Skip to main content Skip to navigation

Bridging the Digital Divide: Small Language Models as a Pathway for Physics and Photonics Education in Underdeveloped Regions

Project Overview

The document highlights the transformative role of Small Language Models (SLMs) in enhancing education, particularly in underdeveloped regions where access to resources and infrastructure is limited. By functioning offline on low-power devices, SLMs facilitate interactive and scalable learning experiences, especially in STEM subjects like physics and photonics. They act as virtual tutors, helping students grasp complex concepts, while also assisting educators in designing curricula tailored to local needs. Importantly, SLMs can be fine-tuned to support native languages, thereby promoting inclusivity in learning. Despite their potential, the document acknowledges significant challenges such as issues with hallucination and varying performance across multiple languages, which must be addressed to ensure effective deployment and maximize educational outcomes. Overall, the findings underscore the promise of generative AI in revolutionizing education and bridging gaps in learning opportunities.

Key Applications

Small Language Models (SLMs) as virtual tutors

Context: Education in underdeveloped regions, targeting students and educators in STEM fields

Implementation: SLMs operate offline on low-power devices, enabling interactive learning without stable internet access.

Outcomes: Increased engagement and understanding of complex topics in physics; support for educators in curriculum design.

Challenges: Hallucination issues in educational contexts; limitations in multilingual performance for low-resource languages.

Implementation Barriers

Infrastructure Barrier

Limited educational infrastructure, including lack of electricity, equipped science laboratories, and access to AI technologies.

Proposed Solutions: Investment in AI technologies and localized AI solutions to bridge the digital divide.

Human Resource Barrier

Shortage of qualified educators trained in STEM subjects, leading to an insufficient workforce to effectively implement generative AI in education.

Proposed Solutions: SLMs can act as virtual tutors to compensate for the lack of trained educators.

Socioeconomic Barrier

Poverty and economic hardship forcing students to prioritize work over education, which can limit their access to STEM education.

Proposed Solutions: Targeted policies and programs to support education and retention in STEM fields.

Cultural Barrier

Societal expectations discouraging women from pursuing STEM careers, which affects overall participation in STEM education.

Proposed Solutions: Mentorship programs and initiatives fostering gender inclusivity in STEM education.

Project Team

Asghar Ghorbani

Researcher

Hanieh Fattahi

Researcher

Contact Information

For information about the paper, please contact the authors.

Authors: Asghar Ghorbani, Hanieh Fattahi

Source Publication: View Original PaperLink opens in a new window

Project Contact: Dr. Jianhua Yang

LLM Model Version: gpt-4o-mini-2024-07-18

Analysis Provider: Openai

Let us know you agree to cookies