Projects
Projects: Custom Search |
||
| Reference Number | NIA2_NGESO064 | |
| Title | Generative AI Discovery | |
| Status | Completed | |
| Energy Categories | Other Power and Storage Technologies (Electricity transmission and distribution) 100%; | |
| Research Types | Applied Research and Development 100% | |
| Science and Technology Fields | PHYSICAL SCIENCES AND MATHEMATICS (Computer Science and Informatics) 100% | |
| UKERC Cross Cutting Characterisation | Not Cross-cutting 100% | |
| Principal Investigator |
Project Contact No email address given National Grid plc |
|
| Award Type | Network Innovation Allowance | |
| Funding Source | Ofgem | |
| Start Date | 01 February 2024 | |
| End Date | 31 March 2024 | |
| Duration | ENA months | |
| Total Grant Value | £36,000 | |
| Industrial Sectors | Power | |
| Region | London | |
| Programme | Network Innovation Allowance | |
| Investigators | Principal Investigator | Project Contact , National Grid plc (99.999%) |
| Other Investigator | Project Contact , National Grid ESO (0.001%) |
|
| Industrial Collaborator | Project Contact , National Grid plc (0.000%) |
|
| Web Site | https://smarter.energynetworks.org/projects/NIA2_NGESO064 |
|
| Objectives | This project aims to discover and qualify how Gen AI can potentially add value to the ESO"s strategic objectives, leveraging Azure"s cloud infrastructure and harnessing the capabilities of Azure OpenAI generative models. The core technical aspects of the project include Identification of more than 15 use cases across the business, where Gen AI can augment current workflows and aid in improving processes. Through this engagement the project will deepen awareness and understanding of technology. Testing and validating: We will prototype Gen AI solutions for 3 high priority use cases. This will enable the testing and validating of the technology and identify ways of working with technology. The prototypes will be demonstrated to a wider audience within ESO to share learnings and inform future actions. Representative public datasets (e.g. ESO data portal) will be used for lab testing. The solutions will be built and tested on supplier"s environment.Importantly from a data governance, security and privacy point of view the data is ring fenced. Therefore, data from our lab tests prompts (inputs) and completions (outputs), embeddings, and training dataare not available to OpenAIare not used to improve OpenAI modelsare not used to improve any Microsoft or 3rd party products or servicesare not used for automatically improving Azure OpenAI modelsFurthermore, to reduce hallucinations (incorrect or misleading results that AI models generate), a Retrieval Augmentation Generation (RAG) will be implemented. RAG is an architecture that augments the capabilities of a Large Language Model (LLM) like ChatGPT by adding an information retrieval system that provides grounding data. Adding an information retrieval system gives control over grounding data used by an LLM when it formulates a response. This project will be delivered within the following work packages:Work package 1 Discovery and write up of 15 use cases:Familiarisation with use case prioritisation framework developed within the AI CoE (NIA2_NGESO021).Identification and research of use cases through workshops with key stakeholders.Conduct use case initial prioritisation with ESO stakeholders.Finalise 15 use cases and identify top 3 priority use cases to take forward for lab testingWork package 2 Lab testing of 3 high priority use cases:Complete lab testing of 3 use cases.Conduct model testing and validation.Run demo sessions to showcase lab test results.Work package 3 Develop high level business case:Document and share final resultsThis project will utilise representative public data sets for the development and lab testing of the use cases. Benchmarking of lab testing outputs will be performed were possible by comparing with existing quantitative or qualitative data available from business owners, considering the time, effort, and quality of existing outputs.In line with the ENA"s ENIP document, the risk rating is scored Low.TRL steps = 1Cost = 1 (<£500k)Suppliers = 1 (1 supplier)Data assumptions = 1 (Defined assumptions & principles)Total = 4 This project will utilise representative public data sets for the development and lab testing of the use cases. Benchmarking of lab testing outputs will be performed were possible by comparing with existing quantitative or qualitative data available from business owners, considering the time, effort, and quality of existing outputs.In-scope: Generating unstructured text data, interpreting plots, graphs and figures embedded within reports, and synthetic data generation/exploration (tabular data containing text and numeric data) are marked as in-scope. Out of Scope: Generating images/audio/video are marked as out-of-scope, as those solutions will require more sophisticated generative models and evaluation. The objective of this project is to identify and test use cases for Gen AI technology across a range of core ESO roles: knowledge management, stakeholder engagement and customer operations. Prototyping use cases will identify additional business uses and ways of working responsibly with the technology and uncover limitations and opportunities of using Gen AI in workflows. | |
| Abstract | Generative AI (Gen AI) creates diverse, realistic artifacts across various domains, including images, video, music, speech, text, etc. Its applications range from common tasks like composing emails to complex data analysis. This project will explore high impact use cases appropriate for Gen AI deployment across the ESO. Open Data initiatives will then be used to lab test three priority use cases on publicly available data. This comprehensive approach will underscore the transformative potential of Gen AI in producing scalable, diverse artifacts reflective of its training data without replication. | |
| Data | No related datasets |
|
| Projects | No related projects |
|
| Publications | No related publications |
|
| Added to Database | 02/10/24 | |
