arXiv subjects

Hongxuan Li

Publications and source records attributed to Hongxuan Li.

11 recordsLinked to original sources

arXiv2026

DrugReason: Dynamic Multi-View Reasoning over Knowledge Graph and Language Evidence for Drug Repurposing

Drug repurposing aims to identify new therapeutic uses for existing compounds and, compared with de novo drug discovery, offers a faster and more cost-effective path to clinical translation. However, the space of candidate drug-disease pairs is enormous and their underlying relationships often depend on complex multi-hop biological mechanisms, making it difficult to reliably predict which pairs represent true therapeutic relationships. Existing approaches tackle this from two directions: knowledge graph-based methods organize curated biomedical evidence into structured relational networks for grounded multi-hop reasoning, while LLM-based methods leverage pretrained knowledge to generate flexible mechanistic rationales. Yet neither is sufficient alone - KGs are confined to observed graph structure while LLMs lack factual grounding and risk hallucination. To address this gap, we propose DrugReason, a multi-view reasoning framework that integrates grounded KG reasoning with LLM-generated mechanistic inference for drug repurposing. DrugReason adaptively routes diverse reasoning paths to specialized experts conditioned on the query context, while a cross-expert distillation objective enables knowledge sharing without sacrificing expert specialization. Experiments on PharmaDB, DDInter, and DrugBank show that DrugReason improves average performance over strong single-view reasoning baselines and achieves competitive or superior results compared with graph-based alternatives, while providing interpretable routing-based predictions.

cs.LG↗

arXiv2026

Towards Precision Therapy in Hepatocellular Carcinoma: A Clinical-Reasoning LLM for Risk Stratification and Treatment Guidance

Hepatocellular carcinoma (HCC) is a common malignancy and a leading cause of cancer-related mortality. Current guidelines and staging systems provide coarse categories, but often miss within-stage heterogeneity and the clinical context in electronic medical records (EMRs). We present HCC-STAR (Hepatocellular Carcinoma Staging, Treatment And pRognosis), a clinically aligned large language model that reads routine EMR narratives and jointly outputs risk score-based staging, ranked guideline-consistent treatments with evidence-based rationales, and individualized survival estimates. We curated about 30,000 HCC cases from SEER and expanded them into EMR-style narrative training data using a clinician-validated, prompt-based augmentation workflow. On this corpus, we developed a knowledge-aligned reasoning framework optimized with a step-verifiable composite reward, moving beyond text-level memorization of clinical guidelines. In a multi-center cohort of 6,668 patients from 12 hospitals in China, HCC-STAR achieved state-of-the-art performance in treatment recommendation and risk stratification compared with clinical guidelines and competitive models, including GPT-5 and Gemini-2.5 Pro. Hypothetical overall-survival analysis showed a median survival of 51 months under adherence to HCC-STAR recommendations, compared with 29 and 32 months under BCLC and CNLC. In clinician-centric evaluations, blinded hepatobiliary specialists rated HCC-STAR's reasoning and evidence-based justifications as trustworthy. The model surpassed resident and attending physicians in treatment accuracy and helped physicians make more accurate decisions faster when used as an assistant. These findings support HCC-STAR as a reliable and verifiable decision-support system for risk stratification and precision therapy in HCC.

cs.AI↗

arXiv2026

Persistent incommensurate amorphous/crystalline meta-interfaces enable engineering-grade superlubricity

Friction dissipates a substantial portion of global energy, motivating the pursuit of superlubricity, a state of near-zero friction, in real-world systems. Conventional approaches rely on crystalline lattice mismatch to suppress periodic energy barriers, but real interfaces invariably contain defects, edges and grain boundaries that restore high-friction states. Here we introduce a materials-agnostic strategy based on amorphous/crystalline heterointerfaces to achieve robust superlubricity under engineering-relevant conditions. Using diamond-like carbon (DLC) and crystalline MoS2 as a model system, we show through experiments and atomistic simulations that their interface remains incommensurate at all orientations and exhibits vanishing energy barriers during friction. In contrast, twisted MoS2 bilayers readily reorient into commensurate, high-friction states. We scale this effect by fabricating laser-patterned arrays of DLC/MoS2 meta-contacts reinforced with Ti3C2Tx MXene, forming hierarchical interfaces that sustain a friction coefficient of ~0.008 over 100000 cycles under combined extreme conditions: millimetre-scale contact size, 12.7 GPa contact pressure and RH 40% air. This unprecedented performance arises from four synergistic factors: intrinsic incommensurability at amorphous/crystalline interface, the rigidity of DLC support, MXene-based mechanical reinforcement and normalized load distribution by geometric patterning. These findings establish a general design paradigm that extends structural superlubricity from nanoscale model systems to practical technologies for sustainable engineering.

