SBIR Phase I: Development of a Semantic Search Engine Using Natural Language Processing to Generate Validated Technique-Based Recommendations for Life Science Research Methodology

这个小企业创新研究（SBIR）第一阶段项目的更广泛的影响/商业潜力是推进人工智能（AI）支持的搜索引擎的开发，以促进生命科学研究的可重复性和效率。拟议技术的开发将使研究人员能够快速获得关于技术和产品的无偏见建议，以推进科学发现。通过简化文献检索过程和优化实验设计阶段的研究，研究人员能够避免在实验室中进行冗长的试错，并加快生产性实验。通过在几分钟内为研究人员提供文献支持和相关的实验建议，拟议的搜索引擎可以节省研究人员花费在不适合其研究目标的实验方法上的时间和资源，同时也使研究人员能够探索有前途的方法可能超出他们的标准操作。这个小企业创新研究第一阶段项目旨在解决实验效率低下的持续问题，生命科学研究进展缓慢。第一阶段的工作将推进概念验证搜索引擎的开发和评估，以推荐与抗体相关的技术，能够细化搜索结果的过滤机制，以及自动生成的图形分析，显示技术使用的关键数据。利用机器学习和自然语言处理（NLP）扫描整个同行评审文献，并提取与基于技术的检索词相关的数据，检索输出将相应地对抗体和方案条件进行排名。为了过滤结果，将通过开发能够识别指示符合强加标准的相关上下文信息的NLP算法来施加约束，例如对设备或靶基因的访问。开发平台的搜索结果的准确性和相关性将与流行的基于研究的搜索引擎进行比较，预计将展示高度精炼的搜索输出和建议，支持改进的实验设计。该奖项反映了NSF的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The broader impact/commercial potential of this Small Business Innovation Research (SBIR) Phase I project is to advance the development of an artificial intelligence (AI)-supported search engine that facilitates reproducibility and efficiency in life science research. Development of the proposed technology will allow researchers to quickly access unbiased recommendations on techniques and products to advance scientific discovery. By streamlining the literature search process and optimizing research at the experimental design stage, researchers are able to avoid lengthy trial-and-error in the laboratory and accelerate productive experiments. By providing researchers with literature-supported and relevant experimental recommendations within minutes, the proposed search engine can spare researchers time and resources spent on experimental methods poorly suited to their research goals, while also enabling researchers to explore promising methods potentially outside their standard operations.This Small Business Innovation Research Phase I project seeks to address the persistent problems of experimental inefficiency and irreproducibility that slow life sciences research. Phase I efforts will advance the development and evaluation of a proof-of-concept search engine for recommendation of techniques associated with antibodies, a filter mechanism capable of refining search results, and automatically generated graphical analytics presenting key data on technique usage. Leveraging machine learning and Natural Language Processing (NLP) to scan the entire body of peer-reviewed literature and extract data relevant to technique-based search terms, search outputs will accordingly rank antibodies and protocol conditions. To filter results, constraints, such as access to equipment or target genes, will be imposed through development of NLP algorithms capable of identifying relevant contextual information indicating conformance to imposed criteria. Accuracy and relevance of the developed platform's search results will be compared to a popular research-based search engine and is expected to demonstrate highly refined search outputs and recommendations, supporting improved experimental design.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.