This work delves into the complexities of machine unlearning in the face of distributional shifts, particularly focusing on the challenges posed by non-uniform feature and label removal. With the advent of regulations like the GDPR emphasizing data privacy and the right to be forgotten, machine learning models face the daunting task of unlearning sensitive information without compromising their integrity or performance. Our research introduces a novel approach that leverages influence functions and principles of distributional independence to address these challenges. By proposing a comprehensive framework for machine unlearning, we aim to ensure privacy protection while maintaining model performance and adaptability across varying distributions. Our method not only facilitates efficient data removal but also dynamically adjusts the model to preserve its generalization capabilities. Through extensive experimentation, we demonstrate the efficacy of our approach in scenarios characterized by significant distributional shifts, making substantial contributions to the field of machine unlearning. This research paves the way for developing more resilient and adaptable unlearning techniques, ensuring models remain robust and accurate in the dynamic landscape of data privacy and machine learning.
Machine unlearning has garnered significant attention due to its ability to selectively erase knowledge obtained from specific training data samples in an already trained machine learning model. This capability enables data holders to adhere strictly to data protection regulations. However, existing unlearning techniques face practical constraints, often causing performance degradation, demanding brief fine-tuning post unlearning, and requiring significant storage. In response, this paper introduces a novel class of machine unlearning algorithms. First method is partial amnesiac unlearning, integration of layer-wise pruning with amnesiac unlearning. In this method, updates made to the model during training are pruned and stored, subsequently used to forget specific data from trained model. The second method assimilates layer-wise partial-updates into label-flipping and optimization-based unlearning to mitigate the adverse effects of data deletion on model efficacy. Through a detailed experimental evaluation, we showcase the effectiveness of proposed unlearning methods. Experimental results highlight that the partial amnesiac unlearning not only preserves model efficacy but also eliminates the necessity for brief post fine-tuning, unlike conventional amnesiac unlearning. Moreover, employing layer-wise partial updates in label-flipping and optimization-based unlearning techniques demonstrates superiority in preserving model efficacy compared to their naive counterparts.
The trustworthy machine learning (ML) community is increasingly recognizing the crucial need for models capable of selectively 'unlearning' data points after training. This leads to the problem of machine unlearning (MU), aiming to eliminate the influence of chosen data points on model performance, while still maintaining the model's utility post-unlearning. Despite various MU methods for data influence erasure, evaluations have largely focused on random data forgetting, ignoring the vital inquiry into which subset should be chosen to truly gauge the authenticity of unlearning performance. To tackle this issue, we introduce a new evaluative angle for MU from an adversarial viewpoint. We propose identifying the data subset that presents the most significant challenge for influence erasure, i.e., pinpointing the worst-case forget set. Utilizing a bi-level optimization principle, we amplify unlearning challenges at the upper optimization level to emulate worst-case scenarios, while simultaneously engaging in standard training and unlearning at the lower level, achieving a balance between data influence erasure and model utility. Our proposal offers a worst-case evaluation of MU's resilience and effectiveness. Through extensive experiments across different datasets (including CIFAR-10, 100, CelebA, Tiny ImageNet, and ImageNet) and models (including both image classifiers and generative models), we expose critical pros and cons in existing (approximate) unlearning strategies. Our results illuminate the complex challenges of MU in practice, guiding the future development of more accurate and robust unlearning algorithms. The code is available at https://github.com/OPTML-Group/Unlearn-WorstCase.
Federated learning (FL), introduced in 2017, facilitates collaborative learning between non-trusting parties with no need for the parties to explicitly share their data among themselves. This allows training models on user data while respecting privacy regulations such as GDPR and CPRA. However, emerging privacy requirements may mandate model owners to be able to \emph{forget} some learned data, e.g., when requested by data owners or law enforcement. This has given birth to an active field of research called \emph{machine unlearning}. In the context of FL, many techniques developed for unlearning in centralized settings are not trivially applicable! This is due to the unique differences between centralized and distributed learning, in particular, interactivity, stochasticity, heterogeneity, and limited accessibility in FL. In response, a recent line of work has focused on developing unlearning mechanisms tailored to FL. This SoK paper aims to take a deep look at the \emph{federated unlearning} literature, with the goal of identifying research trends and challenges in this emerging field. By carefully categorizing papers published on FL unlearning (since 2020), we aim to pinpoint the unique complexities of federated unlearning, highlighting limitations on directly applying centralized unlearning methods. We compare existing federated unlearning methods regarding influence removal and performance recovery, compare their threat models and assumptions, and discuss their implications and limitations. For instance, we analyze the experimental setup of FL unlearning studies from various perspectives, including data heterogeneity and its simulation, the datasets used for demonstration, and evaluation metrics. Our work aims to offer insights and suggestions for future research on federated unlearning.
