved bridges, and whatever no formal system can certify is measured instead, dropping the composed tier to empirical automatically: every claim is certified, partially certified, or empirical, composition inherits the weakest tier, and the tier is decided by the machine. Replayed over $12.4$M entry reads and run under eight-way concurrency with per-request budgets and fail-closed identity attribution, the ledger quantifies the honest trade-off on today's witness and holds its risk budget with zero violations. A fused always-on probe observes a declared one-layer subset under CUDA graphs inside the serving noise floor. Applied to a served DeepSeek-V4 stack with a packed compressed-KV prototype, the same machinery localizes a silent corruption to a precise structural boundary -- exact in the eviction-free, identity-isolated regime, with every observed failure in an eviction or slot-reuse regime -- through a machine-adjudicated discrimination campaign whose calculus rejected two of our own confounded inferences along the way. All artifacts, guards, and the Lean development are released at https://github.com/metask-ai/witprobe-attention-memory; every number in this paper regenerates from the shipped artifacts by one command.

cs.AI↗

Decoding Phenotypes: A Framework for Fusing Genomic Language Models and Neuroimaging

Neuroimaging and genetic testing are two important clinical references for nervous system diseases, offering complementary diagnostic information. However, integrating genomic and neuroimaging data for precise disease diagnosis is challenging due to cross-modality heterogeneity. Existing imaging-genetics approaches mainly encode genetic information as hard-coded labels, which lose the local sequence context around disease-associated variants. To address this limitation, we propose GeneFuse, a multimodal learning framework that aligns genetic representations from pre-trained Genomic Language Models (GLMs) with features extracted from images. GeneFuse integrates two components: (1) Genotype-Conditioned Feature Modulation (GCFM), a FiLM-inspired module that uses genomic embeddings to modulate image feature maps; and (2) Uncertainty-aware Genomic Residual Fusion (U-GRF), a fusion strategy that uses imaging-derived predictive uncertainty to gate the contribution of genotypic features. We evaluate GeneFuse on early cognitive decline identification (NC vs. MCI) and dementia screening (NC vs. AD). In the APOE-centered setting, GeneFuse achieves AUROCs of 0.77 and 0.83, outperforming existing imaging-genetics fusion methods. These results indicate that GLM-derived genomic embeddings provide additional information to imaging.

cs.AI↗