arXiv subjects

Lan Ma

Publications and source records attributed to Lan Ma.

At least 19 recordsLinked to original sources

arXiv2026

LLaVAFlow: Preserving Latent Alignment Flow for Parameter-Efficient Multimodal Fine-Tuning

While Multimodal Large Language Models (MLLMs) exhibit strong generalization, visual instruction tuning for downstream tasks inevitably causes catastrophic forgetting, impairing overall generalization. While existing methods regulate weight updates to reduce forgetting, they overlook the fundamental cross-modal alignment in MLLMs. Based on prior work and our observations, we argue that cross-modal alignment is implicitly captured in the information-compression trajectory. To preserve the alignment flow embedded in the trajectory, we propose LLaVAFlow, an information-theoretic distillation framework. First, we compress the mutual information between the extracted relations and MLLM embeddings, encouraging a learnable module to produce a refined alignment flow that benefits downstream tasks. Second, we maximize the mutual information between the extracted alignment flows of the pretrained and fine-tuned MLLMs, enabling the transfer of compact alignment information. Extensive experiments show that LLaVAFlow is an effective plug-and-play framework that preserves alignment flow and enhances both downstream performance and generalization.

cs.CV↗

arXiv2026

STAR: Structured Tokenization and Target-Aware Interest Representation for PCVR Prediction

Post-click conversion rate (PCVR) prediction is a core ranking task in industrial recommender systems. Modern ranking models must jointly capture heterogeneous non-sequential features, multi-behavior user sequences, and target-item-aware user interests, while remaining robust to high-cardinality sparse features, missing values, and train-inference inconsistencies. In this paper, we present STAR (Structured Tokenization and Target-Aware Interest Representation), a practical framework for the KDD Cup 2026 Tencent UniRec Challenge. STAR combines structured feature tokenization with target-aware interest representation on top of a HyFormer-style multi-sequence backbone. It introduces high-cardinality signal recovery, explicit user-item interaction tokens, target-aware sequence decoding, and a weighted user-item contrastive auxiliary objective inspired by InfoNCE. We further align the training and inference pipelines by reconstructing feature remapping tables and structural hyperparameters from the saved training configuration. Experiments on the challenge dataset identify the components that most reliably improve ranking AUC, while LogLoss is reported as a calibration diagnostic. The main ablation study shows a large gain from temporal context, with smaller but useful contributions from contrastive alignment, target-aware interest encoding, and high-cardinality sequence feature recovery.

cs.IR↗

arXiv2026

EyeMVP: OCT-Informed Fundus Representation Learning via Paired CFP--OCT Pretraining

Color fundus photography (CFP) is the mainstay of large-scale retinal screening, but its diagnostic capacity is limited by the lack of depth-resolved structure, which optical coherence tomography (OCT) provides yet is less accessible at population scale. We present EyeMVP, a cross-modal retinal foundation model that uses paired CFP--OCT pretraining to learn OCT-informed CFP representations while requiring only CFP at inference. Pretrained on 674,893 same-eye same-day CFP--OCT triples from 112,642 patients across eight hospitals, EyeMVP uses cross-modal masked reconstruction to enrich CFP features with OCT-associated supervision, and combines source-constrained cross-attention with CFP-derived structural masks to accommodate the non-aligned geometry of en-face CFP and cross-sectional OCT. Across 15 dataset-level settings spanning classification and segmentation, under both full-data and few-shot regimes, EyeMVP performs on par with or better than representative retinal foundation models, with consistent gains on macular and optic-nerve tasks; it attains AUROCs of 0.923 for macular edema and 0.867 for myopic macular schisis, two conditions poorly resolved in CFP. In an exploratory reader study, EyeMVP surpasses junior and intermediate ophthalmologists but not seniors on macular edema, while exceeding all groups on myopic macular schisis. These results indicate that cross-modal reconstruction can enrich CFP representations with OCT-associated supervision, offering a practical route to stronger CFP-based screening.

