From Words to Search: GoTriple's AI ChatBot for Efficient Research Engagement
Authors/Creators
Description
This presentation introduces an approach to simplify access through the overwhelming volume of GoTriple papers, developing an AI ChatBot.
The project’s core objective is to provide comprehensive AI support to researchers as they embark on new research tasks, offering a two-fold solution.
Firstly, users can initiate general queries with the AI ChatBot, which searches for an answer within the full text of the publications available in GoTriple. This allows users to gain a preliminary, holistic, document-based understanding of the topic they are addressing on the platform. This initial interaction would act as an ice-breaker on a subject, allowing the researcher to get a general idea of what is being discussed and whether concretely there are, at the present time, any publications contained in GoTriple that might be right for him. Secondly, by providing the documentary source from which the unstructured answer was generated, it would allow the searcher to directly access the referenced document to check its content. This feature empowers researchers to delve deeper into papers if they deem it necessary, encouraging in-depth exploration of the content and finer details through the GoTriple platform. This approach would not only provide faster access to the information source but also guide researchers toward targeted, purposeful engagement and with a different approach through the GoTriple platform. Researchers are prompted to explore articles only when the AI ChatBot’s responses align with their research objectives or, possibly, they could refine their original query to be directed to content of more specific interest. This, in a sense, would minimize the time spent in reading many abstracts that, although they may seem interesting at first glance, turn out in the second instance to be not strictly relevant to the use case. Finally, through the identification of reasoned evaluation methods, we will present a method for verifying the quality of the responses obtained and the relevance to the documents surveyed. Conceived in this manner, the AI ChatBot would be able to provide the final researcher a quick response on the content of the research topic and quick reference to resources in the platform, increasing the navigability and the usability of GoTriple environment.
Files
OPERAS_Conference_2024_Poster_50.pdf
Files
(1.6 MB)
| Name | Size | Download all |
|---|---|---|
|
md5:4fcc2f6265ad617dbfec2879becc4bfc
|
1.6 MB | Preview Download |