Hao Sun
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Hao Sun has a total of 203 co-authors. Here are the top co-authors by number of shared publications:
- Yang Liu
University of Chinese Academy of Sciences, Beijing, China
26 shared publications Qi Wang
Zhongguancun Academy
7 shared publicationsRaimondo Betti
Department of Civil Engineering and Engineering Mechanics; Columbia University; New York NY 10027 USA
7 shared publicationsYang Liu
7 shared publicationsChengping Rao
6 shared publications- Ji-Rong Wen
Gaoling School of Artificial Intelligence, Renmin University of China, China
6 shared publications
Latest Publications
- Open Access
- 6 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 6069-6080
Continuous-Time Quantum Walks based Graph Neural Network
- 6 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 6069-6080
Open AccessContinuous-Time Quantum Walks based Graph Neural Network
- Open Access
- 5 November 2026
- Published by Taylor & Francis in Intelligent Computing for Engineering
- Vol. 1 (1)
Physical-field reconstruction from sparse observations: when are diffusion models preferable to deterministic regression?
- 5 November 2026
- Published by Taylor & Francis in Intelligent Computing for Engineering
- Vol. 1 (1)
Open AccessPhysical-field reconstruction from sparse observations: when are diffusion models preferable to deterministic regression?
- Open Access
- 10 September 2026
- Published by Oxford University Press (OUP) in National Science Review
- Vol. 13 (19)
Fusing physics and AI enhances accuracy and diagnostic capabilities in air quality forecasting
Accurate air quality forecasts with diagnostic value are pivotal for emission-control policies and public health, yet chemical transport models (CTMs) are hampered by uncertainties in physicochemical parameterizations, whereas AI-only approaches sacrifice interpretability for predictive power. Here, we introduce ProNet, a coupled forecasting framework that embeds observation-constrained neural correctors directly into the governing equations of CTMs. Rather than treating AI as an external postprocessor or a process-replacing surrogate, ProNet integrates trainable networks into physical simulations to systematically reduce process-level uncertainties, suppress error propagation and maintain physical constraints of mass transport. Applied to 2022–23 data across 338 Chinese cities, ProNet consistently outperforms physical CTMs and AI-only baselines, and better generalizes to extreme events beyond the training data. It also surpasses forecaster-adjusted operational forecasts, which often improve upon raw CTM guidance in severe episodes, improving accuracy by 6.1% for PM2.5 and 3.0% for O3. Notably, ProNet provides error bookkeeping, allowing forecasters to quantitatively attribute simulation errors to specific space and time. In-depth analysis of these process errors reveals systematic biases in current CTM parameterizations, providing actionable insights to refine the physical and chemical modules of numerical models. These results highlight the promise of physics–AI coupling for advancing air quality forecasting and deepening our understanding of the complex Earth system.
- 10 September 2026
- Published by Oxford University Press (OUP) in National Science Review
- Vol. 13 (19)
Open AccessFusing physics and AI enhances accuracy and diagnostic capabilities in air quality forecasting
Accurate air quality forecasts with diagnostic value are pivotal for emission-control policies and public health, yet chemical transport models (CTMs) are hampered by uncertainties in physicochemical parameterizations, whereas AI-only approaches sacrifice interpretability for predictive power. Here, we introduce ProNet, a coupled forecasting framework that embeds observation-constrained neural correctors directly into the governing equations of CTMs. Rather than treating AI as an external postprocessor or a process-replacing surrogate, ProNet integrates trainable networks into physical simulations to systematically reduce process-level uncertainties, suppress error propagation and maintain physical constraints of mass transport. Applied to 2022–23 data across 338 Chinese cities, ProNet consistently outperforms physical CTMs and AI-only baselines, and better generalizes to extreme events beyond the training data. It also surpasses forecaster-adjusted operational forecasts, which often improve upon raw CTM guidance in severe episodes, improving accuracy by 6.1% for PM2.5 and 3.0% for O3. Notably, ProNet provides error bookkeeping, allowing forecasters to quantitatively attribute simulation errors to specific space and time. In-depth analysis of these process errors reveals systematic biases in current CTM parameterizations, providing actionable insights to refine the physical and chemical modules of numerical models. These results highlight the promise of physics–AI coupling for advancing air quality forecasting and deepening our understanding of the complex Earth system.
