arXiv subjects

Yun Liu

Publications and source records attributed to Yun Liu.

At least 19 recordsLinked to original sources

arXiv2026

Towards Multi-Method Evaluation of AI in Health: Integrating Benchmarks, Autoraters, and Expert Evaluation

General-purpose generative AI (GenAI) models are being rapidly adopted in healthcare, but they are challenging to assess with traditional evaluation frameworks. These conceptual, methodological, and practical challenges parallel decades of work in evaluating clinician and health system quality and safety. We describe the adaptation of methods from several fields, combined with emerging technical approaches, into a structured, multi-method evaluation framework designed to give well-rounded assessments of GenAI in health at scale. The approach first anchors evaluation to Critical User Journeys (CUJs) to help prioritize out of a large choice space, and maps these to a phased AI system lifecycle (base model development, product development, and post-deployment monitoring), since evaluation at each stage requires different methodological approaches. In this context, the approach then integrates three distinct evaluation methods (closed-question benchmarks, LLM-as-a-judge autoraters, and expert human review) across three core domains: accuracy, safety, and usefulness. Taken together, this approach provides a strategic framework for bridging the gap between benchmarking results and performance in the real world with the goal of supporting decision-making and accountability across development, procurement, and deployment for product teams, health systems, and regulators.

cs.CY↗

arXiv2026

Predicting Primordial Non-Gaussianity Measurements in the CSST Spectroscopic Survey

Primordial non-Gaussianity (PNG) is a fundamental probe of the physics of the early Universe and inflation. Here we present a comprehensive study of the constraints on the local-type PNG parameter, $f_{\rm NL}$, for the spectroscopic galaxy survey of the upcoming Chinese Space-station Survey Telescope (CSST). Utilizing the high-resolution Jiutian N-body simulation suite, we construct realistic mock catalogs for emission line galaxies (ELGs) at three representative redshifts $z=0.3$, 0.6, and 0.9. The expected CSST observational characteristics are also considered, including redshift uncertainties and selection functions based on signal-to-noise ratios of emission lines. We develop a robust analysis framework for the redshift-space galaxy power spectrum and bispectrum that accounts for redshift-space distortions, scale-dependent bias, and nonlinear effects. For the parameter $p$ in the PNG bias $b_ϕ$ with $p=0.55$, we find $f_{\rm NL}=-4\pm43$ if only using the power spectrum data, and $f_{\rm NL}=3\pm41$ after combining the bispectrum data, reducing the uncertainty by $\sim$5%.For the $p=1$ case, the constraint results are $f_{\rm NL}=-3\pm86$ and $54^{+70}_{-80}$ for comparison. This work demonstrates the potential of the Stage IV surveys like CSST to probe inflationary physics, and highlights the importance of higher-order statistics in extracting information from large-scale structure surveys.

astro-ph.CO↗

arXiv2026

A prospective randomized study of patient-centered AI for clinical visit preparation

Patients often arrive at brief, time-constrained clinical visits unsure of what to expect or ask, and millions of adults now turn to general-purpose AI chatbots to prepare beforehand. Yet whether conversational AI improves visit readiness remains unclear. We designed a large language model (LLM)-based visit preparation agent to guide patients before an appointment by gathering clinical context, setting visit expectations, providing relevant health information, building a prioritized list of questions for their clinician, and generating a summary. In a prospective consented randomized study in the Google Health app, 802 U.S. adults held 866 health-related conversations with the visit preparation agent (standard or concise variant) or a minimally prompted control chatbot built on the same LLM. The visit preparation agent significantly outperformed the control across blinded clinician evaluations and patient self-reports. In clinician reviews of completed sessions (n = 272), the agent achieved a 16.6% higher rate of comprehensive preparation (67.2% vs. 50.6%; P < 0.05) and an 8.5% higher rate of alignment with patients' health context (98.9% vs. 90.4%; P < 0.01), while maintaining high medical safety and appropriateness across all study arms. Patient post-surveys (n = 460) showed 16.8% higher visit preparedness (73.8% vs. 57.0%; P < 0.01), 14.3% higher confidence in knowing what to ask (81.4% vs. 67.1%; P < 0.01), and 12.0% higher overall satisfaction (74.0% vs. 62.0%; P < 0.05). Clinician and participant metrics were not significantly correlated, indicating that expert review and patient self-assessment capture complementary dimensions of patient-facing AI. These findings suggest that LLM chatbots can safely help patients prepare for clinical visits, with greater benefit when following a structured workflow, and motivate future studies on in-visit communication and clinical outcomes.