Understanding and shaping the behaviour of Large Language Models (LLMs) is increasingly important as applications become more powerful and more frequently adopted. This paper introduces a machine unlearning method specifically designed for LLMs. We introduce a selective pruning method for LLMs that removes neurons based on their relative importance on a targeted capability compared to overall network performance. This approach is a compute- and data-efficient method for identifying and removing neurons that enable specific behaviours. Our findings reveal that both feed-forward and attention neurons in LLMs are specialized; that is, for specific tasks, certain neurons are more crucial than others. Code from all experiments is available at https://github.com/nickypro/selective-pruning
The high cost of model training makes it increasingly desirable to develop techniques for unlearning. These techniques seek to remove the influence of a training example without having to retrain the model from scratch. Intuitively, once a model has unlearned, an adversary that interacts with the model should no longer be able to tell whether the unlearned example was included in the model's training set or not. In the privacy literature, this is known as membership inference. In this work, we discuss adaptations of Membership Inference Attacks (MIAs) to the setting of unlearning (leading to their "U-MIA" counterparts). We propose a categorization of existing U-MIAs into "population U-MIAs", where the same attacker is instantiated for all examples, and "per-example U-MIAs", where a dedicated attacker is instantiated for each example. We show that the latter category, wherein the attacker tailors its membership prediction to each example under attack, is significantly stronger. Indeed, our results show that the commonly used U-MIAs in the unlearning literature overestimate the privacy protection afforded by existing unlearning techniques on both vision and language models. Our investigation reveals a large variance in the vulnerability of different examples to per-example U-MIAs. In fact, several unlearning algorithms lead to a reduced vulnerability for some, but not all, examples that we wish to unlearn, at the expense of increasing it for other examples. Notably, we find that the privacy protection for the remaining training examples may worsen as a consequence of unlearning. We also discuss the fundamental difficulty of equally protecting all examples using existing unlearning schemes, due to the different rates at which examples are unlearned. We demonstrate that naive attempts at tailoring unlearning stopping criteria to different examples fail to alleviate these issues.
We present a machine unlearning approach that is both retraining- and label-free. Most existing machine unlearning approaches require a model to be fine-tuned to remove information while preserving performance. This is computationally expensive and necessitates the storage of the whole dataset for the lifetime of the model. Retraining-free approaches often utilise Fisher information, which is derived from the loss and requires labelled data which may not be available. Thus, we present an extension to the Selective Synaptic Dampening algorithm, substituting the diagonal of the Fisher information matrix for the gradient of the l2 norm of the model output to approximate sensitivity. We evaluate our method in a range of experiments using ResNet18 and Vision Transformer. Results show our label-free method is competitive with existing state-of-the-art approaches.
Machine unlearning can be useful for removing harmful capabilities and memorized text from large language models (LLMs), but there are not yet standardized methods for rigorously evaluating it. In this paper, we first survey techniques and limitations of existing unlearning evaluations. Second, we apply a comprehensive set of tests for the robustness and competitiveness of unlearning in the "Who's Harry Potter" (WHP) model from Eldan and Russinovich (2023). While WHP's unlearning generalizes well when evaluated with the "Familiarity" metric from Eldan and Russinovich, we find i) higher-than-baseline amounts of knowledge can reliably be extracted, ii) WHP performs on par with the original model on Harry Potter Q&A tasks, iii) it represents latent knowledge comparably to the original model, and iv) there is collateral unlearning in related domains. Overall, our results highlight the importance of comprehensive unlearning evaluation that avoids ad-hoc metrics.
Machine unlearning (MU) is to forget data from a well-trained model, which is practically important due to the ``right to be forgotten''. The unlearned model should approach the retrained model, where the forgetting data are not involved in the training process and hence do not contribute to the retrained model. Considering the forgetting data's absence during retraining, we think unlearning should withdraw their contribution from the pre-trained model. The challenge is that when tracing the learning process is impractical, how to quantify and detach sample's contribution to the dynamic learning process using only the pre-trained model. We first theoretically discover that sample's contribution during the process will reflect in the learned model's sensitivity to it. We then practically design a novel method, namely MU-Mis (Machine Unlearning by Minimizing input sensitivity), to suppress the contribution of the forgetting data. Experimental results demonstrate that MU-Mis can unlearn effectively and efficiently without utilizing the remaining data. It is the first time that a remaining-data-free method can outperform state-of-the-art (SoTA) unlearning methods that utilize the remaining data.