- Open Access
- 8 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 12772-12783
NPSolver: Neural Poisson Solver with Iterative Physics Supervision
Efficiently solving Poisson equations on complex, irregular domains remains a fundamental challenge in scientific computing, as classical iterative solvers often suffer from prohibitive runtime due to ill-conditioned systems. While neural operators offer a fast alternative, they typically rely on large-scale labeled datasets or struggle with unstable training dynamics when using physics-informed residual losses. We propose NPSolver, a neural Poisson solver trained without solution labels via iterative physics supervision. Instead of relying on fully converged numerical solutions or raw PDE residuals, NPSolver utilizes a small number of preconditioned conjugate gradient (PCG) steps to refine its own predictions, providing a more stable and well-scaled training signal. Theoretical analysis confirms that this iterative supervision serves as a well-conditioned error proxy and that a stop-gradient design is essential for optimization stability. To better capture boundary-driven features under mixed boundary conditions, we further introduce the Boundary-Aware Transolver (BA-Transolver ) architecture that explicitly separates interior and boundary tokenization. Extensive evaluations on 2D and 3D irregular geometries demonstrate that NPSolver outperforms both physics-informed and data-driven baselines. Furthermore, a downstream thermal control task highlights the model's capability for conducting efficient and reliable gradient-based boundary control. We will release our codes and data at https://github.com/intell-sci-comput/NPSolver.
- 8 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 12772-12783
Open AccessNPSolver: Neural Poisson Solver with Iterative Physics Supervision
Efficiently solving Poisson equations on complex, irregular domains remains a fundamental challenge in scientific computing, as classical iterative solvers often suffer from prohibitive runtime due to ill-conditioned systems. While neural operators offer a fast alternative, they typically rely on large-scale labeled datasets or struggle with unstable training dynamics when using physics-informed residual losses. We propose NPSolver, a neural Poisson solver trained without solution labels via iterative physics supervision. Instead of relying on fully converged numerical solutions or raw PDE residuals, NPSolver utilizes a small number of preconditioned conjugate gradient (PCG) steps to refine its own predictions, providing a more stable and well-scaled training signal. Theoretical analysis confirms that this iterative supervision serves as a well-conditioned error proxy and that a stop-gradient design is essential for optimization stability. To better capture boundary-driven features under mixed boundary conditions, we further introduce the Boundary-Aware Transolver (BA-Transolver ) architecture that explicitly separates interior and boundary tokenization. Extensive evaluations on 2D and 3D irregular geometries demonstrate that NPSolver outperforms both physics-informed and data-driven baselines. Furthermore, a downstream thermal control task highlights the model's capability for conducting efficient and reliable gradient-based boundary control. We will release our codes and data at https://github.com/intell-sci-comput/NPSolver.
