Reference Number: ARL-SRP-00019
Project Description
As artificial intelligence (AI) systems continue to grow in capability and complexity, there is increasing interest in methods for selectively removing the influence of specific data, knowledge, or behaviors from trained models. This research project will explore machine unlearning as a potential approach for safeguarding AI systems and improving their adaptability in changing operational or data environments.
The objective of this project is to investigate research questions related to how AI models can be updated, refined, or modified to reduce the impact of targeted information without requiring complete retraining in every case. The participant may support a range of research activities, such as reviewing relevant literature, assisting with data preparation and annotation, helping develop experimental approaches, implementing or adapting prototype methods, conducting tests, and analyzing results.
Depending on project needs, the work may involve machine learning frameworks, data analysis tools, model evaluation techniques, annotation workflows, and other research software commonly used in AI development. Specific methods and tasks may vary based on project direction, technical maturity, and research priorities at the time of execution.
Expected outcomes may include a summary of relevant findings, experimental results, annotated datasets or supporting research artifacts, prototype implementations, or recommendations for future research. The project is intended to provide flexibility for evolving research goals while giving participants exposure to current challenges in machine unlearning and AI safeguarding.
Technical Skills
- Artificial Intelligence/Machine Learning
- Computer Science
- Mathematics
- Natural Language Processing