cs.CV↗

arXiv2026

Continuous Language Diffusion as a Decoder-Interface Problem

Gaussian-corrupted sentence embeddings have no direct linguistic interpretation, yet continuous diffusion language models can generate fluent text from them. We study this puzzle through Embedded Language Flows (ELF) and identify a decoder-basin mechanism: our evidence suggests that denoising becomes reliable when trajectories reach regions where the native decoder can read stable tokens. We introduce a diagnostic protocol for denoisability, semantic recoverability, order sensitivity, decoder compatibility, and trajectory reliability. It exposes failures hidden by scalar metrics: low mean-squared error can discard linguistic content, low perplexity can reflect low-entropy collapse, and clean latent reconstruction can coexist with a narrow decoder basin. A decoder-margin bound explains why token recovery depends on margin and local decoder sensitivity, not latent error alone. Auditing public ELF checkpoints reveals an interface phase diagram: early predictions are weakly readable, mid-trajectory disagreement marks a competition region, and late predictions enter a high-margin decoder basin. Once inside, token realization is surprisingly simple on generated ELF states: frozen T5 (Text-to-Text Transfer Transformer) token-embedding lookup recovers $93$--$96\%$ of native decoder decisions, and a single linear readout reaches $97.9\%$ agreement at 32k samples, leaving an $\approx1.1$--$1.2$ perplexity gap in a structured residual tail. Under conservative held-out gates, a margin rule exits roughly $17$--$28\%$ earlier in denoising steps under an explicit diagnostic monitor. Boundary checks on LangFlow, BitstreamDiffusion, and the Continuous Latent Diffusion Language Model (Cola-DLM) show that the same interface questions remain meaningful when the state object and decoder change. Continuous and latent diffusion language models should therefore be evaluated as representation-decoder systems.

cs.CL↗

arXiv2026

Auditing Training-Free 3D Shape Retrieval with Diffused Geodesic Moments

Reported retrieval scores for training-free shape descriptors conflate local signal design, normalization, aggregation, codebook fitting, and metric choices, making isolated component evaluation difficult. This paper reframes descriptor evaluation as a {\em protocol audit}. We introduce Diffused Geodesic Moments (DGM), a seed-conditioned descriptor that computes sparse implicit heat responses, converts them to distance-like fields, and summarizes each vertex by low-order moments across seeds and scales. DGM is used both as a practical non-spectral baseline and as an instrument for isolating protocol effects. On the registered FAUST benchmark split (FAUST-Reg) and the TOSCA shape collection, aggregation-matched experiments show that an independent Geometric Moment Shape Descriptor baseline built on Heat Kernel Signature features (GMSD-HKS) obtains the highest scores in this implementation ($0.621/0.820$ and $0.865/0.963$ mean average precision (mAP)/top-1), Wave Kernel Signature (WKS) remains a strong classical signal, and DGM is useful mainly when sparse solves, non-spectral deployment, or symmetry-informative seed frames are priorities. The broader finding is methodological: the input field and aggregation protocol can dominate the moment formula. The paper contributes a reproducible protocol-cascade analysis, a cross-shape alignment diagnostic for functional-map compatibility, and concrete recommendations for designing and reporting training-free shape descriptors.

cs.CV↗

arXiv2026

Prime Certificates for Exact Vertex-Coprime Ramsey Numbers

Let $G_n$ be the coprime graph on $\{1,\ldots,n\}$. We prove that the mixed vertex-coloring coprime Ramsey number satisfies \[ \Rcop(k_1,\ldots,k_c)=p_{\sum_{i=1}^c(k_i-1)}, \] where $p_m$ is the $m$-th prime. The proof is elementary: the prime clique $\{1\}\cup\{p\le n:p\text{ prime}\}$ gives the upper bound by pigeonhole, while a prime-bin partition gives the matching lower bound by coloring each composite with a bin containing one of its prime divisors. We reserve $\Rcop$ for this vertex-coloring parameter; the edge-coloring parameter on the same host graph is denoted $\Redge$. The same certificate viewpoint yields several extensions, including a support-disjointness generalization, a polynomial-time certificate-extraction primitive, and an exact reduction of the edge-coloring variant to classical Ramsey numbers: $\Redge(k_1,\ldots,k_c)=p_{\Rcl(k_1,\ldots,k_c)-1}$. These two formulas are rank transfers from the same clique-label certificate. We also prove that the balanced two-color diagonal threshold equals the unrestricted threshold $p_{2k-2}$ for all $k\ge2$, via a deterministic prime-bin split requiring only the weak inequality $2p_m<p_{2m}<3p_m$; for fixed $c$, a Hall argument plus a standard Selberg--Delange estimate gives eventual multicolor balanced certificates.

