Searcharxiv⌕ Search

arXiv subjects

Li Lin

Publications and source records attributed to Li Lin.

At least 19 recordsLinked to original sources

Patient-specific rubric-based evaluation of large language models across longitudinal cancer care

Large language models (LLMs) show growing promise in specialty medicine, yet existing evaluations largely focus on isolated clinical encounters rather than assessing the distinct clinical responsibilities along longitudinal care pathways. We introduce NPC-RubricBench, a phase-aware benchmark for evaluating multistage clinical reasoning across the cancer care continuum, using nasopharyngeal carcinoma as a defined disease context. Curated from real-world records of 405 patients at two cancer centers, it comprises 941 evaluation units across six sequential care stages, from work-up through follow-up, and 17,688 patient-specific criteria across five dimensions. Each assessment is restricted to clinical evidence available at the corresponding decision point. Across 30 general-purpose and medical LLMs, we find strong diagnostic scores did not ensure reliable downstream care. Decision Quality ranged from 56.8 to 94.7 in diagnosis on a 0-100 scale, but dropped to 8.9-56.0 in staging and spanned 32.4-94.2 in treatment. Crucially, failure modes differed across phases: explicit decision errors predominated in anatomical staging (48.2%), whereas decision omissions were more frequent during initial work-up (33.4%) and post-treatment follow-up (30.4%). Varying query granularity revealed a trade-off between decision quality and other clinical requirements. Broad open-ended requests reduced decision quality across all 30 models, whereas more detailed, checklist-like prompts improved decision quality but were accompanied by lower evidence faithfulness and safety adherence. These findings highlight the need to assess clinical reliability against the distinct clinical requirements of each phase along the longitudinal care pathway. Finally, NPC-RubricBench informs the design of stage-aware clinical guardrails and offers a reusable blueprint for evaluating LLMs across other cancers and chronic diseases.

cs.AI↗

Direct Trajectory Optimization for Inference-Time Alignment of Text-to-Image Diffusion Models

Alignment of text-to-image diffusion models with reward functions is a key challenge in controllable generation. Existing approaches typically rely on optimizing model inputs or fine-tuning denoising networks, which requires computationally expensive backpropagation through large denoising models. In this work, we propose Direct Trajectory Optimization (DTO), a training-free framework that performs alignment by directly optimizing the sampling trajectory of diffusion models. Specifically, DTO treats intermediate denoising scores and sampled noise as the optimization variables, eliminating the need for gradients through the denoising network. To mitigate the out-of-distribution problem, we introduce Reward-Optimized Score Guidance, which interpolates between the optimized scores and the pretrained model predictions to stabilize sampling, while providing a provable theoretical guarantee of reward improvement under constrained distributions. Empirically, DTO improves alignment performance across three diffusion models on two benchmarks, outperforming existing inference-time alignment methods while maintaining high efficiency. These results suggest that trajectory-level optimization is a promising direction for efficient and effective diffusion model alignment.

cs.CV↗

Beyond the Flat-Band Paradigm: Open Orbits as a Route to Correlated States in Dispersive Moiré Minibands

The realization of fractional Chern insulators has largely relied on isolated flat bands in complex bilayer or multilayer moiré systems with ideal quantum geometry. Whether such phases can emerge in genuinely dispersive bands, and by what mechanism, remains an open question. Here, we propose a mechanism for magnetic-field-induced strong correlations in the highly dispersive bands of monolayer graphene on hexagonal boron nitride. In the low-field classical-to-quantum crossover regime of the second moiré miniband, we show that the interplay of the moiré potential and weak magnetic fields (B < 1.5 T) profoundly alters electron dynamics. Saddle-point van Hove singularities produce extended open orbits that strongly suppress carrier propagation, potentially favoring a regime in which Coulomb interactions become comparatively more important, while weak magnetic fields lift valley degeneracy and trigonal warping redistributes Berry curvature. High-resolution magnetotransport and temperature-dependent measurements reveal fractional-slope transport features and signatures of interaction-driven incompressibility. Because the Hall conductivity is not fully quantized and the longitudinal resistance remains finite, we describe these observations as incipient FCI behavior: transport signatures consistent with a developing fractional Chern insulating state, without claiming a fully developed FCI phase established by complete Hall quantization and vanishing longitudinal resistance. Our results suggest a possible fermiology-driven route toward correlated topological states in dispersive bands in the classical-to-quantum crossover, distinct from the conventional flat-band scenario.

cond-mat.mes-hall↗

A Hybrid Attention Model Learning Unified Time-aware Patch Representation for Irregular Multivariate Time Series Forecasting

