Project Title: Sustainable AI model for predicting nanoparticle toxicity and antimicrobial activity: Towards safer nanotechnology and responsible innovation
Description:
While the applications of nanotechnology continue to expand, physically testing the
safety, efficacy, and environmental impact of every new nanoparticle design is
becoming impractical. To address this bottleneck, we developed Nano-Shield, a
sustainable, AI-driven platform for nanomaterial assessment. The platform operates
as a comprehensive virtual laboratory utilizing a two-step screening process. This
process is divided across four specialized modules: NanoToxNet (NTN) for
predicting toxicity, NanoGuardNet (NGN) for assessing antimicrobial activity, and
two sustainability modules (Batch Assessment and Oxide Assessment) for evaluating
resource consumption and environmental impact.
We built and evaluated the framework using the Nanoparticle Toxicity Dataset and
NanoAntimicrobialDB. After standardizing the data and preparing the features, we
tested four supervised machine learning algorithms: Random Forest (RF), K-Nearest
Neighbors (KNN), Support Vector Machines (SVM), and Decision Trees (DT).
Relying on its tree voting consensus, Random Forest significantly outperformed the
others, achieving 97.35% accuracy for toxicity prediction and 90.01% accuracy for
antimicrobial activity.
Beyond its predictive capabilities, Nano-Shield calculates the real-world
environmental footprint of potential physical experiments. It evaluates energy and
water consumption, chemical waste, carbon emissions, and financial costs, converting
these metrics into tangible equivalents—such as kilometers driven by a car or standard
water bottles consumed. By combining predictive machine learning with Safe-and
Sustainable-by-Design (SSbD) principles, Nano-Shield allows researchers to identify
the safest and most sustainable nanoparticle candidates before stepping foot in a
physical lab, ultimately reducing experimental waste and promoting responsible
innovation.
Uploaded At: 2026-07-03 05:47:35