SWSong Wang

Song Wang

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Song Wang has a total of 204 co-authors. Here are the top co-authors by number of shared publications:

  • Jun Jiang

    University of Science and Technology of China

    23 shared publications
  • Donglai Zhou

    State Key Laboratory of Precision and Intelligent Chemistry University of Science and Technology of China Hefei Anhui China

    9 shared publications
  • Guokun Yang

    State Key Laboratory of Precision and Intelligent Chemistry University of Science and Technology of China Hefei Anhui China

    6 shared publications
  • Shuo Feng

    State Key Laboratory of Precision and Intelligent Chemistry University of Science and Technology of China Hefei Anhui China

    6 shared publications
  • Yan Huang

    State Key Laboratory of Precision and Intelligent Chemistry University of Science and Technology of China Hefei Anhui China

    6 shared publications
  • Yi Feng

    State Key Laboratory of Precision and Intelligent Chemistry University of Science and Technology of China Hefei Anhui China

    6 shared publications

Latest Publications

  • Generation of Tailorable Multi‐Property Molecules via Multimodal Spectral Fusion Evolution

    by Shijie Tao, Meng Huang, Ledu Wang, Shuo Feng, Yulan Han, Jing He, Yan Huang, Zijin Jia, Donglai Zhou, Yi Feng, et al.

    Molecular generation stands at the forefront of intelligent molecular design, promising transformative advances in the discovery of molecules that meet drug‐likeness or catalysis requirements. However, existing generative models typically focus on optimizing a single target property at a time, hindered by the intrinsic difficulty of simultaneously encoding and evolving multiple structural features. Herein, we developed MUltimodal Spectral fusion Evolution (MUSE)—a spectroscopy‐guided generative framework that unifies multimodal spectroscopic descriptors as a universal chemical language for multi‐property molecular creation. A Spectra‐to‐Molecule (S2M) model, as one of MUSE's core modules, achieved a high success rate in molecular structure generation by training on a self‐built density functional theory (DFT) level multimodal spectral database. Leveraging the continuously variable nature of spectral descriptors, MUSE achieved synergistic integration of structural, vibrational, and electronic information by combining complementary spectroscopic modalities. Through evolutionary algorithms that constrain the generation direction of multi‐property molecules, the framework enables up to 700‐fold enrichment of rare quadruple‐property candidates. This work provides a scalable pathway for multi‐objective discovery in chemistry and materials science.

  • Generation of Tailorable Multi‐Property Molecules via Multimodal Spectral Fusion Evolution

    by Shijie Tao, Meng Huang, Ledu Wang, Shuo Feng, Yulan Han, Jing He, Yan Huang, Zijin Jia, Donglai Zhou, Yi Feng, et al.

    Molecular generation stands at the forefront of intelligent molecular design, promising transformative advances in the discovery of molecules that meet drug‐likeness or catalysis requirements. However, existing generative models typically focus on optimizing a single target property at a time, hindered by the intrinsic difficulty of simultaneously encoding and evolving multiple structural features. Herein, we developed MUltimodal Spectral fusion Evolution (MUSE)—a spectroscopy‐guided generative framework that unifies multimodal spectroscopic descriptors as a universal chemical language for multi‐property molecular creation. A Spectra‐to‐Molecule (S2M) model, as one of MUSE's core modules, achieved a high success rate in molecular structure generation by training on a self‐built density functional theory (DFT) level multimodal spectral database. Leveraging the continuously variable nature of spectral descriptors, MUSE achieved synergistic integration of structural, vibrational, and electronic information by combining complementary spectroscopic modalities. Through evolutionary algorithms that constrain the generation direction of multi‐property molecules, the framework enables up to 700‐fold enrichment of rare quadruple‐property candidates. This work provides a scalable pathway for multi‐objective discovery in chemistry and materials science.

  • Open Access

    Bridging the Theory–Experiment Gap in Vibrational Spectra for Molecular Structure Elucidation