Time series foundation models (TSFMs) have recently delivered impressive zero-shot performance across diverse forecasting tasks. However, real-world decision-making frequently relies on \emph{irregular multivariate time series} (IMTS), where inconsistent inter-observation intervals and asynchronous sampling across variables coexist with informative missingness. Existing TSFMs handle such inputs either through imputation that injects spurious values or through index-based positional encodings that ignore continuous time. There is still a gap in the foundation model that follows the original IMTS patterns. In this paper, we propose a hybrid attention model that learns a unified time-aware patch representation for IMTS forecasting. We first design a \emph{time-aware patch encoding} that maps a variable number of intra-patch timestamps into a fixed-size embedding, producing a uniform format for irregular patches without resorting to imputation. We then introduce a \emph{time bias attention} mechanism that calibrates inter-patch temporal misalignment and asynchronous cross-channel dependencies as auxiliary attention offset. Finally, on top of a decoder-only Transformer backbone, we adopt a \emph{hybrid causal mask} that preserves a bidirectional full view over the historical context while keeping the forecast horizon strictly autoregressive. To support large-scale pretraining under irregular settings, we also curate VersaTSA, an archive of $30$B observations that retains the native sampling sparsity of its sources. Experiments on three IMTS benchmarks and a standard regular-MTS benchmark show that our model achieves state-of-the-art zero-shot performance on IMTS and remains competitive when transferred to regular forecasting.

cs.LG↗

IDProxy: CTR Prediction with Multimodal LLMs for Cold-Start Recommendation at Xiaohongshu

Content-driven platforms such as Xiaohongshu often leverage click-through rate (CTR) prediction models for recommendation. However, these models depend heavily on item ID embeddings, which perform poorly in item cold-start settings. In this paper, we present IDProxy, a production-scale system developed at Xiaohongshu to address this challenge. IDProxy leverages multimodal large language models (MLLMs) to generate proxy embeddings from rich content signals, enabling CTR prediction for new items in the absence of usage data. Through a lightweight coarse-to-fine mechanism, these proxies are aligned with the ID embedding space and trained end-to-end with the ranking model, allowing seamless integration into production-facing pipelines. Extensive offline and online experiments demonstrate the effectiveness of the method, which has been deployed in 2025 in Xiaohongshu's Content Feed and Display Ads features, reaching hundreds of millions of users daily.

cs.IR↗

FibVLA: An Efficient Temporal Vision-Language-Action Model with Fibonacci Sampling

Vision-language-action models (VLAs), which leverage the cognition of multimodal information to infer physical-world actions, provide a generalized solution for embodied AI applications. Conventional VLAs usually concentrate on current digital cognition. While some efforts are made to enhance VLAs' reasoning capabilities by capturing temporal information, encoding the long-context history causes an efficiency-decreasing issue. To reconcile the conflict between capturing temporal information and maintaining inference efficiency in VLAs, this paper introduces FibVLA, an efficient framework featuring temporal perception of long-context history. Specifically, we leverage logarithmic hindsight sampling to both proprioceptive states and visual frames to capture long-term temporal dependencies with minimal redundancy. For the action expert, we introduce the flow matching to produce action distributions, and the Fibonacci recurrent inference strategy to generate long-range planning steps based on real-time closed-loop feedback. Experiments demonstrate that FibVLA significantly improves action smoothness and success rates without retraining large-scale visual encoders. Efficiency analysis demonstrates superior real-time responsiveness compared to video-based baselines in real-world evaluations.

cs.RO↗

Laser power transmission in space: Plasma-based power cell

Laser power beaming offers a route to space energy delivery, but semiconductor laser photovoltaic receivers face thermalization, joule heat, and radiative recombination waste, etc. Here we propose a gas-phase plasma power cell that converts vacuum-ultraviolet photons into electrical output through xenon photoionization and magnetically biased charge separation. Particle-in-cell Monte Carlo simulations of a low-pressure xenon chamber driven by a 58.4 nm pulsed laser predict a steady-state laser-to-electrical conversion efficiency of 82.25% at 2000 W/m2 average incident power. Energy accounting closes to 1%, with 7.52% photon escape, 8.95% boundary loss, and 1.29% chamber-stored energy. Parameter scans over bias voltage, magnetic field, and gas pressure identify photon absorption and electron confinement as controlling design factors. These results are a proof-of-concept gas-phase receiver architecture for space laser power beaming.