cs.HC↗

arXiv2026

Training language models in simulation for psychiatric intake

In psychiatric practice, the clinical intake interview remains a primary diagnostic instrument, requiring the clinician to establish rapport while disentangling complex, overlapping presentations. Here, we show that multi-turn reinforcement learning (RL) in simulation environments improves language model performance on structured text-based psychiatric intake, spanning history-taking, differential diagnosis, and initial management planning. In a randomized, double-blind crossover study ($N = 60$ matched scenario pairs, 120 encounters) comparing the RL-trained agent against its identically prompted base model, 32 board-certified psychiatrists role-played standardized patients while a separate panel of 15 independent, board-certified psychiatrists scored the de-identified transcripts. The trained agent outperformed the base model in 76.7% of matched scenario pairs ($p < 0.0001$), with gains on an adapted clinical-examination rubric ($d_z = 0.77$, $p < 0.0001$), patient-centered communication (82.0% vs. 65.2%, $p < 0.001$), and differential diagnosis appropriateness (68.3% vs. 33.3%, $p < 0.001$). Across 127 psychiatrist-validated simulated intake scenarios, the trained agent nearly doubled its base model's clinical rubric score (0.706 vs. 0.379; composite reward 0.564 vs. 0.294, $p < 0.0001$). To systematically probe conversational safety, we established two Delphi-consensus taxonomies, of clinical safety risks and relational competencies, from which 256 challenging adversarial presentations were derived; we find that training in simulation environments more than halved under-triage of acute safety risks (12.9% vs. 30.5%, $p < 0.0001$). Simulation environments thus provide a controlled testbed for training and stress-testing psychiatric conversational agents, with gains that transferred to psychiatrist-enacted encounters, though residual gaps require safeguards and clinician oversight.

cs.AI↗

arXiv2026

AxisQ: Depth-Adaptive Channel--Frequency Quantization for Camouflaged Object Detection

Camouflaged object detection (COD) depends on weak texture and boundary cues that can be destroyed by aggressive post-training quantization (PTQ). Existing PTQ methods mainly optimize quantizers or reconstruct network blocks, while using a single precision-allocation domain throughout hierarchical backbones. Instead, we introduce AxisQ, a depth-adaptive bit-allocation framework motivated by the observation that the useful allocation axis changes with feature depth. Under a fixed per-layer average activation-bit budget, AxisQ allocates precision in the channel domain for shallow features and in the local-frequency domain for deep features, using COD-aware sensitivity to better preserve subtle foreground regions and ambiguous boundaries. We also formulate a boundary-aware Hessian approximation for reconstruction by incorporating boundary-weighted gradient and error statistics. Experiments across multiple COD models and diverse benchmarks demonstrate improved accuracy over the strongest baselines at low bit widths, with relative weighted F-measure gains of 12.7% for CFRN and 33.3% for CamoFormer-P on CAMO under W2A2 quantization. These results support adapting the allocation domain to feature depth as a transferable quantization strategy for hierarchical COD models.

cs.CV↗

arXiv2026

A Formation-Stage Bottleneck for Exomoons around Close-in Rocky Planets

Massive moons around rocky exoplanets are expected outcomes of giant impacts, yet no exomoon has been securely confirmed. We address this problem using meshless finite-mass simulations of giant impacts between differentiated rocky planets that include stellar tidal and Coriolis forces in a local co-rotating frame for orbital periods of 1--300 days, and compare these simulations with otherwise identical collisions in isolation. We find that stellar perturbations impose a severe formation-stage bottleneck on moon-forming disks. Ultra-short-period impacts leave essentially no surviving circumplanetary disk, whereas 10-day systems retain only strongly depleted, radially truncated, and dynamically compact disks. High impact velocities, expected for close-in planets, are devastating for disk generation, while retrograde impacts can slightly compensate for disk mass depletion. These results show that stellar perturbations can suppress exomoon formation around close-in rocky planets, reshaping expectations for the demographics of exomoons.

