Searcharxiv⌕ Search

arXiv subjects

Zhi Liu

Publications and source records attributed to Zhi Liu.

At least 19 recordsLinked to original sources

ReForge: Keeping ABR Algorithms Never Finished with Verified Large Language Model Revisions

Designing an ABR algorithm for a network scenario takes an engineer months, and large language models (LLMs) now do this work in hours. But either way, the algorithm is fixed once it is deployed. It fits the networks seen during its design and falls behind on the networks that arrive afterwards. We ask whether an ABR algorithm can keep being designed after deployment, revised as each new network arrives, without becoming worse on any network it already serves. We propose ReForge, which keeps a pool of existing ABR policies unchanged and lets an LLM revise only the rules that decide which policy chooses the next bitrate. When a new network arrives, the LLM reads how the current rules and every policy perform on the networks served so far and rewrites the rules. The rewrite is deployed only if a replay in simulation shows that it improves the served networks as a whole and makes none of them worse. A fast GPU simulator that replays every episode exactly makes this acceptance test cheap enough to run for every revision. We evaluate ReForge on nine real-world network datasets, with new networks arriving one at a time. The revisions raise the mean QoE from 1.43 to 1.79, past the best single policy at 1.66 and close to the 1.86 of selecting the best policy for each dataset, while the QoE on the networks served first does not drop. The gain also holds in a real video player. All code, data, and records will be available.

cs.AI↗

Booster-based beam recycling for swap-out injection at the High Energy Photon Source

Fourth-generation synchrotron light sources employ ultralow-emittance storage rings with stringent injection requirements. On-axis swap-out injection alleviates the dependence on storage-ring dynamic aperture, but high-charge operation requires an efficient injector architecture capable of producing high-charge replacement bunches. This paper presents the accelerator physics design and performance analysis of a booster-based beam-recycling swap-out injection scheme implemented at the High Energy Photon Source (HEPS). In this approach, the full-energy booster serves as both an injector and a high-energy accumulator. An extracted storage-ring bunch is returned to the booster, merged with a low-charge bunch previously injected from the linac and accelerated to full energy. Following high-energy damping, the merged bunch is reinjected into the original storage-ring bucket. The scheme avoids the need for a dedicated accumulator ring while enabling high-charge bunch replacement. The recycling scheme was commissioned through staged machine studies. Full recycling-chain simulations, commissioning studies, and measured performance analysis are presented. The measured results characterize the recycling operation and quantify the transmission efficiency and performance limitations of the complete recycling loop. These results demonstrate the feasibility of the booster-based beam-recycling architecture and establish its operational basis for high-charge swap-out injection in future fourth-generation synchrotron light sources.

physics.acc-ph↗

Silent Neuron Theory and Plasticity Preservation for Deep Reinforcement Learning in Adaptive Video Streaming

Adaptive video streaming optimizes Quality of Experience (QoE) metrics by selecting appropriate bitrates according to varying network bandwidth and user demands. In practice, however, real-world network bandwidth often exhibits heterogeneity relative to training environments. Current methods predominantly tackle this problem through learning-based approaches designed to improve generalization performance. Our systematic investigation reveals a critical limitation: neural networks suffer from plasticity loss, significantly impeding their ability to adapt to heterogeneous network conditions. Through theoretical analysis of neural propagation mechanisms, we demonstrate that existing dormant neuron metrics inadequately characterize neural plasticity loss. To address this limitation, we have developed the Silent Neuron theory, which provides a more comprehensive framework for understanding plasticity degradation. Based on these theoretical insights, we propose the Reset Silent Neuron (ReSiN), which preserves neural plasticity through strategic neuron resets guided by both forward and backward propagation states. Moreover, we establish a tighter performance bound for ReSiN under non-stationary network conditions. In our implementation of an adaptive video streaming system, ReSiN shows substantial gains under non-stationary conditions in which the available bandwidth alternates between high and low regimes, improving the QoE by 178%, and it remains beneficial under stationary real-world traces, where it improves the QoE by 1.6%. Code will be available: https://tinyzqh.github.io/ReSiN/

cs.LG↗

An evidence-traceable benchmark of clinical reasoning in large language models

Benchmark scores for medical large language models (LLMs) are difficult to interpret when scoring criteria cannot be traced to clinical evidence. Here we developed Med Evidence Bench (MEB; https://www.medevidencebench.org.cn/), comprising 1,536 physician-constructed items across three levels: guideline knowledge, standardized clinical scenarios and real-world laboratory scenarios. Each criterion stating a medical claim is linked to a versioned source-text span. Across ten LLMs, the Evidence Trace Score (ETS) was 0.975-0.997 for guideline knowledge, 0.292-0.658 for standardized scenarios and 0.182-0.447 for real-world scenarios, and rankings varied across levels. For real-world scenarios, scores for identifying clinical information (mean, 0.770) exceeded those for matching evidence (0.165) or formulating recommendations (0.160). On 1,296 criterion-level judgments, automated grading agreed with the majority vote of three physicians (macro-F1, 0.930; 95% confidence interval, 0.915-0.945), although agreement was lower for evidence application. These findings from Chinese-language materials indicate that guideline recall does not ensure patient-specific evidence use and do not establish clinical effectiveness.