math.CO↗

arXiv2026

DepthArb: Training-Free Depth-Arbitrated Generation for Occlusion-Robust Image Synthesis

Text-to-image models often struggle to synthesize correct occlusion relationships among multiple objects, especially in densely overlapping regions. Many training-free layout-guided methods enforce 2D spatial constraints but do not explicitly resolve depth-dependent attention competition, which can cause concept mixing and implausible occlusion. To address this problem, we propose DepthArb, a training-free framework that formulates occlusion generation as attention arbitration within a unified denoising trajectory. DepthArb employs two core occlusion-control mechanisms: Attention Arbitration Modulation suppresses background-object attention within foreground support according to relative depth, while Spatial Compactness Control limits attention dispersion to preserve object coherence. Because interference varies during generation, Occlusion Conflict Estimation constructs a shared spatial conflict field to adaptively weight both objectives. Through a unified spatial-text attention interface, DepthArb operates on U-Net cross-attention and the image-to-text component of MMDiT joint attention without model retraining. We further introduce OcclBench, a benchmark with continuous relative-depth specifications and occlusion-specific evaluation metrics. Experiments on OcclBench and public benchmarks show that DepthArb improves several layout and occlusion metrics over the evaluated baselines while maintaining competitive text-image alignment.

cs.CV↗

arXiv2026

Enabling Real-Time Colonoscopic Polyp Segmentation on Commodity CPUs via Ultra-Lightweight Architecture

Real-time polyp segmentation is essential for early colorectal cancer detection, yet clinical deployment remains blocked by GPU dependency. We introduce the UltraSeg family, a set of CPU-native segmentation models operating below 0.3M parameters. UltraSeg-108K (0.108M) establishes the extreme-compression frontier, while UltraSeg-130K (0.130M) integrates cross-layer lightweight fusion for enhanced multi-center generalization. The architecture replaces parameter-heavy components with grouped multi-rate dilated convolutions and attention-gated cross-layer fusion, achieving real-time throughput on a single CPU core (exceeding 50 FPS at 256*256 and 30 FPS at 352*352) without sacrificing clinical-grade accuracy. Evaluated on seven public datasets, UltraSeg-130K attains Dice scores exceeding 0.8 at both resolutions, substantially outperforming all existing sub-0.3M competitors. Notably, it approaches or exceeds UNet-Medium (7.76M parameters) on zero-shot external validations while using only 1.7% of its parameters, establishing the first strong baseline for CPU-native real-time polyp segmentation. When scaled to 4.38M parameters, UltraSeg achieves accuracy competitive with heavyweight state-of-the-art models while maintaining an order-of-magnitude parameter advantage, demonstrating that the proposed design principles yield intrinsic representational gains across the entire efficiency spectrum. By delivering the first clinically deployable, CPU-native real-time solution, this work provides an immediately usable tool for resource-limited settings and a reproducible blueprint for real-time medical AI beyond endoscopy. Source code is publicly available.

cs.CV↗

arXiv2025

InfoCLIP: Bridging Vision-Language Pretraining and Open-Vocabulary Semantic Segmentation via Information-Theoretic Alignment Transfer

Recently, the strong generalization ability of CLIP has facilitated open-vocabulary semantic segmentation, which labels pixels using arbitrary text. However, existing methods that fine-tune CLIP for segmentation on limited seen categories often lead to overfitting and degrade the pretrained vision-language alignment. To stabilize modality alignment during fine-tuning, we propose InfoCLIP, which leverages an information-theoretic perspective to transfer alignment knowledge from pretrained CLIP to the segmentation task. Specifically, this transfer is guided by two novel objectives grounded in mutual information. First, we compress the pixel-text modality alignment from pretrained CLIP to reduce noise arising from its coarse-grained local semantic representations learned under image-text supervision. Second, we maximize the mutual information between the alignment knowledge of pretrained CLIP and the fine-tuned model to transfer compact local semantic relations suited for the segmentation task. Extensive evaluations across various benchmarks validate the effectiveness of InfoCLIP in enhancing CLIP fine-tuning for open-vocabulary semantic segmentation, demonstrating its adaptability and superiority in asymmetric transfer.