astro-ph.EP↗

arXiv2026

Planetary Geospatial Foundation Models: A New Paradigm for Global Public Health

The efficacy of traditional disease prediction is limited by spatial gaps and temporal lags, which impact the timing and targets of resource deployments. Outbreaks escalate undetected, chronic disease burdens are quantified years later, and at-risk populations in data-sparse regions remain unaddressed. Planetary geospatial foundation models complement existing epidemiological workflows to provide operational improvements, encoding multimodal search, mobility, and environmental signals into generalizable place representations. As illustrations of this complementarity, we present independent global health case studies of Google Earth AI's Population Dynamics Foundation Model (PDFM) -- a foundation model for geospatial inference -- across four domains (vaccine-preventable, communicable, noncommunicable, maternal mental health), five tasks (spatial extrapolation, interpolation/nowcasting, probabilistic forecasting, prospective forecasting, risk stratification), and four countries (USA, Canada, Mexico, and the Democratic Republic of the Congo). Across these case studies, PDFM addresses critical surveillance gaps across domains: improving US-Canada border MMR vaccination coverage predictions by capturing cross-border behavioral spillovers domestic models miss; nowcasting cardiovascular disease to accelerate data availability; enhancing short-term municipal Mexican dengue forecasts for timely outbreak vector control; improving forecasts of cholera hotspots; and adding a transferable signal to individual-level postpartum-depression risk prediction in US states the model had never seen, while not replacing individual socioeconomic data or closing demographic screening gaps. Together, these results showcase capabilities of geospatial foundation models for public health surveillance.

cs.LG↗

arXiv2026

ResidencyRL: Reinforcement Learning in Simulated Clinical Environments

In medical education, physicians convert academic knowledge into clinical expertise through residency: years of training across thousands of encounters, with diverse sources of feedback and progressively greater autonomy. Much of clinical reasoning relies on the patient encounter, a dialogue in which a clinician elicits history, refines diagnostic hypotheses, and decides management under uncertainty. While large language models (LLMs) excel on static medical benchmarks, methods to optimize the full sequence of clinical decisions remain underdeveloped. We present ResidencyRL, a reinforcement learning (RL) method for training clinical artificial intelligence (AI) agents through simulated multi-turn clinical encounters (up to 60 dialogue turns and 8 tool calls per trajectory). ResidencyRL pairs the policy agent with LLM simulators capable of complex, adversarial behaviors, training against a structured reward aligned to diagnostic accuracy, management quality, communication, documentation, and safety. On held-out evaluations, the ResidencyRL agent improves diagnostic accuracy by 7.0% under adversarial conditions (88.0% vs. 81.0%) and reduces missed red flag rates by 31%, demonstrating rigorous mitigation of premature closure. Blinded expert clinicians validated these gains, preferring the trained agent in 87.6% of side-by-side comparisons. The procedural competencies transfer to unseen benchmarks: the agent outperforms the base model across all six clinical axes of the AMIE multi-visit benchmark, and shows consistent directional improvements on AgentClinic and CRAFT-MD. Our findings demonstrate that sequential clinical decision-making can be effectively learned through multi-turn RL in simulation, yielding robust, generalizable capabilities, paving the way towards clinical mastery. Prospective validation with real-world workflows remains necessary to establish clinical utility.

cs.AI↗

arXiv2026

Robust and Efficient Noisy-Label Time-Series Classification via Dynamic Time Warping Based Granular Ball Computing

Dynamic Time Warping (DTW)-based Nearest-Neighbor (NN) classifiers are effective for time-series classification but are vulnerable to mislabeled training samples and require numerous DTW computations during inference. We propose DTW-based Granular Ball Computing (DTW-GBC), which organizes temporally similar training samples into granular balls and performs classification at the granule level. We further develop two granular-ball construction strategies for DTW-GBC. Experiments on four benchmark datasets with symmetric label noise show that the two DTW-GBC variants generally mitigate the performance degradation caused by label noise while requiring substantially fewer comparisons than DTW-based 1-NN during inference. These findings suggest that DTW-GBC provides a favorable balance between classification robustness and inference efficiency.