physics.plasm-ph↗

LEDOM: Reverse Language Model

Autoregressive language models are trained exclusively left-to-right. We explore the complementary factorization, training right-to-left at scale, and ask what reasoning patterns emerge when a model conditions on future context to predict the past. We train LEDOM, an open-source purely reverse autoregressive language model (2B/7B parameters, 435B tokens), and find it develops capabilities distinct from forward models, including abductive inference, question synthesis, and natural resolution of the reversal curse. We then explore one application of the reverse model: combining forward likelihood $P(y \mid x)$ with reverse posterior $P(x \mid y)$ through noisy channel duality. We propose Reverse Reward, which reranks forward outputs using reverse posterior estimates, and prove that bidirectional scoring penalizes hallucinated reasoning chains whose backward reconstruction degrades. Reverse Reward yields gains of up to 6.6\% on AIME 2024 and 15\% on AMC 2023 across multiple strong baselines. We release all models, code, and data here: https://github.com/Arvid-pku/LEDOM.

cs.CL↗

A Set-Theoretic Approach to Detecting Logic Bugs in DBMS Inner Join Optimizations

The query optimizer is a fundamental component of database management systems that determines the most efficient execution strategy for a given query by evaluating alternative query plans. Among its tasks, join optimization plays a central role, as the order of joins in multi-table queries can significantly affect execution performance. However, due to the inherent complexity of join optimization, logical bugs are inevitable and often difficult to detect. While existing fuzzing tools have shown notable success in uncovering crash- and performance-related errors, effectively identifying logical bugs -- cases in which the system produces incorrect query results -- remains largely unresolved. In this paper, we propose a metamorphic testing approach to detect DBMS bugs related to INNER JOIN optimization through the lens of set theory. For each testing case, equivalent queries are generated based on a basic set operation -- intersection -- and three semantics-preserving transformation rules, i.e., symmetric join transformation, asymmetric difference transformation, and symmetric difference transformation, are introduced. These rules rewrite a simple NATURAL/INNER JOIN query into a more complex, yet semantically equivalent, form. We implement this design in JoinEquiv, which serves as a testing oracle to systematically uncover logical inconsistencies in DBMS query processing by comparing the results of original and transformed queries. Using JoinEquiv, we uncovered 29 previously unknown issues in mainstream DBMSs (MySQL, TiDB, DuckDB, and Percona), and 27 of them were officially confirmed. JoinEquiv reveals deep logical flaws in DBMS optimizers and executors, underscoring its value in enhancing DBMS robustness.

cs.DB↗

ScaleSweep: Accurate NVFP4 Post-Training Quantization of LLMs via Block Scale Initialization

NVFP4 is a recently introduced hardware-supported FP4 format that improves the fidelity of 4-bit quantization through fine-grained block scales. However, existing NVFP4 scale initialization methods still primarily rely on AbsMax initialization, which leaves a noticeable gap to the optimal solution. To address this, we propose ScaleSweep, a simple and efficient scale optimization method that sweeps over feasible block scale candidates and selects the candidate that minimizes a target objective. We further provide a theoretical analysis of NVFP4 quantization and derive both lower and upper bounds for the required sweep range under mean square error (MSE) and weighted mean square error (WMSE) between the original tensor and the quantized reconstructed tensor. The proposed bounds substantially reduce the sweep space while preserving the optimal candidate, enabling negligible overhead compared with the baseline quantization operators. Experiments on Llama and Qwen models demonstrate that ScaleSweep consistently improves quantization performance over existing initialization methods and further narrows the gap to full precision. In particular, under aggressive end-to-end quantization of weights, activations, KV cache, and query states, ScaleSweep preserves more than 93% of the full-precision performance.

cs.LG↗

NeUQI: Near-Optimal Uniform Quantization Parameter Initialization for Low-Bit LLMs

Large language models (LLMs) achieve impressive performance across domains but face significant challenges when deployed on consumer-grade GPUs or personal devices such as laptops, due to high memory consumption and inference costs. Post-training quantization (PTQ) of LLMs offers a promising solution that reduces their memory footprint and decoding latency. In practice, PTQ with uniform quantization representation is favored due to its efficiency and ease of deployment, as uniform quantization is widely supported by mainstream hardware and software libraries. Recent studies on low-bit uniform quantization have led to noticeable improvements in post-quantization model performance; however, they mainly focus on quantization methodologies, while the initialization of quantization parameters remains underexplored and still relies on the conventional Min-Max formula. In this work, we identify the limitations of the Min-Max formula, move beyond its constraints, and propose NeUQI, a method that efficiently determines near-optimal initialization for uniform quantization. Our NeUQI simplifies the joint optimization of the scale and zero-point by deriving the zero-point for a given scale, thereby reducing the problem to a scale-only optimization. Benefiting from the improved quantization parameters, our NeUQI consistently outperforms existing methods in the experiments with the LLaMA and Qwen families on various settings and tasks. Furthermore, when combined with a lightweight distillation strategy, NeUQI even achieves superior performance to PV-tuning, a considerably more resource-intensive method.