cs.CV↗

arXiv2025

Generic Construction of Optimal-Access Binary MDS Array Codes with Smaller Sub-packetization

A $(k+r,k,l)$ binary array code of length $k+r$, dimension $k$, and sub-packetization $l$ is composed of $l\times(k+r)$ matrices over $\mathbb{F}_2$, with every column of the matrix stored on a separate node in the distributed storage system and viewed as a coordinate of the codeword. It is said to be maximum distance separable (MDS) if any $k$ out of $k+r$ coordinates suffice to reconstruct the whole codeword. The repair problem of binary MDS array codes has drawn much attention, particularly for single-node failures. In this paper, given an arbitrary binary MDS array code with sub-packetization $m$ as the base code, we propose two generic approaches (Generic Construction I and II) for constructing binary MDS array codes with optimal access (or repair) bandwidth for single-node failures. For every $s\leq r$, a $(k+r,k,ms^{\lceil \frac{k+r}{s}\rceil})$ code $\mathcal{C}_1$ with optimal access bandwidth can be constructed by Generic Construction I. Repairing a failed node of $\mathcal{C}_1$ requires connecting to $d = k+s-1$ helper nodes, in which $s-1$ helper nodes are designated and $k$ are free to select. $\mathcal{C}_1$ generally achieves smaller sub-packetization and provides greater flexibility in the selection of its coefficient matrices. For even $r\geq4$ and $s=\frac{r}{2}$ such that $s+1$ divides $k+r$, a $(k+r, k,ms^{\frac{k+r}{s+1}})$ code $\mathcal{C}_2$ with optimal repair bandwidth can be constructed by Generic Construction II, with $\frac{s}{s+1}(k+r)$ out of $k+r$ nodes having the optimal access property. To the best of our knowledge, $\mathcal{C}_2$ possesses the smallest sub-packetization among existing binary MDS array codes with optimal repair bandwidth known to date.

cs.IT↗

arXiv2025

TINC: Trusted Intelligent NetChain

Blockchain technology facilitates the development of decentralized systems that ensure trust and transparency without the need for expensive centralized intermediaries. However, existing blockchain architectures particularly consortium blockchains face critical challenges related to scalability and efficiency. State sharding has emerged as a promising approach to enhance blockchain scalability and performance. However, current shard-based solutions often struggle to guarantee fair participation and a balanced workload distribution among consortium members. To address these limitations, we propose Trusted Intelligent NetChain (TINC), a multi-plane sharding architecture specifically designed for consortium blockchains. TINC incorporates intelligent mechanisms for adaptive node assignment and dynamic workload balancing, enabling the system to respond effectively to changing network conditions while maintaining equitable shard utilization. By decoupling the control and data planes, TINC allows control nodes to focus on consensus operations, while data nodes handle large-scale storage, thus improving overall resource efficiency. Extensive experimental evaluation and formal analysis demonstrate that TINC significantly outperforms existing shard-based blockchain frameworks. It achieves higher throughput, lower latency, balanced node and transaction distributions, and reduced transaction failure rates. Furthermore, TINC maintains essential blockchain security guarantees, exhibiting resilience against Byzantine faults and dynamic network environments. The integration of Dynamic Decentralized Identifiers (DDIDs) further strengthens trust and security management within the consortium network.

cs.NI↗

arXiv2025

LLaVA-RE: Binary Image-Text Relevancy Evaluation with Multimodal Large Language Model

