SafeHelper: Developing LLM-Powered Multimodal Systems for Human-Centric Safety-Critical Decision-Making
Research Project, 2024 – 2029

The objective of this project is to create SafeHelper, a framework for designing and building LLM-based systems tailored for human-centric, safety-critical applications. SafeHelper aims to offer a structured approach to both the design and evaluation of these systems. Specifically, the SafeHelper framework includes three integral LLM components: SafeHelper.Interpreter, SafeHelper.Analyst, and SafeHelper.Guard, each dedicated to a unique function within the system to ensure reliable analysis, interpretation, and risk assessment.

Participants

Yinan Yu (contact)

Chalmers, Computer Science and Engineering (Chalmers), Functional Programming

Robert Feldt

Chalmers, Computer Science and Engineering (Chalmers), Software Engineering (Chalmers)

Collaborations

Sahlgrenska University Hospital

Gothenburg, Sweden

Funding

Wallenberg AI, Autonomous Systems and Software Program

Funding Chalmers participation during 2024–2029

Related Areas of Advance and Infrastructure

Information and Communication Technology

Areas of Advance

Health Engineering

Areas of Advance

More information

Latest update

9/27/2024