This study investigates the concept of the `right to be forgotten' within the context of large language models (LLMs). We explore machine unlearning as a pivotal solution, with a focus on pre-trained models--a notably under-researched area. Our research delineates a comprehensive framework for machine unlearning in pre-trained LLMs, encompassing a critical analysis of seven diverse unlearning methods. Through rigorous evaluation using curated datasets from arXiv, books, and GitHub, we establish a robust benchmark for unlearning performance, demonstrating that these methods are over $10^5$ times more computationally efficient than retraining. Our results show that integrating gradient ascent with gradient descent on in-distribution data improves hyperparameter robustness. We also provide detailed guidelines for efficient hyperparameter tuning in the unlearning process. Our findings advance the discourse on ethical AI practices, offering substantive insights into the mechanics of machine unlearning for pre-trained LLMs and underscoring the potential for responsible AI development.
Machine Learning models increasingly face data integrity challenges due to the use of large-scale training datasets drawn from the Internet. We study what model developers can do if they detect that some data was manipulated or incorrect. Such manipulated data can cause adverse effects including vulnerability to backdoored samples, systemic biases, and reduced accuracy on certain input domains. Realistically, all manipulated training samples cannot be identified, and only a small, representative subset of the affected data can be flagged. We formalize Corrective Machine Unlearning as the problem of mitigating the impact of data affected by unknown manipulations on a trained model, only having identified a subset of the corrupted data. We demonstrate that the problem of corrective unlearning has significantly different requirements from traditional privacy-oriented unlearning. We find most existing unlearning methods, including retraining-from-scratch without the deletion set, require most of the manipulated data to be identified for effective corrective unlearning. However, one approach, Selective Synaptic Dampening, achieves limited success, unlearning adverse effects with just a small portion of the manipulated samples in our setting, which shows encouraging signs for future progress. We hope our work spurs research towards developing better methods for corrective unlearning and offers practitioners a new strategy to handle data integrity challenges arising from web-scale training. Code is available at https://github.com/drimpossible/corrective-unlearning-bench.
The technological advancements in diffusion models (DMs) have demonstrated unprecedented capabilities in text-to-image generation and are widely used in diverse applications. However, they have also raised significant societal concerns, such as the generation of harmful content and copyright disputes. Machine unlearning (MU) has emerged as a promising solution, capable of removing undesired generative capabilities from DMs. However, existing MU evaluation systems present several key challenges that can result in incomplete and inaccurate assessments. To address these issues, we propose UnlearnCanvas, a comprehensive high-resolution stylized image dataset that facilitates the evaluation of the unlearning of artistic styles and associated objects. This dataset enables the establishment of a standardized, automated evaluation framework with 7 quantitative metrics assessing various aspects of the unlearning performance for DMs. Through extensive experiments, we benchmark 9 state-of-the-art MU methods for DMs, revealing novel insights into their strengths, weaknesses, and underlying mechanisms. Additionally, we explore challenging unlearning scenarios for DMs to evaluate worst-case performance against adversarial prompts, the unlearning of finer-scale concepts, and sequential unlearning. We hope that this study can pave the way for developing more effective, accurate, and robust DM unlearning methods, ensuring safer and more ethical applications of DMs in the future. The dataset, benchmark, and codes are publicly available at https://unlearn-canvas.netlify.app/.
Through pretraining on a corpus with various sources, Large Language Models (LLMs) have gained impressive performance. However, the impact of each component of the pretraining corpus remains opaque. As a result, the organization of the pretraining corpus is still empirical and may deviate from the optimal. To address this issue, we systematically analyze the impact of 48 datasets from 5 major categories of pretraining data of LLMs and measure their impacts on LLMs using benchmarks about nine major categories of model capabilities. Our analyses provide empirical results about the contribution of multiple corpora on the performances of LLMs, along with their joint impact patterns, including complementary, orthogonal, and correlational relationships. We also identify a set of ``high-impact data'' such as Books that is significantly related to a set of model capabilities. These findings provide insights into the organization of data to support more efficient pretraining of LLMs.
The rapid advancement of Large Language Models (LLMs) has demonstrated their vast potential across various domains, attributed to their extensive pretraining knowledge and exceptional generalizability. However, LLMs often encounter challenges in generating harmful content when faced with problematic prompts. To address this problem, existing work attempted to implement a gradient ascent based approach to prevent LLMs from producing harmful output. While these methods can be effective, they frequently impact the model utility in responding to normal prompts. To address this gap, we introduce Selective Knowledge negation Unlearning (SKU), a novel unlearning framework for LLMs, designed to eliminate harmful knowledge while preserving utility on normal prompts. Specifically, SKU is consisted of two stages: harmful knowledge acquisition stage and knowledge negation stage. The first stage aims to identify and acquire harmful knowledge within the model, whereas the second is dedicated to remove this knowledge. SKU selectively isolates and removes harmful knowledge in model parameters, ensuring the model's performance remains robust on normal prompts. Our experiments conducted across various LLM architectures demonstrate that SKU identifies a good balance point between removing harmful information and preserving utility.