Multimodal generative AI usually involves generating image or text responses given inputs in another modality. The evaluation of image-text relevancy is essential for measuring response quality or ranking candidate responses. In particular, binary relevancy evaluation, i.e., ``Relevant'' vs. ``Not Relevant'', is a fundamental problem. However, this is a challenging task considering that texts have diverse formats and the definition of relevancy varies in different scenarios. We find that Multimodal Large Language Models (MLLMs) are an ideal choice to build such evaluators, as they can flexibly handle complex text formats and take in additional task information. In this paper, we present LLaVA-RE, a first attempt for binary image-text relevancy evaluation with MLLM. It follows the LLaVA architecture and adopts detailed task instructions and multimodal in-context samples. In addition, we propose a novel binary relevancy data set that covers various tasks. Experimental results validate the effectiveness of our framework.

cs.CV↗

arXiv2025

Court of LLMs: Evidence-Augmented Generation via Multi-LLM Collaboration for Text-Attributed Graph Anomaly Detection

The natural combination of intricate topological structures and rich textual information in text-attributed graphs (TAGs) opens up a novel perspective for graph anomaly detection (GAD). However, existing GAD methods primarily focus on designing complex optimization objectives within the graph domain, overlooking the complementary value of the textual modality, whose features are often encoded by shallow embedding techniques, such as bag-of-words or skip-gram, so that semantic context related to anomalies may be missed. To unleash the enormous potential of textual modality, large language models (LLMs) have emerged as promising alternatives due to their strong semantic understanding and reasoning capabilities. Nevertheless, their application to TAG anomaly detection remains nascent, and they struggle to encode high-order structural information inherent in graphs due to input length constraints. For high-quality anomaly detection in TAGs, we propose CoLL, a novel framework that combines LLMs and graph neural networks (GNNs) to leverage their complementary strengths. CoLL employs multi-LLM collaboration for evidence-augmented generation to capture anomaly-relevant contexts while delivering human-readable rationales for detected anomalies. Moreover, CoLL integrates a GNN equipped with a gating mechanism to adaptively fuse textual features with evidence while preserving high-order topological information. Extensive experiments demonstrate the superiority of CoLL, achieving an average improvement of 13.37% in AP. This study opens a new avenue for incorporating LLMs in advancing GAD.

cs.LG↗

arXiv2025

mRNA Design and Optimization with Deep Knowledge-Infused Approach

The mRNA optimization is essential for mRNA vaccines, therapies, and industrial protein production. Based on current explorations, an ideal optimization approach should simultaneously (i) prevent unintended amino-acid changes, (ii) optimize multiple, biologically relevant objectives, and (iii) retain computational efficiency. However, existing methods are forced to trade off between these perspectives, forming an "impossible triangle." We present RNop, a knowledge-infused Transformer that integrates mechanism-aligned losses to address this problem. By encoding biological prior knowledge in losses, RNop makes knowledge infusion explicit and controllable across optimization focus. Trained on over 6 million sequences, in silico analyses show RNop resolves the "impossible triangle" of mRNA optimization with absolute sequence fidelity, significantly improved biological metrics, and high throughput. In in vitro validation, it can deliver up to 2.28-fold expression gain. Ablation studies reveal how each prior contributes to targeted improvements, yielding mechanism-level interpretability. RNop represents a shift in mRNA optimization methodology: by infusing explicit and interpretable knowledge, the "black-box" mRNA design can be transformed into a predictable, explainable engineering problem. RNop is designed as an extensible platform: additional biological priors can be incorporated as modular, mechanism-aligned loss functions, enabling future development and adaptation to related sequence design problems.

q-bio.QM↗

arXiv2025

Multitemporal Latent Dynamical Framework for Hyperspectral Images Unmixing

Multitemporal hyperspectral unmixing can capture dynamical evolution of materials. Despite its capability, current methods emphasize variability of endmembers while neglecting dynamics of abundances, which motivates our adoption of neural ordinary differential equations to model abundances temporally. However, this motivation is hindered by two challenges: the inherent complexity in defining, modeling and solving problem, and the absence of theoretical support. To address above challenges, in this paper, we propose a multitemporal latent dynamical (MiLD) unmixing framework by capturing dynamical evolution of materials with theoretical validation. For addressing multitemporal hyperspectral unmixing, MiLD consists of problem definition, mathematical modeling, solution algorithm and theoretical support. We formulate multitemporal unmixing problem definition by conducting ordinary differential equations and developing latent variables. We transfer multitemporal unmixing to mathematical model by dynamical discretization approaches, which describe the discreteness of observed sequence images with mathematical expansions. We propose algorithm to solve problem and capture dynamics of materials, which approximates abundance evolution by neural networks. Furthermore, we provide theoretical support by validating the crucial properties, which verifies consistency, convergence and stability theorems. The major contributions of MiLD include defining problem by ordinary differential equations, modeling problem by dynamical discretization approach, solving problem by multitemporal unmixing algorithm, and presenting theoretical support. Our experiments on both synthetic and real datasets have validated the utility of our work

