Generative AI Support for Research

AI for Research

Research Technology supports the use of generative AI for research including:

  • Data Processing: Web scraping or downloading data from online archives, assisting in formatting text, audio or video files for machine learning/AI model development.

  • Text Generation: We can assist researchers of all skill levels to get started using either closed-source large language models like ChatGPT or open source alternatives. This includes traditional natural language processing (NLP) approaches such as sentiment analysis, named entity recognition and entity linking, as well as more complicated tasks like adding text markup or topic modeling.

  • Transcription: Automatic speech recognition and speaker recognition (diarization) can save you the time of manually transcribing interviews or other audio sources. We can also incorporate line-by-line translation, allowing you to work in several languages at once.

  • Qualitative Data Analysis: Methods for automatic qualitative coding offer a time-saver for you or your team. We can also perform further statistical analysis or data visualization similar to NVivo or Dedoose, alongside supporting those programs.

  • Literature Review: AI can search the web for you! Given a research goal, we can have AI choose sources for a literature review or survey of an academic field.

  • Study Material Generation: With a clearly defined goal, we can develop applications with a limited scope to generate first-draft study material for human review from course material including textbooks, articles and course objectives.

  • Translation: We can work with many different languages both using proprietary services like Google Translate or DeepL and open-source alternatives to translate your documents in whatever form best fits your purposes.

  • Optical Character Recognition: Transforming images of text into a machine actionable format can be difficult with out help. Let us assist you with rendering these sources into sharable digital editions, and connect you with our contacts in the Libraries.

  • Reproducibility: We can help you create a reproducible environment for your research, to deploy your generative AI model to colleagues. We are happy to enable reproducibility and meet data sharing requirements from funders.

  • Consultations: This includes, but is not limited to, the following research areas:

    • Bioinformatics: We also support generative AI workflows in bioinformatics, including use of Alpha Fold for generative protein analysis.

    • Data Science: We can help support conceptual approaches for building generative AI models, and general coding and implementation questions, and provide in depth consultations. Additionally, we can help you scale your model from your local machine to the Tufts High-Performance Compute cluster (HPC).

    • Digital Humanities/NLP: We can assist with any projects that require the analysis of large corpora of textual data. In a consultation, we can discuss the specific methods that best fit your needs.