arXiv subjects

Shizhong Yang

Publications and source records attributed to Shizhong Yang.

3 recordsLinked to original sources

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↗

arXiv2025

AllSpark: A Multimodal Spatio-Temporal General Intelligence Model with Ten Modalities via Language as a Reference Framework

Leveraging multimodal data is an inherent requirement for comprehending geographic objects. However, due to the high heterogeneity in structure and semantics among various spatio-temporal modalities, the joint interpretation of multimodal spatio-temporal data has long been an extremely challenging problem. The primary challenge resides in striking a trade-off between the cohesion and autonomy of diverse modalities. This trade-off becomes progressively nonlinear as the number of modalities expands. Inspired by the human cognitive system and linguistic philosophy, where perceptual signals from the five senses converge into language, we introduce the Language as Reference Framework (LaRF), a fundamental principle for constructing a multimodal unified model. Building upon this, we propose AllSpark, a multimodal spatio-temporal general artificial intelligence model. Our model integrates ten different modalities into a unified framework. To achieve modal cohesion, AllSpark introduces a modal bridge and multimodal large language model (LLM) to map diverse modal features into the language feature space. To maintain modality autonomy, AllSpark uses modality-specific encoders to extract the tokens of various spatio-temporal modalities. Finally, observing a gap between the model's interpretability and downstream tasks, we designed modality-specific prompts and task heads, enhancing the model's generalization capability across specific tasks. Experiments indicate that the incorporation of language enables AllSpark to excel in few-shot classification tasks for RGB and point cloud modalities without additional training, surpassing baseline performance by up to 41.82\%. The source code is available at https://github.com/GeoX-Lab/AllSpark.

cs.AI↗

arXiv2024

Generalized Representative Structures for Atomistic Systems

A new method is presented to generate atomic structures that reproduce the essential characteristics of arbitrary material systems, phases, or ensembles. Previous methods allow one to reproduce the essential characteristics (e.g. chemical disorder) of a large random alloy within a small crystal structure. The ability to generate small representations of random alloys, with the restriction to crystal systems, results from using the fixed-lattice cluster correlations to describe structural characteristics. A more general description of the structural characteristics of atomic systems is obtained using complete sets of atomic environment descriptors. These are used within for generating representative atomic structures without restriction to fixed lattices. A general data-driven approach is provided utilizing the atomic cluster expansion(ACE) basis. The N-body ACE descriptors are a complete set of atomic environment descriptors that span both chemical and spatial degrees of freedom and are used within for describing atomic structures. The generalized representative structure(GRS) method presented within generates small atomic structures that reproduce ACE descriptor distributions corresponding to arbitrary structural and chemical complexity. It is shown that systematically improvable representations of crystalline systems on fixed parent lattices, amorphous materials, liquids, and ensembles of atomic structures may be produced efficiently through optimization algorithms. We highlight reduced representations of atomistic machine-learning training datasets that contain similar amounts of information and small 40-72 atom representations of liquid phases. The ability to use GRS methodology as a driver for informed novel structure generation is also demonstrated. The advantages over other data-driven methods and state-of-the-art methods restricted to high-symmetry systems are highlighted.

cond-mat.mtrl-sci↗