cs.LG↗

arXiv2026

LEGO: Synergizing Expert GraphRAG and Expert Chain-of-Thought for Legal Reasoning

Large language models are increasingly applied to high-risk domains such as law, yet complex legal reasoning remains limited by two structural challenges. First, existing RAG and GraphRAG methods emphasize lexical or semantic similarity while overlooking normative relations among legal provisions. Second, vanilla Chain-of-Thought prompting may generate plausible rationales without enforcing the normative structure of legal reasoning. To deal with the bottleneck of pipelines in the legal reasoning domain, we propose LEGO, a dual-module framework that synergizes Legal Expert GraphRAG and expert Chain-of-thought for complex legal reasoning. ExpertGraphRAG uses an expert-annotated civil code graph encoding these normative relations with a greedy normative-coverage retrieval algorithm to dynamically extract instance-specific provision subgraphs, while ExpertCoT organizes the retrieved provisions and case facts into structured Provision-Fact-Conclusion reasoning. With a Qwen3-8B backbone, LEGO achieves 40.53% exact-match accuracy on LawExamQA_Civil, outperforming the evaluated RAG and CoT baselines and performing comparably to the evaluated larger models, while remaining robust on multi-hop questions. It also achieves the best results among the evaluated baselines on the open-ended benchmarks. Ablation studies confirm the individual and complementary contributions of both modules, demonstrating LEGO's effectiveness in improving LLMs' complex legal reasoning ability. Code and dataset can be found in the link: https://github.com/BLK-WHT/LEGO

cs.CL↗

arXiv2026

Cross-correlating Squared kSZ and HI Intensity Fields: A Map-level ACT--MeerKAT Forecast

The kinetic Sunyaev--Zel'dovich (kSZ) effect probes the line-of-sight momentum of free electrons, but its blackbody spectrum and velocity-dependent sign make it difficult to isolate and cause conventional cross-correlations with density tracers to vanish. In this work, we investigate the cross-correlation of the squared kSZ field with the projected squared HI intensity field using mock ACT and MeerKAT observations constructed from a Jiutian-1G simulation lightcone. We generate halo-based kSZ maps and MeerKAT-like HI brightness-temperature cubes over a $13.5^\circ\times13.5^\circ$ field. The simulations incorporate ACT beam smoothing and effective internal-linear-combination noise, together with the frequency-dependent MeerKAT beam, scanning strategy, map-making, and inhomogeneous thermal noise. We analyze signal-only, idealized amplitude-reconstructed, and Wiener-filtered kSZ maps. The resulting squared-field cross-correlations have diagnostic cumulative signal-to-noise ratios of approximately $3$-$4$. Our map-level results demonstrate the detectability of the ${\rm kSZ}^2\times {\rm HI}^2$ correlation in future joint ACT--MeerKAT observations.

astro-ph.CO↗

arXiv2026

Ruby-ASR: Evidence-Preserving Supervision for Joint Orthographic and Lexical-Reading Recognition

Conventional Japanese automatic speech recognition (ASR) is supervised by an orthographic transcript, although the same written form can correspond to different lexical readings realized in speech. Such utterances receive an identical target, so their reading distinction is absent from the supervision interface and cannot be recovered reliably by post-hoc text-only grapheme-to-phoneme conversion. We present Ruby-ASR, which refines the conventional target into a span-bound orthographic--lexical-reading sequence. Unlike separate full-sentence orthographic and phonological outputs, the ruby representation locally binds each written span to its realized reading and permits deterministic recovery of both views. We instantiate the target under subtitle-style and verbatim-style transcription conventions using a Qwen3-ASR backbone; a mora-level CTC objective provides auxiliary monotonic reading supervision. The experimental results across five Japanese benchmarks show that refining the recognition target can improve lexical-reading recovery without sacrificing readable orthographic transcription. We release the checkpoints and inference code.

