| Departments |
Computer Science, Zienkiewicz Institute for Modelling, Data and AI
|
|---|
| Description | There is significant current interest in hybrid approaches to AI, which combine the strengths of a logic-based approach for accurate, explanable, and transparent reasoning with a neural network-based approach such as ChatGPT, which leverages and quickly processes information found in large language models. The aim of this project is to produce a computational representation of parts of the UK Highway Code, so that one can reason and execute actions as in an Autonomous Vehicle. The project takes an integrated hybrid approach using an BLAWX, which is an integrated open-source graphical interface to a logic-based programming language in constraint Answer Set Programing with a neural network-based, Large Language Model based system such as ChatGPT. Blawx simplifies the interface to the programming language, making programming more generally accessible. |
|---|
| Preparation | The project is based on the following article and materials, which should be consulted to have a sense of the project:
Rules as Code vs. ChatGPT: Lessons from Converting Canadian Federal Legislation into Code using Blawx
https://www.google.com/url?sa=t&source=web&rct=j&opi=89978449&url=https://site.unibo.it/hypermodelex/en/publications/2024-12-perron-logie.pdf/%40%40download/file/2024-12-PERRON-LOGIE.pdf
Github Blawx
https://github.com/Lexpedite/blawx
Jason Morris on LinkedIn
https://www.linkedin.com/posts/jason-morris-09684023_rulesascode-rulesascode-activity-7280001618677260288-iRn3/ |
|---|
| Project Categories |
Artificial Intelligence (AI), Data Science, January Cohort, Law, Modelling, Theorical Computer Science |
|---|
| Project Keywords |
Logic, Neural Networks, Scientific Modelling, Text Analysis |
|---|
Level of Studies
|
| Level 6 (Undergraduate Year 3) |
yes
|
|---|
| Level 7 (Masters) |
yes
|
|---|
| Level 8 (PhD) |
yes
|
|---|