Projects

Current Research Project:

PI of the project Enhancing Graphical Abstracts for Research Communication: A Human-in-the-Loop Artificial Intelligence (AI) Approach (AIGA)

Summary

The increasing complexity of scientific communication calls for innovative approaches to ensure that research findings are both accessible and visually compelling. Graphical abstracts have emerged as an important tool to summarize and communicate the main contributions of scientific articles, but their production is often time-intensive and limited by the authors¿ design skills. This project examines the potential of large language models (LLMs) to aid researchers in generating graphical abstracts from article abstracts, without relying on full-text content to circumvent copyright issues. Central to this project is a human-in-the-loop approach, where human expertise remains crucial in guiding semantic extraction, refining generated prototypes, and ensuring interpretability and accuracy of the content. The methodology is structured around three steps: first, combining bibliometric techniques and LLM-based semantic analysis to extract and represent key concepts from abstracts; second, generating graphical prototypes in Scalable Vector Graphics (SVGs) format; and third, improving this prototype with AI image generation and editing models to improve quality and visual appeal. Human validation and fine-tuning will be integral at every stage, ensuring both scientific accuracy and clear communication. The project will conclude with a verification survey involving experts assessing users’ perceptions of the generated graphic abstracts, without any comparison with existing copyrighted materials. By advancing the integration of LLMs into research communication practices, this work seeks to establish a framework for hybrid human-AI co-creation in science communication.

Participant in the project entitled Designing the Sapientia Observatory on University Sustainability (PI prof. Daraio)

Current Third mission Project:

Participant in the project entitled Sostenibile o insostenibile, questo è il problema! È più nobile vivere il presente o temere il futuro? (SINE METU, PI prof. Placidi)

Other projects:

Sciento AI

ScientoAI applies AI Bibliometric Grounded Analysis to bridge the gap between traditional bibliometrics and modern AI capabilities. It does not replace classical bibliometric methods but complements them by adding a semantic layer to the analysis.

The software leverages Large Language Models (LLMs) to understand the context of research papers, making it an essential tool for constructing Narrative CVs by extracting and visualizing the underlying narrative of a researcher’s work.

Available at Sciento AI App on Hugging Face