Infoscience presenting at the Society for Professional Data Managers (SPDM) June 2023

Infoscience is delighted to present at the Society for Professional Data Managers (SPDM) mid year event next week which is on the theme “Data Management in an AI World – Becoming Obsolete or Increasingly Vital?”.

The abstract is titled “Unstructured Data Management and Artificial Intelligence: Two sides of the Same Coin?”.

Abstract

Data may be seen as the beating heart of the Artificial Intelligence (AI) system. As AI use increases, the transparency and quality of the training data it is based on becomes paramount. The predictions and ability to generalise of any model are intrinsically linked with its training data. These data may include single truth reference data, taxonomies, labelled and unlabelled data. The data scientist may just concern themselves with an F1 score of the model, but the data manager will also be concerned with the underlying model datasets provenance, restrictions, representativeness, bias, format and use for reproducibility of published models. Algorithm accountability may emerge as one of the key themes for our digital society and one where the data manager has an important role to play.

AI can automatically and autonomously manipulate its input (data), into derivative products to be used for insights and actions. Where AI is targeted towards the data management process, this may relate to automating some aspect of tasks relating to data quality & provenance, classification and cataloguing, integration, summarisation, accessibility, security, search & discovery, publishing, archival and deletion. The days of true AI emergence may (or may not) arrive, but for now it is for the data manager to exploit AI to automate tasks to speed up workflows and enhance data discovery and security. AI has already replaced in some cases narrow data management tasks that used to be done manually, and is used to assist others. As the volume, velocity and veracity of data grows, so do the possibilities for innovation in the data management process with the data manager well positioned to help orchestrate this digital transformation.

This duality of data management and AI will be explored further, using specific examples from the subsurface & wells discipline including ChatGPT as it relates to AI and unstructured data.

The full programme can be found here: https://www.societypdm.org/events

#subsurface #datamanagement #osdu #unstructureddata #naturallanguageprocessing #languagemodels #chatgpt #artificialintelligence