cs.CL↗

arXiv2026

Accelerator and Brake: Dynamic Persuasion with Dead Ends

This paper studies dynamic persuasion in a strategic-experimentation relationship in which the principal has a single-peaked preference over the agent's stopping time. Excessive experimentation may end in a dead end. The principal privately observes project quality, which determines the agent's payoff conditional on success, while both parties learn about feasibility only through the agent's experimentation. We show that an optimal policy uses at most two one-shot disclosures: an accelerator before the principal's ideal stopping time and a brake afterward. A local Arrow--Pratt comparison of induced payoffs over stopping time determines whether the accelerator is concentrated or gradual. Under common discounting, the comparison yields a one-shot accelerator. Under heterogeneous discounting, the one-shot result remains robust unless the agent is sufficiently more impatient than the principal, in which case the ranking reverses over an interval and the accelerator can take a one-shot--gradual--one-shot form.

econ.TH↗

arXiv2026

Quasi-flag manifolds and moment graphs

We introduce and study a new class of topological $G$-spaces generalizing the classical flag manifolds $G/T$ of compact connected Lie groups. These spaces, which we call the $m$-quasi-flag manifolds $ F_m = F_m(G,T) $, are topological realizations of the algebras $ Q_k(W) $ of $k$-quasi-invariant polynomials of the Weyl group $ W $ in the sense that their (even-dimensional) $G$-equivariant cohomology $ H_G(F_m, {\mathbb C}) $ is naturally isomorphic to $ Q_k(W) $, where $ m $ is a $W$-invariant integer-valued multiplicity function on the system of roots of $W$ and $ k = \frac{m}{2}$ or $ \frac{m+1}{2}$ depending on whether $m$ is even or odd. Many topological properties and algebraic structures related to the flag manifolds can be extended to quasi-flag manifolds. We compute the cohomology of quasi-flag manifolds by constructing their rational algebraic models in terms of coaffine stacks -- a certain kind of derived stacks introduced by B.Toën and J. Lurie to provide an algebro-geometric framework for rational homotopy theory. Besides cohomology, we also compute the equivariant K-theory of quasi-flag manifolds and extend some of our cohomological results to the multiplicative setting. On the topological side, our approach is strongly influenced by the classical work on homotopy decompositions of classifying spaces of compact Lie groups; however, the diagrams that we use in our decompositions do not arise from collections of subgroups of $G$ but rather from moment graphs -- combinatorial objects introduced in a different area of topology called the GKM theory.

math.AT↗

arXiv2026

A phenomenological approach to direct ${\rm{K}}^{*}$ production and hadronic medium effects in nucleus-nucleus collisions at high baryon density

Short-lived hadron resonances serve as sensitive probes of the late-stage hadronic medium in heavy-ion collisions. Using the AMPT-HC model, we study ${\rm{K}}^{*}(892)$ production and its hadronic medium effects in Au+Au collisions at $\sqrt{s_{\rm{NN}}} = 3$ GeV, a region of high baryon density. We introduce a phenomenological direct-production mechanism for ${\rm K}^{*}$ by replacing a fraction of the final-state kaons produced in the ${\rm NN} \to {\rm NYK}$ and ${\rm MN} \to {\rm YK}$ channels with ${\rm K}^{*}$ resonances, with the substitution fraction controlled by a parameter $α$ while conserving four-momentum. The direct ${\rm K}^{*}$ is produced early, at about 6 fm/$c$, with little centrality dependence, whereas resonance fusion via ${\rm K}+π\to{\rm K}^{*}$ occurs later, with the mean production time increasing from about 8 to 10 fm/$c$ toward central collisions. Consequently, direct ${\rm K}^{*}$ mesons suffer stronger daughter rescattering, leading to a pronounced decrease in reconstruction efficiency toward central collisions, while the ${\rm K}^{*}$ survival rate remains close to unity. Elastic scattering of the daughters also shifts the invariant mass away from the resonance peak, contributing to the background-like component. The ${\rm K}^{*}/{\rm K}$ centrality dependence reflects the competition between direct production and resonance fusion and is sensitive to $α$. At 3 GeV, a moderate direct-production contribution may result in an increasing ${\rm K}^{*}/{\rm K}$ ratio toward central collisions, providing a testable prediction for future measurements.