- Open Access
- 8 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 4742-4753
Spectral-Inspired Neural Operator Learning with Limited Data and Unknown Physics
- 8 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 4742-4753
Open AccessSpectral-Inspired Neural Operator Learning with Limited Data and Unknown Physics
- 22 May 2026
- Published by Elsevier in Neurocomputing
- Vol. 696
Generative spatial downscaling of global ocean wind speed profiles via a diffusion model
- 22 May 2026
- Published by Elsevier in Neurocomputing
- Vol. 696
Generative spatial downscaling of global ocean wind speed profiles via a diffusion model
- 20 May 2026
- Published by Elsevier in Computer Physics Communications
- Vol. 326
PhySimNet: Accelerating PDE simulations with physics-integrated learning
- 20 May 2026
- Published by Elsevier in Computer Physics Communications
- Vol. 326
PhySimNet: Accelerating PDE simulations with physics-integrated learning
- Open Access
- 6 May 2026
- Published by Springer Nature in Nature Communications
- Vol. 17 (1)
Siamese foundation models for crystal structure prediction
Predicting crystal structures from chemical compositions is a fundamental challenge in materials discovery, complicated by complex 3D geometries that distinguish it from fields like protein folding. Here, we present Diffusion-based crystAl Omni (DAO), a pretrain-finetune framework for crystal structure prediction integrating two Siamese foundation models: a structure generator and an energy predictor. The generator is pretrained via a two-stage pipeline on a vast dataset of stable and unstable structures, leveraging the predictor to relax unstable configurations and guide the generative sampling. Across two well-known benchmarks, pretraining significantly enhances performance across multiple backbone architectures. Ablation studies confirm that the synergy between the generator and predictor mutually benefits both components. We further validate DAO on three real-world superconductors (Cr 6 Os 2 , Zr 16 Rh 8 O 4 , and Zr 16 Pd 8 O 4 ) typically inaccessible to conventional computation. For Cr 6 Os 2 , DAO achieves a 100% match rate with experimental references and an atomic-position error of 0.0012 under 20-shot generation, performing over 2000 × faster per iteration than DFT-based structure predictors. These compelling results collectively highlight the potential of our approach for advancing materials science research.
- 6 May 2026
- Published by Springer Nature in Nature Communications
- Vol. 17 (1)
Open AccessSiamese foundation models for crystal structure prediction
Predicting crystal structures from chemical compositions is a fundamental challenge in materials discovery, complicated by complex 3D geometries that distinguish it from fields like protein folding. Here, we present Diffusion-based crystAl Omni (DAO), a pretrain-finetune framework for crystal structure prediction integrating two Siamese foundation models: a structure generator and an energy predictor. The generator is pretrained via a two-stage pipeline on a vast dataset of stable and unstable structures, leveraging the predictor to relax unstable configurations and guide the generative sampling. Across two well-known benchmarks, pretraining significantly enhances performance across multiple backbone architectures. Ablation studies confirm that the synergy between the generator and predictor mutually benefits both components. We further validate DAO on three real-world superconductors (Cr 6 Os 2 , Zr 16 Rh 8 O 4 , and Zr 16 Pd 8 O 4 ) typically inaccessible to conventional computation. For Cr 6 Os 2 , DAO achieves a 100% match rate with experimental references and an atomic-position error of 0.0012 under 20-shot generation, performing over 2000 × faster per iteration than DFT-based structure predictors. These compelling results collectively highlight the potential of our approach for advancing materials science research.
- Open Access
- 20 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 1066-1077
Spatiotemporal Graph Learning with Direct Volumetric Information Passing and Feature Enhancement
Data-driven learning of physical systems has attracted significant attention, where many neural models have been developed. In particular, mesh-based graph neural networks (GNNs) have demonstrated considerable potential in modeling spatiotemporal dynamics across arbitrary geometric domains. However, the existing node-edge message-passing and aggregation mechanism in GNNs limits the representation learning capability. In this paper, we propose a dual-module framework, Cell-embedded and Feature-enhanced Graph Neural Network (CeFeGNN), for learning spatiotemporal dynamics. Specifically, we embed learnable cell attributions to the common node-edge message passing process, thereby better capturing the spatial dependency of regional features. Such a strategy essentially upgrades the local aggregation scheme from first order (e.g., from edge to node) to a higher order (e.g., from volume and edge to node), which takes advantage of volumetric information in message passing. Meanwhile, a novel feature-enhanced block is designed to further improve the model's performance and alleviate the over-smoothing problem. Extensive experiments on various PDE systems and a real-world dataset demonstrate that CeFeGNN achieves superior performance compared with other baselines.