cs.CL↗

Deep Plug-and-Play Proximal Distance Algorithm for Robust Principal Component Analysis

Robust Principal Component Analysis (RPCA) is essential in signal processing, with applications in hyperspectral imaging, face recognition, and video background modeling. Traditional RPCA methods often suffer from insufficient utilization of data priors and struggle to capture complex image structures. Conversely, advanced deep denoisers like DnCNN, FFDNet, and DRUNet excel at leveraging image data priors but lack physical model constraints and demonstrate poor robustness against structured noise in RPCA tasks. This paper proposes a novel Plug-and-Play (PnP) method for multi-channel RPCA, integrating both the physical generative model and cutting-edge deep priors. Extensive experiments on RGB and hyperspectral datasets show that PnP-PDA achieves better quantitative performance than conventional RPCA methods and state-of-the-art methods, including deep denoisers and tensor RPCA approaches, while preserving visual details.

eess.IV↗

Functional Inference for the Mean Curve of Intraday Spot Correlation

In this paper, we propose estimators for the mean of the daily spot-correlation curves, which are robust to the presence of finite-variation jumps in the price processes. The related feasible functional central limit theorems for the proposed estimators have been established under a joint infill and long-span setting. Based on the asymptotic theory, we develop a test procedure for flatness of the curves. Monte Carlo simulations confirm the strong finite-sample performance of the estimators. In an empirical application to 30 U.S. stocks from 2010 to 2019, we find a pronounced end-of-day surge in the mean of correlation curves across diverse stock groupings, alongside substantial calendar-time non-flatness in the curves.

math.ST↗

SparkAlign: A Star-Hopping Heuristic for Efficient Unsupervised Plain Graph Alignment

Unsupervised plain graph alignment (UPGA) seeks corresponding nodes between two graphs from topology alone, without attributes or seed anchors. The dominant embed-then-match paradigm is expensive on both fronts: it repeatedly retrains node embeddings and then solves a costly global assignment, so it scales poorly to large graphs. We present SparkAlign, a training-free heuristic that, inspired by star-hopping in celestial navigation, locates a node's counterpart by measuring its heat-diffusion proximity to a shared set of high-confidence landmark matches. Although heuristic in spirit, the descriptor is principled: diffusion makes structurally equivalent nodes comparable across graphs, and a short stability analysis justifies a finite diffusion depth. Matching then uses two sparse, highly parallel strategies that avoid dense global assignment, one favoring speed and one preserving one-to-one consistency on a restricted candidate set. On three benchmarks SparkAlign attains accuracy approaching a structural identifiability reference while running up to two orders of magnitude faster than the strongest baseline and seeding markedly more accurate pseudo-anchors. Our code is available at https://github.com/MaxQ545/SparkAlign

cs.SI↗

DERM-3R: A Resource-Efficient Multimodal Agents Framework for Dermatologic Diagnosis and Treatment in Real-World Clinical Settings

Skin diseases impose a substantial and growing global health burden. Modern dermatologic therapies control acute manifestations rapidly, but single-target treatment paradigms, recurrent disease courses, and neglected systemic comorbidities limit their long-term effectiveness. Traditional Chinese medicine (TCM) offers a complementary, holistic approach through syndrome differentiation and individualized treatment, yet its clinical practice is hindered by non-standardized knowledge, incomplete multimodal records, and the difficulty of scaling expert-driven reasoning. We propose DERM-3R, a resource-efficient multimodal agents framework that models TCM dermatologic diagnosis and treatment under limited data and computational resources. We decompose and redefine real-world dermatologic decision-making into three essential issues: fine-grained lesion recognition, multi-view lesion representation with specialist-level pathogenesis modeling, and holistic clinical reasoning for syndrome differentiation and treatment planning. DERM-3R comprises three collaborative agents, DERM-Rec, DERM-Rep, and DERM-Reason, each addressing one of these issues. Built on a lightweight multimodal large language model and fine-tuned by partial-parameter finetuning on 103 real-world TCM psoriasis cases, DERM-3R performs strongly across multiple dermatologic reasoning tasks. Automatic metrics, LLM-as-a-Judge assessments, and human doctors show that, despite extremely limited training data and parameter updates, DERM-3R matches or even surpasses hundred-billion-parameter general-purpose multimodal models such as GPT-5.1 and Gemini-3-Flash. Our results indicate that structured, domain-aware multi-agent modeling is an effective alternative to brute-force scaling for complex clinical tasks, offering a practical and scalable paradigm for multimodal AI in dermatology and integrative medicine.