nucl-th↗

arXiv2026

Propagating construction-time knowledge quality into medical question answering: A framework grounded in clinical guidelines

Large language models have facilitated knowledge graph (KG) construction from clinical guidelines, but extracted triples vary in structural validity and evidential support. Meanwhile, graph-augmented question answering (QA) systems typically optimize query relevance during retrieval, with limited reuse of quality information produced during KG construction. This creates a disconnect between construction-time quality control and inference-time evidence use. We investigate whether construction-time triple quality can serve as a persistent signal for downstream evidence selection and presentation. We propose a quality-aware framework that models structural conformance (SchemaConf) and evidential support (EvidScore) as complementary dimensions and fuses them into a per-triple quality signal, Q(t). Rather than using quality solely for filtering, the framework retains Q(t) and derived quality tiers as graph attributes and propagates them into quality-weighted subgraph retrieval and tier-conditioned evidence prompting, while preserving passage-level provenance. Experiments on Chinese diabetes clinical guidelines show that the utility of the quality signal is distribution dependent. Under cross-version and cross-model shift, the fused Q(t) provides stronger triple-quality discrimination than either component alone (AUC 0.748 vs. 0.703 for EvidScore and 0.645 for SchemaConf). In guideline-grounded QA, propagating construction-time quality reduces required-knowledge omission from 16.3% to 5.3% and conflicting outputs from 16.3% to 2.7%, with an evidence-grounded precision of 81.6% and near-zero invalid citations. Blinded clinician ratings favor the full framework over no retrieval (4.68 vs. 4.21 on a five-point scale) and approach the oracle condition (4.80), while cross-generator experiments show consistent trends.

cs.AI↗

arXiv2026

RegimeFormer: A Large Protein Model of Global Perturbation Regimes

Protein language models organize sequence and structure at scale, but a global representation of how proteins respond to mutation remains lacking. We present RegimeFormer, a large protein perturbation model coupled to RegimeAtlas, constructed by harmonizing and indexing 202,556,313 non-redundant protein sequences across the tree of life. A diversity-preserving one-million-protein subset provides the high-resolution training and inference layer, with 995,995 proteins yielding residue-level summaries across 407,048,356 residues and substitution-specific predictions available on demand. Across experimental deep mutational scanning, molecular benchmarks, structural confidence and evolutionary constraint, RegimeFormer identifies reproducible protein-level perturbation regimes that organize residue fragility, adaptability and predictive uncertainty. Regime conditioning improves substitution-specific prediction, with the largest relative gains under unseen-protein, unseen-family and low-homology evaluation. RegimeFormer-derived molecular priors further improve downstream transcriptomic and drug-response modelling. Together, RegimeFormer and RegimeAtlas provide a scalable framework for mapping, predicting and querying protein perturbation landscapes across global sequence space.

q-bio.QM↗

arXiv2026

Hyper-FSAD: Training-Free and Language-Free Few-Shot Anomaly Detection via Sparse Hyper Matching

Few-shot anomaly detection (FSAD) is particularly valuable when only a few normal images are available in a new target domain, while anomalous cases are rare, diverse, and difficult to enumerate in advance. However, existing methods often still require task-specific fitting or language prompts, and their patch-level retrieval commonly relies on brittle nearest-neighbor or fixed Top-$p$ rules. We propose \textbf{Hyper-FSAD}, a training-free and language-free framework that performs support-only inference with a frozen visual encoder. We formulate FSAD as support-only scoring in frozen feature space and establish the selective stability of sparse retrieval and sufficient conditions for normal--anomaly separation. Guided by this analysis, \textbf{Sparse Hyper Matching} uses \textit{sparsemax} to adaptively select support patches for each query patch, exactly suppressing below-threshold distractors without a manually specified retrieval hyperparameter. \textbf{Dual-Branch Image Scoring} further combines local reconstruction discrepancies with support-conditioned global deviation. Across four industrial and two medical benchmarks, Hyper-FSAD achieves the best overall performance across the 1/2/4-shot settings, while requiring only 52.6\,ms per image and 0.89\,GB GPU memory. The code will be released.

cs.CV↗