cs.LG↗

ImViD: Immersive Volumetric Videos for Enhanced VR Engagement

User engagement is greatly enhanced by fully immersive multi-modal experiences that combine visual and auditory stimuli. Consequently, the next frontier in VR/AR technologies lies in immersive volumetric videos with complete scene capture, large 6-DoF interaction space, multi-modal feedback, and high resolution & frame-rate contents. To stimulate the reconstruction of immersive volumetric videos, we introduce ImViD, a multi-view, multi-modal dataset featuring complete space-oriented data capture and various indoor/outdoor scenarios. Our capture rig supports multi-view video-audio capture while on the move, a capability absent in existing datasets, significantly enhancing the completeness, flexibility, and efficiency of data capture. The captured multi-view videos (with synchronized audios) are in 5K resolution at 60FPS, lasting from 1-5 minutes, and include rich foreground-background elements, and complex dynamics. We benchmark existing methods using our dataset and establish a base pipeline for constructing immersive volumetric videos from multi-view audiovisual inputs for 6-DoF multi-modal immersive VR experiences. The benchmark and the reconstruction and interaction results demonstrate the effectiveness of our dataset and baseline method, which we believe will stimulate future research on immersive volumetric video production.

cs.CV↗

MUSE: Resolving Manifold Misalignment in Visual Tokenization via Topological Orthogonality

Unified visual tokenization faces a fundamental trade-off between high-fidelity pixel reconstruction (spatial equivariance) and semantic abstraction (conceptual invariance). We attribute this conflict to Manifold Misalignment: naive joint optimization induces opposing gradients, creating a zero-sum game between reconstruction and perception. To address this, we propose MUSE, a framework based on Topological Orthogonality. By treating Structure as an orthogonal bridge, MUSE decouples optimization within Transformers: structural gradients refine attention topology, while semantic gradients update feature values. This turns destructive interference into Mutual Reinforcement. Experiments show that MUSE breaks the trade-off, achieving state-of-the-art generation quality (gFID 3.08) and surpassing its teacher InternViT-300M in linear probing (85.2\% vs. 82.5\%), demonstrating that structurally aligned reconstruction can enhance semantic perception. Code is available at https://github.com/PanqiYang1/MUSE.

cs.CV↗

MINOS: A Multimodal Evaluation Model for Bidirectional Generation Between Image and Text

Evaluation is important for multimodal generation tasks, while traditional multimodal evaluation metrics suffer from several limitations. With the rapid progress of MLLMs, there is growing interest in applying MLLMs to build general evaluation systems. However, existing researches often simply collect large-scale evaluation data for training, while overlooking the quality of evaluation data. What's more, current proposed evaluation models often struggle to achieve consistently strong performance across both image-to-text (I2T) and text-to-image (T2I) tasks. In this paper, through rigorous quality control strategies, we construct a comprehensive multimodal evaluation dataset, Minos-57K, with evaluation samples across 15 datasets, for developing the multimodal evaluation model Minos with SFT and preference alignment training strategies. Notably, despite using less than half the scale of the training data of prior work, our model achieves state-of-the-art evaluation performance across 16 out-of-domain datasets covering both I2T and T2I tasks among all open-source multimodal evaluation models and remain competitive with closed-source models. Extensive experiments demonstrate the importance of leveraging quality control process, jointly training on evaluation data from both I2T and T2I generation tasks and further preference alignment.

cs.CL↗

Rippled graphene pores as fluidic memristive devices with synaptic and neuromorphic functionalities