cs.AI↗

TCMClinicalReason-Bench: Can Language Models Reason from Pathogenesis to Prescription over Real-World Clinical Cases?

Large language models (LLMs) can generate clinical narratives that are insufficiently grounded in patient-specific evidence. In traditional Chinese medicine (TCM), errors can propagate from etiology and pathogenesis through syndrome diagnosis and treatment principles to prescription generation. We developed TCMClinicalReason-Bench using 2,000 multicenter electronic health record cases to distinguish case-grounded responses from fluent but unsupported diagnostic and therapeutic conclusions. Five general-purpose and two TCM-specific LLMs were evaluated in zero-shot settings. An evidence-constrained rubric assessed seven diagnostic and therapeutic components and three cross-block relations, allowing case-supported alternatives. Qwen3.7-Plus with TCM retrieval served as the automated judge, alongside parallel blinded ratings by five senior TCM clinicians on a 600-case subset. Structural completeness was nearly saturated (99.3-100.0%), but normalized content scores ranged from 40.7% to 54.1%. The five general-purpose models averaged 50.0%, versus 41.3% for the two smaller TCM-specific models. Cross-block logic consistency ranged from 60.8% to 66.8% and correlated moderately with content across cases (Pearson's r = 0.515-0.656). Deficits were greatest in prescription generation, prescription analysis, and symptom-guided modification. In judge stress testing on 100 independent cases, perturbation detection rates across the three relations were 57%, 56%, and 31%, with contradictions detected more reliably than omissions. Separating component quality from cross-block consistency localizes failures missed by endpoint and completeness metrics and identifies where clinician oversight remains necessary.

cs.AI↗

One Rollout Is All You Get: Fully Test-Time Adaptation for GUI Agents

GUI agents are deployed with frozen weights and discard everything they experience on the job. Existing ways to update an agent's weights assume something deployment withholds: ground truth, rollouts beyond the single attempt (retries, samples, practice runs), or a learning phase other than deployment. Because GUI actions can be irreversible, a deployed agent gets one attempt per task occurrence, in arrival order, and every attempt counts. No ground truth is available at any point. We define fully test-time adaptation for GUI agents by these constraints and pair it with a minimal weight-space method, SOLO. Auxiliary models read each episode: a judge selects the episodes it deems successful, and a proposer-verifier pair relabels a failed episode's prefix with the subtask that prefix completed. Admitted episodes enter a short sliding window, and each admission updates a small adapter by top-K self-distillation on the agent's own predictions, provided the window holds a judged success. On recurring task streams built from WebArena, VisualWebArena and MobileWorld, SOLO improves on the frozen agent with both UI-TARS-7B and Qwen3-VL-8B, by three to six points of success rate, and exceeds two in-setting memory methods on the web streams.

cs.LG↗

MotionQ: Operator-Conditioned Motion Quotients for Cross-Observation WiFi Gesture Recognition

WiFi gesture recognition is accurate in fixed deployments but often degrades when user orientation, available links, or transceiver placement changes. Unlike ordinary domain shifts, these changes alter the wireless observation operator, so the same motion is expected to produce different measurements. Existing methods nevertheless pursue domain-invariant features and largely overlook changing layouts and observation configurations. Yet changing the observation operator also changes which task-relevant motion cues are physically observable, rather than merely altering the appearance of a fixed set of cues. Under a local linearization of the WiFi forward process, we derive a common task-observability condition under which a strict common linear representation is recoverable from every geometry-induced operator while preserving the gesture task. When the condition fails, enforcing stronger alignment across additional heterogeneous source operators may discard task-relevant cues still observable under individual operators. We therefore present MotionQ, which generates an operator-conditioned two-support motion measure for each candidate geometry. A motion quotient removes only the arbitrary ordering of its unlabeled supports and is represented by permutation-invariant central moments. Rather than matching quotients across operators, single-link-retention interventions encourage each view to retain information sufficient for gesture recognition. Extensive evaluations show that MotionQ is robust to extrapolative observation operators.