cond-mat.mtrl-sci↗

arXiv2026

A biomimetic feedback loop for sustaining self-lubrication and wear resistance

Intelligent materials that self-sense and self-regulate are an emerging frontier in sustainable technology. Here we introduce Cu(Au)/C nanocomposite films that act as bioinspired self-adjusting lubricants. In these films, frictional heating triggers melting and migration of soft metal nanoparticles (NPs) such as Cu or Au along nano-pores to the friction interface, where the metal catalyzes the in-situ formation of ordered carbon nano-structures. Real-time monitoring of friction coefficient, electrical resistance(R), and metal release confirms an autonomous cycle: high friction coefficient generates heat, melting the metal NPs; the migrating metal then lowers friction coefficent by creating low-friction nanostructures, which reduces heat and arrests further migration until friction rises again. This self-limiting feedback enables stable ultra-low friction (~0.04) and an exceptional wear life (>40 km) even in high vacuum. By utilizing friction-derived heat as an intrinsic activation signal, our system establishes a general paradigm for intelligent, self-regulating materials with applications extending beyond tribology.

cond-mat.mtrl-sci↗

arXiv2026

FlashSchNet: Fast and Accurate Coarse-Grained Neural Network Molecular Dynamics

Graph neural network (GNN) potentials such as SchNet improve the accuracy and transferability of molecular dynamics (MD) simulation by learning many-body interactions, but remain slower than classical force fields due to fragmented kernels and memory-bound pipelines that underutilize GPUs. We show that a missing principle is making GNN-MD IO-aware, carefully accounting for reads and writes between GPU high-bandwidth memory (HBM) and on-chip SRAM. We present FlashSchNet, an efficient and accurate IO-aware SchNet-style GNN-MD framework built on four techniques: (1) flash radial basis, which fuses pairwise distance computation, Gaussian basis expansion, and cosine envelope into a single tiled pass, computing each distance once and reusing it across all basis functions; (2) flash message passing, which fuses cutoff, neighbor gather, filter multiplication, and reduction to avoid materializing edge tensors in HBM; (3) flash aggregation, which reformulates scatter-add via CSR segment reduce, reducing atomic writes by a factor of feature dimension and enabling contention-free accumulation in both forward and backward passes; (4) channel-wise 16-bit quantization that exploits the low per-channel dynamic range in SchNet MLP weights to further improve throughput with negligible accuracy loss. On a single NVIDIA RTX PRO 6000, FlashSchNet achieves 1000 ns/day aggregate simulation throughput over 64 parallel replicas on coarse-grained (CG) protein containing 269 beads (6.5x faster than CGSchNet baseline with 80% reduction of peak memory), surpassing classical force fields (e.g. MARTINI) while retaining SchNet-level accuracy and transferability.

cs.LG↗

arXiv2025

Point Cloud Quantization through Multimodal Prompting for 3D Understanding

Vector quantization has emerged as a powerful tool in large-scale multimodal models, unifying heterogeneous representations through discrete token encoding. However, its effectiveness hinges on robust codebook design. Current prototype-based approaches relying on trainable vectors or clustered centroids fall short in representativeness and interpretability, even as multimodal alignment demonstrates its promise in vision-language models. To address these limitations, we propose a simple multimodal prompting-driven quantization framework for point cloud analysis. Our methodology is built upon two core insights: 1) Text embeddings from pre-trained models inherently encode visual semantics through many-to-one contrastive alignment, naturally serving as robust prototype priors; and 2) Multimodal prompts enable adaptive refinement of these prototypes, effectively mitigating vision-language semantic gaps. The framework introduces a dual-constrained quantization space, enforced by compactness and separation regularization, which seamlessly integrates visual and prototype features, resulting in hybrid representations that jointly encode geometric and semantic information. Furthermore, we employ Gumbel-Softmax relaxation to achieve differentiable discretization while maintaining quantization sparsity. Extensive experiments on the ModelNet40 and ScanObjectNN datasets clearly demonstrate the superior effectiveness of the proposed method.

cs.CV↗

arXiv2025

A Probability Space at Inception of Stochastic Process

Recently, progress has been made in the theory of turbulence, which provides a framework on how a deterministic process changes to a stochastic one owing to the change in thermodynamic states. It is well known that, in the framework of Newtonian mechanics, motions are dissipative; however, when subjected to periodic motion, a system can produce nondissipative motions intermittently and subject to resonance. It is in resonance that turbulence occurs in fluid flow, solid vibration, thermal transport, etc. In this, the findings from these physical systems are analyzed in the framework of statistics with their own probability space to establish their compliance to the stochastic process. In particular, a systematic alignment of the inception of the stochastic process with the signed measure theory, signed probability space, and stochastic process was investigated. It was found that the oscillatory load from the dissipative state excited the system and resulted in a quasi-periodic probability density function with the negative probability regimes. In addition, the vectorial nature of the random velocity splits the probability density function along both the positive and negative axes with slight asymmetricity. By assuming that a deterministic process has a probability of 1, we can express the inception of a stochastic process, and the subsequent benefit is that a dynamic fractal falls on the probability density function. Moreover, we leave some questions of inconsistency between the physical system and the measurement theory for future investigation. We believe that the establishment of the probability density function of resonance nondissipative dynamics in contemporary statistics should make many mathematical tools available and the analytical formulas for the random velocity and probability density function can provide a convenient platform for the development of statistics.