Nanofluidic memristive devices work with nanoscale pores and ions dissolved in water, which harness the ionic memory effect aiming to store and process information. These devices share the same charge carriers as biological systems and bring hope for better emulating the neural functions and developing ionic circuits for neuromorphic applications. Specially, theory and experiments suggest that nanoconfinement is essential for inducing a memory effect, which places limit on the pore size to nm-scale or smaller. Such devices are difficult to scale up with precision and operate with long-term stability. Here, we show that a micrometer size pore, generally expected to exhibit a linear ion transport, can display a pronounced memory effect, if its rim is wrapped by strongly curved and tightly stacked graphene. We attribute the observation to slow ion dynamics confined in the rippled graphene edges. The devices are easy to scale up and integrate into fluidic circuits. The memory effect is ion-selective and exhibits long endurance comparable to the lifetime of synaptic proteins, which enables reversible modification of the conductance states using programmable voltage spikes and various electrolytes over a long time, akin to biological synaptic plasticity. Thanks to this plasticity, our devices and their integrated circuits enable storing, transmitting and processing information with high reliability, fidelity and accuracy, as evidenced in the identification of both greyscale and color images, and in the real-time analysis of emulated neural signals. Our results highlight nanoscale morphology of the pore wall as an important parameter regulating ion transport and indicate that the stringent nanoconfinement for ionic memory can be lifted from restricting the pore size to designing its rim structure. The devices and their integrated circuits may find use in ionic neuromorphic applications.

cond-mat.mtrl-sci↗

LUMOS: Universal Semi-Supervised OCT Retinal Layer Segmentation with Hierarchical Reliable Mutual Learning

Optical Coherence Tomography (OCT) layer segmentation faces challenges due to annotation scarcity and heterogeneous label granularities across datasets. While semi-supervised learning helps alleviate label scarcity, existing methods typically assume a fixed granularity, failing to fully exploit cross-granularity supervision. This paper presents LUMOS, a semi-supervised universal OCT retinal layer segmentation framework based on a Dual-Decoder Network with a Hierarchical Prompting Strategy (DDN-HPS) and Reliable Progressive Multi-granularity Learning (RPML). DDN-HPS combines a dual-branch architecture with a multi-granularity prompting strategy to effectively suppress pseudo-label noise propagation. Meanwhile, RPML introduces region-level reliability weighing and a progressive training approach that guides the model from easier to more difficult tasks, ensuring the reliable selection of cross-granularity consistency targets, thereby achieving stable cross-granularity alignment. Experiments on six OCT datasets demonstrate that LUMOS largely outperforms existing methods and exhibits exceptional cross-domain and cross-granularity generalization capability.

cs.CV↗

Financial Relativity: An Information-Geometric Interpretation of Asset Pricing

Classical asset pricing relies on the risk-neutral measure $Q$ for valuation, yet its economic interpretation is typically anchored in a physical measure $P$. This creates an inherent asymmetry: pricing is governed by $Q$, while meaning resides in $P$, making it difficult to provide a unified account of asset pricing within a single conceptual framework. This paper proposes an alternative perspective based on information geometry, termed Financial Relativity. Its central principle is the relativity of probabilistic reference frames: $P$ and $Q$ have no intrinsic hierarchy, but instead represent geometric structures induced by different informational constraints. Terminal structural information shapes probability geometry, which in turn governs how information is expressed in prices. Within this framework, the risk-neutral measure is reinterpreted as a posterior probability geometry. Asset prices are characterized as geometric projections of terminal payoffs onto information subspaces, and their dynamics reflect the progressive manifestation of structural information under evolving geometry. We develop both discrete and continuous financial field equations to describe the formation of probability geometry and derive geodesic price dynamics in which volatility arises endogenously from posterior uncertainty. The framework provides a unified explanation for price fluctuations, event-driven behavior, and risk premia, and yields testable implications, including structural links between volatility and posterior variance and measures of price informational efficiency. By integrating structural information, probability measures, and price dynamics within a single geometric framework, the paper offers a coherent, extensible, and empirically tractable reinterpretation of asset pricing.

q-fin.MF↗

Universal reconstructive polarimetry with graphene-metal infrared photodetectors

Recent advent of smart photodetectors, where in-situ tuning of responsivity enables the reconstruction of light intensity, polarization and spectrum by a single device, has revolutionized the field of optoelectronics. So far, most such reconstructive detectors were realized with non-scalable technology of van der Waals stacking. Here, we demonstrate the infrared reconstructive polarimetry with photodetectors based on conventional gated graphene-metal junctions. The reconstruction exploits the gate tuning of polarization contrast, which enables the determination of both infrared power and polarization angle from photovoltage measurements at two different gate voltages. The physics enabling the polarimetry lies in polarization-dependent shift of the electron hot spot near the contact, and the gate tuning of photosensitive barrier width. We further show the universality of polarization reconstruction, i.e. its feasibility with different geometries of the junction, and with graphene of different quality, from boron-nitride encapsulated flakes to the scalable chemical vapor deposited films.

cond-mat.mes-hall↗