- 20 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 1066-1077
Open AccessSpatiotemporal Graph Learning with Direct Volumetric Information Passing and Feature Enhancement
Data-driven learning of physical systems has attracted significant attention, where many neural models have been developed. In particular, mesh-based graph neural networks (GNNs) have demonstrated considerable potential in modeling spatiotemporal dynamics across arbitrary geometric domains. However, the existing node-edge message-passing and aggregation mechanism in GNNs limits the representation learning capability. In this paper, we propose a dual-module framework, Cell-embedded and Feature-enhanced Graph Neural Network (CeFeGNN), for learning spatiotemporal dynamics. Specifically, we embed learnable cell attributions to the common node-edge message passing process, thereby better capturing the spatial dependency of regional features. Such a strategy essentially upgrades the local aggregation scheme from first order (e.g., from edge to node) to a higher order (e.g., from volume and edge to node), which takes advantage of volumetric information in message passing. Meanwhile, a novel feature-enhanced block is designed to further improve the model's performance and alleviate the over-smoothing problem. Extensive experiments on various PDE systems and a real-world dataset demonstrate that CeFeGNN achieves superior performance compared with other baselines.
- Top Cited PapersOpen Access
- 7 March 2026
- Published by Springer Nature in Nature Communications
- Vol. 17 (1)
Evaluating LLMs' divergent thinking capabilities for scientific idea generation with minimal context
Large Language Models (LLMs) demonstrate remarkable capabilities in scientific tasks such as literature analysis and experimental design. For instance, these models excel at accurately extracting key findings from papers or generating coherent experimental procedures. However, existing evaluation benchmarks primarily assess performance using rich contextual inputs. We introduce LiveIdeaBench, a comprehensive benchmark evaluating LLMs’ scientific idea generation by assessing divergent thinking capabilities using single-keyword prompts. Drawing from Guilford’s creativity theory, our benchmark employs a dynamic panel of state-of-the-art LLMs to assess generated ideas across five dimensions: originality, feasibility, fluency, flexibility, and clarity. Through experimentation with over 40 leading models across 1180 keywords spanning 22 scientific domains, we reveal that the scientific idea generation capabilities measured by our benchmark are poorly predicted by standard metrics of general intelligence scores. Our results demonstrate that models like QwQ-32B-preview achieve creative performance comparable to models such as claude-3.7-sonnet: thinking, despite significant gaps in their general intelligence scores. These findings highlight the need for specialized evaluation benchmarks for scientific idea generation and suggest that enhancing these idea generation capabilities in LLMs may require different training strategies than those used for improving general problem-solving abilities. Such strategies could potentially enable a wider range of AI tools tailored for different stages of the scientific process.
- 7 March 2026
- Published by Springer Nature in Nature Communications
- Vol. 17 (1)
Top Cited PapersOpen AccessEvaluating LLMs' divergent thinking capabilities for scientific idea generation with minimal context
Large Language Models (LLMs) demonstrate remarkable capabilities in scientific tasks such as literature analysis and experimental design. For instance, these models excel at accurately extracting key findings from papers or generating coherent experimental procedures. However, existing evaluation benchmarks primarily assess performance using rich contextual inputs. We introduce LiveIdeaBench, a comprehensive benchmark evaluating LLMs’ scientific idea generation by assessing divergent thinking capabilities using single-keyword prompts. Drawing from Guilford’s creativity theory, our benchmark employs a dynamic panel of state-of-the-art LLMs to assess generated ideas across five dimensions: originality, feasibility, fluency, flexibility, and clarity. Through experimentation with over 40 leading models across 1180 keywords spanning 22 scientific domains, we reveal that the scientific idea generation capabilities measured by our benchmark are poorly predicted by standard metrics of general intelligence scores. Our results demonstrate that models like QwQ-32B-preview achieve creative performance comparable to models such as claude-3.7-sonnet: thinking, despite significant gaps in their general intelligence scores. These findings highlight the need for specialized evaluation benchmarks for scientific idea generation and suggest that enhancing these idea generation capabilities in LLMs may require different training strategies than those used for improving general problem-solving abilities. Such strategies could potentially enable a wider range of AI tools tailored for different stages of the scientific process.