Chinese Researchers Unveil Fengshu 2.0, a Vertical Large Model for Soybean Breeding

Professor Wang Xiaobo of Anhui Agricultural University unveiled the vertical large model "Fengshu" 2.0 for soybean research at the 32nd National Soybean Research and Production Seminar in Nanning, Guangxi. By fusing multimodal data and mitigating single-model bias, this tool accelerates the integration of soybean data into breeding programs.
Collaborative multi-model analysis minimizes single-model bias
"Fengshu" 2.0 constructs a structured knowledge system by integrating soybean germplasm resources, multi-omics data (genome, proteome, transcriptome), disease samples, breeding trial records, and global literature. Unlike its predecessor, version 2.0 employs a multi-model collaborative mechanism that leverages general large model interfaces to address specific professional queries independently.
A dedicated summary module then synthesizes these outputs by evaluating consensus, discrepancies, and evidence completeness against domain knowledge and private data. This approach generates more reliable conclusions, significantly enhancing the stability of soybean knowledge services and preventing misleading results caused by single-model hallucinations.
Six integrated modules support comprehensive breeding workflows
The platform supports critical agronomic tasks, including disease diagnosis, parent selection, virtual cross design, phenotypic analysis, marker-assisted breeding, and candidate gene function studies. It features six core modules—"Soy Encyclopedia," "Soy Molecule," "Soy Literature," "Soy Disease," "Soy Phenotype," and "Soy Breeding"—to bridge the gap between multidimensional data systems and practical breeding applications.
Powered by over 10 million words of specialized text, 10,000 scientific papers and patents, and a domain knowledge graph comprising 20,000 entities and 100,000 relationships, the model also incorporates 8,000 genome re-sequencing datasets, approximately 40,000 germplasm resources, and 3,000 variety phenotypic records. This extensive data foundation establishes a robust framework for intelligent soybean breeding.
Related article
Relay AI Shutdown: Founder Moves to Google Chrome
Relay, an AI-driven workflow automation platform, has officially shut down. Founder and CEO Jacob Bank, along with select team members, are transitioning to the Google Chrome division. Bank confirmed that paid subscriptions will cease on September 14
Alibaba, DeepSeek Drive China’s AI Model Race to Lower Costs
Alibaba has introduced Qwen3.8-Max, its most powerful AI model yet, as DeepSeek’s new V4-Flash gains attention for offering inference pricing lower than several competitors.With 2.4 trillion parameters, Qwen3.8-Max employs a mixture-of-experts archit
Former Spotify Exec Launches Vinyl Bar Startup in Shibuya to Redefine Music Apps
As generative AI continues to reshape the creative landscape, a new wave of startups is empowering users to produce music through AI prompts, mimicking the style of any artist. A Vinyl Bar in Shibuya, a startup with an indie band-inspired name, is de
Related Special Topic Recommendations
Comments (1)
0/500

Professor Wang Xiaobo of Anhui Agricultural University unveiled the vertical large model "Fengshu" 2.0 for soybean research at the 32nd National Soybean Research and Production Seminar in Nanning, Guangxi. By fusing multimodal data and mitigating single-model bias, this tool accelerates the integration of soybean data into breeding programs.
Collaborative multi-model analysis minimizes single-model bias
"Fengshu" 2.0 constructs a structured knowledge system by integrating soybean germplasm resources, multi-omics data (genome, proteome, transcriptome), disease samples, breeding trial records, and global literature. Unlike its predecessor, version 2.0 employs a multi-model collaborative mechanism that leverages general large model interfaces to address specific professional queries independently.
A dedicated summary module then synthesizes these outputs by evaluating consensus, discrepancies, and evidence completeness against domain knowledge and private data. This approach generates more reliable conclusions, significantly enhancing the stability of soybean knowledge services and preventing misleading results caused by single-model hallucinations.
Six integrated modules support comprehensive breeding workflows
The platform supports critical agronomic tasks, including disease diagnosis, parent selection, virtual cross design, phenotypic analysis, marker-assisted breeding, and candidate gene function studies. It features six core modules—"Soy Encyclopedia," "Soy Molecule," "Soy Literature," "Soy Disease," "Soy Phenotype," and "Soy Breeding"—to bridge the gap between multidimensional data systems and practical breeding applications.
Powered by over 10 million words of specialized text, 10,000 scientific papers and patents, and a domain knowledge graph comprising 20,000 entities and 100,000 relationships, the model also incorporates 8,000 genome re-sequencing datasets, approximately 40,000 germplasm resources, and 3,000 variety phenotypic records. This extensive data foundation establishes a robust framework for intelligent soybean breeding.
Relay AI Shutdown: Founder Moves to Google Chrome
Relay, an AI-driven workflow automation platform, has officially shut down. Founder and CEO Jacob Bank, along with select team members, are transitioning to the Google Chrome division. Bank confirmed that paid subscriptions will cease on September 14
Alibaba, DeepSeek Drive China’s AI Model Race to Lower Costs
Alibaba has introduced Qwen3.8-Max, its most powerful AI model yet, as DeepSeek’s new V4-Flash gains attention for offering inference pricing lower than several competitors.With 2.4 trillion parameters, Qwen3.8-Max employs a mixture-of-experts archit
Former Spotify Exec Launches Vinyl Bar Startup in Shibuya to Redefine Music Apps
As generative AI continues to reshape the creative landscape, a new wave of startups is empowering users to produce music through AI prompts, mimicking the style of any artist. A Vinyl Bar in Shibuya, a startup with an indie band-inspired name, is de





Home






