View Proposal

---

Proposer

Andrew Ireland

Title

A Neuro-symbolic Approach to Explaining System Failures and Accidents

Goal

Build a tool that assists engineers in finding explanations of system failures and accidents.

Description

Problem frames is a technique that uses common patterns that occur within problems to capture requirements within the context of software engineering. A feature of the problem frames approach is that it draws a distinction between system requirements (world view) and software requirements (software view). The consistency between the world and software views can then be verified mechanically, e.g., via an off-the-shelf SAT Solver. Note that inconsistencies at this level have historically led to catastrophic system failures and associated accidents. Verifying consistency will typically rely upon assumptions about the 'world' in which the intended system will operate. Problems frames does not help with ensuring the validating of such assumptions. This is where LLMs might help, i.e., we propose to use LLMs to explain the validity (or otherwise) of a given set of assumptions. Ultimately, it will be the responsibility of a human engineer to decide if they believe the LLM's explanations. But a tool that combines symbolic 'consistency verification' with neuro (i.e., LLM) explanations of 'assumption validating' could increase the productivity of an engineer. It would be a cool tool too! For more details, see the URL slot below that provides an External Link to a note that more fully describes the proposal and includes an example.

Resources

LLM and an appropriate programming language, i.e., library support for propositional reasoning. Problem Frames: https://en.wikipedia.org/wiki/Problem_frames_approach http://www.jacksonworkbench.co.uk/stevefergspages/pfa/index.html SAT Solver: https://simewu.com/SAT-solver/ SAT Solver: https://homes.cs.washington.edu/~kevinz/sat-solver/

Background

Logic and Proof, AI, Software Engineering.

Url

Difficulty Level

High

Ethical Approval

None

Number Of Students

1

Supervisor

Andrew Ireland

Keywords

requirements engineering, llms, explaining accidents

Degrees

Bachelor of Science in Computer Science
Master of Engineering in Software Engineering
Master of Science in Software Engineering