nlin.CD↗

arXiv2025

VTD-CLIP: Video-to-Text Discretization via Prompting CLIP

Vision-language models bridge visual and linguistic understanding and have proven to be powerful for video recognition tasks. Existing approaches primarily rely on parameter-efficient fine-tuning of image-text pre-trained models, yet they often suffer from limited interpretability and poor generalization due to inadequate temporal modeling. To address these, we propose a simple yet effective video-to-text discretization framework. Our method repurposes the frozen text encoder to construct a visual codebook from video class labels due to the many-to-one contrastive alignment between visual and textual embeddings in multimodal pretraining. This codebook effectively transforms temporal visual data into textual tokens via feature lookups and offers interpretable video representations through explicit video modeling. Then, to enhance robustness against irrelevant or noisy frames, we introduce a confidence-aware fusion module that dynamically weights keyframes by assessing their semantic relevance via the codebook. Furthermore, our method incorporates learnable text prompts to conduct adaptive codebook updates. Extensive experiments on HMDB-51, UCF-101, SSv2, and Kinetics-400 have validated the superiority of our approach, achieving more competitive improvements over state-of-the-art methods. The code will be publicly available at https://github.com/isxinxin/VTD-CLIP.

cs.CV↗

arXiv2025

A Survey on Mechanistic Interpretability for Multi-Modal Foundation Models

The rise of foundation models has transformed machine learning research, prompting efforts to uncover their inner workings and develop more efficient and reliable applications for better control. While significant progress has been made in interpreting Large Language Models (LLMs), multimodal foundation models (MMFMs) - such as contrastive vision-language models, generative vision-language models, and text-to-image models - pose unique interpretability challenges beyond unimodal frameworks. Despite initial studies, a substantial gap remains between the interpretability of LLMs and MMFMs. This survey explores two key aspects: (1) the adaptation of LLM interpretability methods to multimodal models and (2) understanding the mechanistic differences between unimodal language models and crossmodal systems. By systematically reviewing current MMFM analysis techniques, we propose a structured taxonomy of interpretability methods, compare insights across unimodal and multimodal architectures, and highlight critical research gaps.

cs.LG↗

arXiv2024

Navigating the Dual Facets: A Comprehensive Evaluation of Sequential Memory Editing in Large Language Models

Memory Editing (ME) has emerged as an efficient method to modify erroneous facts or inject new facts into Large Language Models (LLMs). Two mainstream ME methods exist: parameter-modifying ME and parameter-preserving ME (integrating extra modules while preserving original parameters). Regrettably, previous studies on ME evaluation have two critical limitations: (i) evaluating LLMs with single edit only, neglecting the need for continuous editing, and (ii) evaluations focusing solely on basic factual triples, overlooking broader LLM capabilities like logical reasoning and reading understanding. This study addresses these limitations with contributions threefold: (i) We explore how ME affects a wide range of fundamental capabilities of LLMs under sequential editing. Experimental results reveal an intriguing phenomenon: Most parameter-modifying ME consistently degrade performance across all tasks after a few sequential edits. In contrast, parameter-preserving ME effectively maintains LLMs' fundamental capabilities but struggles to accurately recall edited knowledge presented in a different format. (ii) We extend our evaluation to different editing settings, such as layers to edit, model size, instruction tuning, etc. Experimental findings indicate several strategies that can potentially mitigate the adverse effects of ME. (iii) We further explain why parameter-modifying ME damages LLMs from three dimensions: parameter changes after editing, language modeling capability, and the in-context learning capability. Our in-depth study advocates more careful use of ME in real-world scenarios.

cs.CL↗

arXiv2023

THInImg: Cross-modal Steganography for Presenting Talking Heads in Images

Cross-modal Steganography is the practice of concealing secret signals in publicly available cover signals (distinct from the modality of the secret signals) unobtrusively. While previous approaches primarily concentrated on concealing a relatively small amount of information, we propose THInImg, which manages to hide lengthy audio data (and subsequently decode talking head video) inside an identity image by leveraging the properties of human face, which can be effectively utilized for covert communication, transmission and copyright protection. THInImg consists of two parts: the encoder and decoder. Inside the encoder-decoder pipeline, we introduce a novel architecture that substantially increase the capacity of hiding audio in images. Moreover, our framework can be extended to iteratively hide multiple audio clips into an identity image, offering multiple levels of control over permissions. We conduct extensive experiments to prove the effectiveness of our method, demonstrating that THInImg can present up to 80 seconds of high quality talking-head video (including audio) in an identity image with 160x160 resolution.

cs.CV↗