    by Shijie Tao, Ledu Wang, Yi Feng, Song Wang, Wei Hu, Jun Jiang, Meng Huang

    Structure elucidation from experimental vibrational spectra is a fundamental task in chemistry, yet data-driven approaches remain limited by scarce paired experimental data and systematic deviations between computed and measured spectra. Here, we bridge this theory–experiment gap through two coupled advances. First, we introduce Spec2Spec, a U-Net translator that fuses theoretical vibrational spectra with three-dimensional molecular embeddings from a pretrained Uni-Mol encoder via cross-attention, learning a structure-conditioned, nonlinear mapping from computed to experimental infrared spectra. Applying Spec2Spec to the 238,869-molecule USTC-FG26 database yields FG26AUG, a large-scale pseudo-experimental dataset. Second, we pretrain the spectrum-based molecular generative model Spec2Mol on FG26AUG and fine-tune it on limited experimental data, markedly reducing the reliance on large paired experimental datasets. On two independent benchmarks (NIST gas-phase spectra and OCR-digitized literature spectra), this pipeline consistently outperforms counterparts pretrained on purely theoretical corpora. Gradient-weighted class activation mapping (Grad-CAM) analyses indicate that the fine-tuned model attends to chemically meaningful spectral regions. Together, these results establish a practical pathway for molecular structure generation directly from experimental vibrational spectra.

  • Open Access

    Generation of Tailorable Multi-Property Molecules via Multimodal Spectral Fusion Evolution

    by Shijie Tao, Meng Huang, Ledu Wang, Shuo Feng, Yulan Han, Jing He, Yan Huang, Zijin Jia, Donglai Zhou, Yi Feng, et al.

    Molecular generation stands at the forefront of intelligent molecular design, promising transformative advances in the discovery of molecules meeting druglikeness or catalysis requirements. However, existing generative models typically focus on optimizing a single target property at a time, hindered by the intrinsic difficulty of simultaneously encoding and evolving multiple structural features. Herein, we developed MUltimodal Spectral fusion Evolution (MUSE)—a spectroscopy-guided generative framework that unifies multimodal spectroscopic descriptors as a universal chemical language for multi-property molecular creation. A Spectra-to-Molecules (S2M) model, as one of MUSE's core modules, achieved high success rate in molecular structure generation by training on a self-built density functional theory (DFT) levelmultimodal spectral database. Leveraging the continuously variable nature of spectral descriptors, MUSE achieved synergistic integration of structural, vibrational, and electronic information by combining complementary spectroscopic modalities. Through evolutionary algorithms that constrain the generation direction of multiproperty molecules, the framework enables an up to 700-fold increase in the enrichment of rare quadruple-property candidates. This work provides a scalable pathway for multiobjective discovery in chemistry and materials science.


  • Open Access

    The Natural Coding of Language: SPECTRA is All You Need

    by Yi Feng, Shijie Tao, Xiaolong Zhang, Meng Huang, Donglai Zhou, Linjiang Chen, Daobin Liu, Guokun Yang, Hengyu Xiao, Yang Wang, et al.

    In the data-driven era, the recording and preservation of information have become increasingly crucial. To surpass the binary storage capabilities of silicon-based materials, several novel storage media have been proposed in recent years, such as synthetic DNA and polymers with sequence-encoding characteristics. Herein, we present an alternative strategy that uses spectra of small molecules as a natural code for linking human language to molecular structures. By assigning words to codes derived from selected regions of experimental spectra, we demonstrate a workflow that generates spectra conditioned on language inputs and translates them into molecules. The molecules produced in this way yield computed and measured spectra that match their intended target codes, confirming the fidelity of the approach. These findings introduce spectra as a compact and interpretable channel − “SPECTRA 1 is All You Need” − for storing and expressing information, advancing data storage, cryptography, and the emerging field of chemistry-inspired information technologies.

  • Noncovalent and precise loading of small molecule drugs on DNA nanocarriers

    by Xinyu Feng, Mingtao Zhao, Huichao Chen, Rui She, Yangying Wang, Leyi Jia, Wenting Li, Mengmeng Li, Shuheng Wei, Yunxiao Ma, et al.

    Cordycepin is loaded into aptamer-functionalized DNA to demonstrate a new method for the noncovalent and precise loading of small molecule drugs onto DNA nanocarriers. This system shows remarkable tumor suppression in melanoma-bearing mice.


  • A practical inverse design approach for high-entropy catalysts using generative AI

    by Donglai Zhou, Ruyu Yang, Zijin Jia, Yuhai Cai, Luyuan Zhao, Lulu Guo, Guilin Ye, Song Wang, Linjiang Chen, Daobin Liu, et al.


  • Supramolecular surface modification of ceria nanoparticles for enhanced chemical mechanical planarization performance in shallow trench isolation

    by Haining Zhang, Chenyan Zhao, Xueli Sun, Chengrui Xin, Fangwei Lu, Shuming Wu, Xin-Ping Qu, Song Wang, Simin Li, Hui Shen, et al.

    A supramolecular approach utilizing β-cyclodextrin was developed to engineer high-performance ceria-based slurries, demonstrating enhanced chemical mechanical polishing efficiency for shallow trench isolation applications.