cs.HC↗

On the Regularization Landscape for the Linear Recommendation Models

Recently, a wide range of recommendation algorithms inspired by deep learning techniques have emerged as the performance leaders on several standard recommendation benchmarks. While these algorithms were built on different DL techniques (e.g., dropouts, autoencoder), they have similar performance and even similar cost functions. This paper studies whether the models' comparable performance are sheer coincidence, or they can be unified under a single framework. We find that all linear performance leaders effectively add only a nuclear-norm based regularizer, or a Frobenius-norm based regularizer. The former ones possess a (surprising) rigid structure that limits the models' predictive power but their solutions are low rank and have closed form. The latter ones are more expressive and more efficient for recommendation but their solutions are either full-rank or require executing hard-to-tune numeric procedures such as ADMM. Along this line of finding, we further propose two low-rank, closed-form solutions, derived from carefully generalizing Frobenius-norm based regularizers. The new solutions get the best of both nuclear-norm and Frobenius-norm world.

cs.AI↗

Redesigning Service Level Agreements: Equity and Efficiency in City Government Operations

We study how government should allocate resources (e.g., maintenance of public infrastructure) over time, efficiently and equitably -- when these desiderata may conflict. In particular, we consider the design of Service Level Agreements (SLAs) in city government operations: promises that incidents such as potholes and fallen trees will be responded to within a certain time. We model the problem of designing a set of SLAs as an optimization problem with equity and efficiency objectives under a queueing framework; the city has two decision levers: how to allocate response budgets to different neighborhoods, and how to schedule responses to individual incidents. We first theoretically analyze a stylized model, deriving when the equity-efficiency trade-off is substantial in absolute terms and centralizing response efforts leads to relatively smaller improvements over optimal decentralized operations. We then develop a simulation-optimization framework to design and analyze policies in practice. Applying our framework to data from NYC, we find: (a) in practice, despite large potential trade-offs, equity improvements are possible with relatively small decreases in efficiency; (b) there exist policies that are both more efficient and equitable than status quo inspections; (c) the gain from centralization is {relatively small compared to the potential efficiency-equity tradeoff}. Our results provide widely applicable predictions about the scale of the equity-efficiency trade-off in similar resource allocation settings, and our data-driven simulation-optimization framework enables specific policy recommendations in government operations. Our code is available at https://github.com/nikhgarg/sla_design. Our theoretical statements have been formalized in Lean, available at https://github.com/nikhgarg/AppliedModelingLib/blob/main/papers/LG24ServiceLevelAgreements/.

cs.CY↗

Aplaud: Adaptive Personalized Low-Rank Decomposition for User-Specific LLM

In this paper, we study the problem of personalized survey response prediction using fine-tuned large language models (LLMs). This task poses unique challenges: limited per-user training data, scalability of model storage, and the need to exploit shared structure across survey questions. To address these issues, we propose Aplaud (Adaptive Personalized Low-rank and User-specific Nested Decomposition), a lightweight and scalable framework for LLM personalization. Aplaud extends the LoRA paradigm by separating adaptation into a frozen, shared low-rank basis and a compact user-specific correction, augmented with a rank-one residual for finer personalization. To further reduce per-user parameter cost and mitigate overfitting, the correction matrix can be factorized into an even lower-rank form. Empirical results demonstrate that Aplaud achieves efficient, scalable personalization across users while outperforming state-of-the-art LoRA-based personalized LLM approaches in both generalization and inference efficiency.

cs.AI↗

Checkerboard: Closed-Form and Data-Independent Trigger Design for Clean-Label Backdoor Attacks

Backdoor attacks threaten the deep-learning supply chain by poisoning a small fraction of the training data so that a model behaves normally on clean inputs but maps triggered inputs to an attacker-chosen class. Clean-label backdoor attacks are especially difficult to audit because poisoned examples preserve their semantic labels. Yet existing clean-label attacks often require surrogate model training, auxiliary data access, or iterative optimization. In this paper, we present \emph{Checkerboard}, a clean-label backdoor attack with closed-form, data-independent trigger design. We formulate trigger design through an input-space Fisher-separability objective and, under a ridge four-neighbor local-smoothness prior for natural images, obtain the pixel-wise checkerboard as a closed-form maximizer of the resulting design proxy without data access, model training, or optimization. Across four benchmark datasets, Checkerboard outperforms the evaluated norm-bounded clean-label attacks and achieves state-of-the-art performance under low global poisoning rates. For example, on CIFAR-10, under a trigger perturbation of $10/255$, poisoning 20 training samples achieves $95.72\%$ Attack Success Rate (ASR). On IN-100, a global poisoning rate of only $0.4\%$ yields over $83\%$ ASR without degrading clean accuracy. The proposed attack also remains effective against state-of-the-art backdoor defenses and shows resistance to adaptive defenses under simple modification.