eess.IV↗

arXiv2025

AI-driven Prediction of Insulin Resistance in Normal Populations: Comparing Models and Criteria

Insulin resistance (IR) is a key precursor to diabetes and a significant risk factor for cardiovascular disease. Traditional IR assessment methods require multiple blood tests. We developed a simple AI model using only fasting blood glucose to predict IR in non-diabetic populations. Data from the NHANES (1999-2020) and CHARLS (2015) studies were used for model training and validation. Input features included age, gender, height, weight, blood pressure, waist circumference, and fasting blood glucose. The CatBoost algorithm achieved AUC values of 0.8596 (HOMA-IR) and 0.7777 (TyG index) in NHANES, with an external AUC of 0.7442 for TyG. For METS-IR prediction, the model achieved AUC values of 0.9731 (internal) and 0.9591 (external), with RMSE values of 3.2643 (internal) and 3.057 (external). SHAP analysis highlighted waist circumference as a key predictor of IR. This AI model offers a minimally invasive and effective tool for IR prediction, supporting early diabetes and cardiovascular disease prevention.

cs.LG↗

arXiv2025

CardioLive: Empowering Video Streaming with Online Cardiac Monitoring

Online Cardiac Monitoring (OCM) emerges as a compelling enhancement for the next-generation video streaming platforms. It enables various applications including remote health, online affective computing, and deepfake detection. Yet the physiological information encapsulated in the video streams has been long neglected. In this paper, we present the design and implementation of CardioLive, the first online cardiac monitoring system in video streaming platforms. We leverage the naturally co-existed video and audio streams and devise CardioNet, the first audio-visual network to learn the cardiac series. It incorporates multiple unique designs to extract temporal and spectral features, ensuring robust performance under realistic video streaming conditions. To enable the Service-On-Demand online cardiac monitoring, we implement CardioLive as a plug-and-play middleware service and develop systematic solutions to practical issues including changing FPS and unsynchronized streams. Extensive experiments have been done to demonstrate the effectiveness of our system. We achieve a Mean Square Error (MAE) of 1.79 BPM error, outperforming the video-only and audio-only solutions by 69.2% and 81.2%, respectively. Our CardioLive service achieves average throughputs of 115.97 and 98.16 FPS when implemented in Zoom and YouTube. We believe our work opens up new applications for video stream systems. We will release the code soon.

cs.HC↗

arXiv2024

Acquire Precise and Comparable Fundus Image Quality Score: FTHNet and FQS Dataset

The retinal fundus images are utilized extensively in the diagnosis, and their quality can directly affect the diagnosis results. However, due to the insufficient dataset and algorithm application, current fundus image quality assessment (FIQA) methods are not powerful enough to meet ophthalmologists` demands. In this paper, we address the limitations of datasets and algorithms in FIQA. First, we establish a new FIQA dataset, Fundus Quality Score(FQS), which includes 2246 fundus images with two labels: a continuous Mean Opinion Score varying from 0 to 100 and a three-level quality label. Then, we propose a FIQA Transformer-based Hypernetwork (FTHNet) to solve these tasks with regression results rather than classification results in conventional FIQA works. The FTHNet is optimized for the FIQA tasks with extensive experiments. Results on our FQS dataset show that the FTHNet can give quality scores for fundus images with PLCC of 0.9423 and SRCC of 0.9488, significantly outperforming other methods with fewer parameters and less computation complexity.We successfully build a dataset and model addressing the problems of current FIQA methods. Furthermore, the model deployment experiments demonstrate its potential in automatic medical image quality control. All experiments are carried out with 10-fold cross-validation to ensure the significance of the results.

eess.IV↗