"Sage Publishing", "Sage Texti", their logos and relevant iconography remain copyright of Sage Publications Inc. and are show here for demonstration purposes.
Sage Publishing, a leading academic publisher, recognized the need to empower its data scientists with advanced tools for extracting insights from vast corpora of scientific papers. In response, they embarked on a collaborative project with our team from September to December 2020, resulting in the development of Sage Texti - a bespoke, AI-driven solution revolutionizing scientific data analysis.
Objective and Scope
The primary objective of Sage Texti was to provide researchers with a user-friendly platform to extract information from scientific papers without the need for extensive data science expertise. The scope encompassed building an intuitive interface for querying the text corpus, as well as incorporating AI-driven data processing capabilities to streamline analysis tasks.
AI Integration, Functionality, and Features
Sage Texti leverages state-of-the-art machine learning algorithms integrated into a custom Python-based framework. While specifics regarding the AI implementation are covered under NDA, the solution paved the way for efficient data analysis akin to modern language models like ChatGPT.
The Sage Texti platform comprised a React-based frontend and an ASP .NET Core-based backend, providing users with seamless access to their accounts, data processing functionalities, and action runs. Container-based environments ensured scalability and security, with data processing performed in ephemeral containers to safeguard privacy.
User Experience and Interface
Users interact with an intuitive web application, where they can drag and drop predefined data science actions to build workflows. This WYSIWYG editor abstracts away complex Python code, enabling users to design custom analyses effortlessly. Upon submission, data and actions are processed in containerised environments, with results promptly delivered to users via the frontend.
Collaboration and Communication
Close collaboration with Dr. Daniela Duca, Sage Publishing’s senior project manager, ensured alignment with project goals and requirements. Weekly meetings facilitated feedback loops and feature discussions, enabling iterative improvements and timely adjustments based on user needs.
Data Security and Privacy
Data security and privacy were paramount considerations in the design of Sage Texti. Container-based data processing environments minimized the attack surface, while ephemeral containers ensured data confidentiality. Users were responsible for securely managing their input data, as processed data was not retrievable after processing.
Project Timelinem Outcomes and Impact
The MVP stage of Sage Texti spanned three months, during which rapid development was essential to meet project milestones. Collaboration with Sage Publishing’s internal resources ensured timely progress and adherence to deadlines.
Following the MVP stage, Sage Texti was launched as a service by Sage Publishing, enabling researchers to access advanced data analysis capabilities for scientific papers. While specific outcomes post-implementation are not available due to our limited involvement post-MVP, the availability of Sage Texti marks a significant milestone in empowering researchers with cutting-edge tools.
Future Developments
While we have had no direct involvement with Sage Texti since the end of the MVP stage, its launch as a service suggests ongoing developments and enhancements by Sage Publishing. This interview with Dr. Duca provides insights into the project’s evolution and future directions, indicating continued innovation and exploration within Sage Publishing.
Conclusion
Sage Texti stands as a testament to the collaborative efforts between Sage Publishing and our team, exemplifying the power of technology in advancing scientific research. By democratizing access to advanced data analysis capabilities, Sage Texti paves the way for transformative insights and discoveries in the scientific community.