cs.CR↗

US-VLA: An Ultrasound Vision-Language-Action Model for Embodied Abdomina

Artificial intelligence-assisted ultrasound scanning enhances diagnostic reliability and efficiency by providing real-time guidance for standardized image acquisition and reducing operator dependence. However, existing reinforcement learning and learning-assisted ultrasound scanning methods typically rely on carefully designed reward functions or extensive interaction data, which limits their generalization ability and stability across different devices, patient populations, and complex clinical scenarios. To address these challenges, we propose an ultrasound vision-language-action model (US-VLA) for automated ultrasound scanning that explicitly encodes clinical semantic goals and generates sequential probe manipulation actions under real-time ultrasound feedback. In particular, we first design an ultrasound-aware expert fusion module to jointly integrate ultrasound observations with auxiliary contextual information, enabling semantic ultrasound feedback to effectively guide the scanning process. Then, we construct US-VLA-Data, a real-world dataset covering liver and kidney examinations, which includes five clinically defined standard planes and comprises 320 expert scanning trajectories with approximately 80,000 synchronized timesteps. Extensive experiments demonstrate that US-VLA achieves competitive performance in ultrasound probe manipulation tasks, indicating its effectiveness and promising generalization within the evaluated abdominal ultrasound setting. The source code is available at https://github.com/VMVLab/US-VLA.

cs.RO↗

ForceU-VLA: A Force-Aware Vision-Language-Action Model for Embodied Ultrasound Scanning

Embodied intelligent ultrasound scanning enables the automation and standardization of the ultrasound examination process by integrating perception, decision-making, and execution capabilities. However, existing methods suffer from loosely coupled modeling between force and ultrasound modalities and lack awareness of scanning stages, which limits their ability to capture dynamic probe-tissue interactions. To address these issues, we propose ForceU-VLA, a force-aware Vision-Language-Action model for autonomous embodied ultrasound scanning, which leverages force signals and ultrasound image feedback throughout the scanning process to enable accurate and high-quality ultrasound acquisition. Firstly, we propose a Force-Ultrasound Synergistic Fusion Module (FUSFM) that synergistically fuses ultrasound visual and force-feedback information to provide stable, reliable guidance for probe motion. Secondly, a Stage-Adaptive Modulation Mechanism (SAMM) is proposed to accommodate the task requirements across different scanning stages by adaptively modulating multimodal features to enhance their representation quality. Additionally, we introduce ForceU-VLA-Data, a real-world, force-aware embodied ultrasound dataset that integrates visual, force, and action signals, including data from two organs across five representative clinical scanning views, and comprising 450 expert-collected trajectories with approximately 100,000 synchronized multimodal frames. Extensive experimental results demonstrate that ForceU-VLA significantly improves contact stability and probe pressure regulation in embodied ultrasound scanning, thereby effectively enhancing task execution quality and overall system reliability. The source code is available at https://github.com/VMVLab/ForceU-VLA.

cs.RO↗

TAGR: Task-Adaptive Geo-Semantic Retrieval across Heterogeneous Urban Delivery Tasks

As urban residents demand higher-quality instant delivery services, demand forecasting, estimated time of arrival prediction, and next-stop prediction have become critical tasks in urban delivery systems. However, the three tasks rely on historical evidence with substantially different structures and granularities. Meanwhile, their prediction targets differ in output form, making it difficult to integrate retrieved evidence through a single fusion mechanism without causing mismatch or negative transfer. These issues can increase prediction errors and subsequently lead to imbalanced resource allocation, where some regions experience unmet demand. In contrast, others suffer from underutilization, ultimately degrading overall efficiency and service quality. To address these challenges, we propose TAGR, a Task-Adaptive Geo-Semantic Retrieval framework across heterogeneous urban delivery tasks. TAGR constructs an LLM-based geo-semantic shared prior and task-specific retrieval banks to share regional knowledge while preserving heterogeneous task-specific evidence. The LLM-derived regional representations encode geographic context and semantic relations among regions, providing transferable knowledge for all three tasks without requiring online LLM inference. It further introduces a task-adaptive retrieval augmentation module to accommodate heterogeneous output structures, thereby mitigating information mismatch and negative transfer caused by a uniform retrieval-fusion strategy. Experiments on three city datasets demonstrate that TAGR consistently outperforms state-of-the-art baselines across multiple tasks.

cs.IR↗