Machine Unlearning

Description

Machine unlearning enables removal of specific training data's influence from trained models without complete retraining. This technique addresses privacy rights like GDPR's right to be forgotten by selectively erasing learned patterns associated with particular data points, individuals, or sensitive attributes. Methods include exact unlearning (provably equivalent to retraining without the data), approximate unlearning (efficient algorithms that closely approximate retraining), and certified unlearning (providing formal guarantees about information removal).

Example Use Cases

Privacy

Responding to user deletion requests in a social media recommendation system by removing all influence of that user's historical interactions, ensuring GDPR compliance and verifiable data removal.

Fairness

Removing specific patient records from a hospital's diagnostic model after a patient withdraws consent, ensuring the model no longer reflects patterns from that individual's medical history while maintaining clinical accuracy for other patients.

Transparency

Enabling a financial institution to demonstrate regulatory compliance by providing cryptographic proof that a former customer's transaction history has been completely removed from credit risk assessment models following account closure.

Limitations

Exact unlearning for complex models like deep neural networks is computationally expensive, often nearly as costly as full retraining.
Approximate unlearning methods may not provide strong guarantees that information has been fully removed, potentially leaving residual influence.
Difficult to verify unlearning effectiveness, as adversaries might extract information about supposedly removed data through membership inference or other attacks.
Repeated unlearning requests can degrade model performance significantly, especially if many data points are removed from the training distribution.
Requires access to model architecture and training process details (white-box access), making it difficult to apply to third-party models or models where internal structure is proprietary.
Particularly challenging for very large foundation models where even storing checkpoints for potential retraining is infeasible, limiting practical applicability to smaller, domain-specific models.

Resources

Research Papers

Right to be forgotten in the era of large language models

Dawen Zhang et al.•Sep 10, 2024

Digital forgetting in large language models: A survey of unlearning methods

Alberto Blanco-Justicia et al.•Jan 13, 2025

Machine Unlearning for Traditional Models and Large Language Models: A Short Survey

Yi Xu•Apr 1, 2024

With the implementation of personal data privacy regulations, the field of machine learning (ML) faces the challenge of the "right to be forgotten". Machine unlearning has emerged to address this issue, aiming to delete data and reduce its impact on models according to user requests. Despite the widespread interest in machine unlearning, comprehensive surveys on its latest advancements, especially in the field of Large Language Models (LLMs) is lacking. This survey aims to fill this gap by providing an in-depth exploration of machine unlearning, including the definition, classification and evaluation criteria, as well as challenges in different environments and their solutions. Specifically, this paper categorizes and investigates unlearning on both traditional models and LLMs, and proposes methods for evaluating the effectiveness and efficiency of unlearning, and standards for performance measurement. This paper reveals the limitations of current unlearning techniques and emphasizes the importance of a comprehensive unlearning evaluation to avoid arbitrary forgetting. This survey not only summarizes the key concepts of unlearning technology but also points out its prominent issues and feasible directions for future research, providing valuable guidance for scholars in the field.

Software Packages

awesome-machine-unlearning

Jul 7, 2022

Awesome Machine Unlearning (A Survey of Machine Unlearning)

machine_unlearning

Sep 11, 2020

Existing Literature about Machine Unlearning

Related Techniques

Name	Description	Assurance Goals
Influence Functions	Influence functions quantify how much each training example influenced a model's predictions by computing the change in prediction that would occur if that training example were removed and the model retrained. Using calculus and the implicit function theorem, they approximate this 'leave-one-out' effect without actually retraining the model by computing gradients and Hessian information. This mathematical approach reveals which specific training examples were most responsible for pushing the model toward or away from particular predictions, enabling practitioners to trace problematic outputs back to their root causes in the training data.	Explainability Fairness Privacy
Federated Learning	Federated learning enables collaborative model training across multiple distributed parties (devices, organisations, or data centres) without requiring centralised data sharing. Participants train models locally on their private datasets and only share model updates (gradients, weights, or aggregated statistics) with a central coordinator. This distributed approach serves multiple purposes: preserving data privacy and sovereignty, reducing communication costs, enabling learning from diverse data sources, improving model robustness through heterogeneous training, and facilitating compliance with data protection regulations whilst maintaining model performance comparable to centralised training.	Privacy Reliability Safety Fairness
Homomorphic Encryption	Homomorphic encryption allows computation on encrypted data without decrypting it first, producing encrypted results that, when decrypted, match the results of performing the same operations on the plaintext. This enables secure outsourced computation where sensitive data remains encrypted throughout processing. By allowing ML operations on encrypted data, it provides strong privacy guarantees for applications involving highly sensitive information.	Privacy Safety Transparency Security
Membership Inference Attack Testing	Membership inference attack testing evaluates whether adversaries can determine if specific data points were included in a model's training set. This technique simulates attacks where adversaries use model confidence scores, prediction patterns, or loss values to distinguish training data from non-training data. Testing measures privacy leakage by calculating attack success rates, precision-recall trade-offs, and advantage over random guessing. Results inform decisions about privacy-enhancing techniques like differential privacy or regularisation.	Privacy Security Transparency