We explore machine unlearning (MU) in the domain of large language models (LLMs), referred to as LLM unlearning. This initiative aims to eliminate undesirable data influence (e.g., sensitive or illegal information) and the associated model capabilities, while maintaining the integrity of essential knowledge generation and not affecting causally unrelated information. We envision LLM unlearning becoming a pivotal element in the life-cycle management of LLMs, potentially standing as an essential foundation for developing generative AI that is not only safe, secure, and trustworthy, but also resource-efficient without the need of full retraining. We navigate the unlearning landscape in LLMs from conceptual formulation, methodologies, metrics, and applications. In particular, we highlight the often-overlooked aspects of existing LLM unlearning research, e.g., unlearning scope, data-model interaction, and multifaceted efficacy assessment. We also draw connections between LLM unlearning and related areas such as model editing, influence functions, model explanation, adversarial training, and reinforcement learning. Furthermore, we outline an effective assessment framework for LLM unlearning and explore its applications in copyright and privacy safeguards and sociotechnical harm reduction.
We introduce a novel machine unlearning framework founded upon the established principles of the min-max optimization paradigm. We capitalize on the capabilities of strong Membership Inference Attacks (MIA) to facilitate the unlearning of specific samples from a trained model. We consider the scenario of two networks, the attacker $\mathbf{A}$ and the trained defender $\mathbf{D}$ pitted against each other in an adversarial objective, wherein the attacker aims at teasing out the information of the data to be unlearned in order to infer membership, and the defender unlearns to defend the network against the attack, whilst preserving its general performance. The algorithm can be trained end-to-end using backpropagation, following the well known iterative min-max approach in updating the attacker and the defender. We additionally incorporate a self-supervised objective effectively addressing the feature space discrepancies between the forget set and the validation set, enhancing unlearning performance. Our proposed algorithm closely approximates the ideal benchmark of retraining from scratch for both random sample forgetting and class-wise forgetting schemes on standard machine-unlearning datasets. Specifically, on the class unlearning scheme, the method demonstrates near-optimal performance and comprehensively overcomes known methods over the random sample forgetting scheme across all metrics and multiple network pruning strategies.
Large-scale diffusion models, known for their impressive image generation capabilities, have raised concerns among researchers regarding social impacts, such as the imitation of copyrighted artistic styles. In response, existing approaches turn to machine unlearning techniques to eliminate unsafe concepts from pre-trained models. However, these methods compromise the generative performance and neglect the coupling among multi-concept erasures, as well as the concept restoration problem. To address these issues, we propose a Separable Multi-concept Eraser (SepME), which mainly includes two parts: the generation of concept-irrelevant representations and the weight decoupling. The former aims to avoid unlearning substantial information that is irrelevant to forgotten concepts. The latter separates optimizable model weights, making each weight increment correspond to a specific concept erasure without affecting generative performance on other concepts. Specifically, the weight increment for erasing a specified concept is formulated as a linear combination of solutions calculated based on other known undesirable concepts. Extensive experiments indicate the efficacy of our approach in eliminating concepts, preserving model performance, and offering flexibility in the erasure or recovery of various concepts.
This paper explores Machine Unlearning (MU), an emerging field that is gaining increased attention due to concerns about neural models unintentionally remembering personal or sensitive information. We present SeUL, a novel method that enables selective and fine-grained unlearning for language models. Unlike previous work that employs a fully reversed training objective in unlearning, SeUL minimizes the negative impact on the capability of language models, particularly in terms of generation. Furthermore, we introduce two innovative evaluation metrics, sensitive extraction likelihood (S-EL) and sensitive memorization accuracy (S-MA), specifically designed to assess the effectiveness of forgetting sensitive information. In support of the unlearning framework, we propose efficient automatic online and offline sensitive span annotation methods. The online selection method, based on language probability scores, ensures computational efficiency, while the offline annotation involves a two-stage LLM-based process for robust verification. In summary, this paper contributes a novel selective unlearning method (SeUL), introduces specialized evaluation metrics (S-EL and S-MA) for assessing sensitive information forgetting, and proposes automatic online and offline sensitive span annotation methods to support the overall unlearning framework and evaluation process.
Tree-based machine learning models, such as decision trees and random forests, have been hugely successful in classification tasks primarily because of their predictive power in supervised learning tasks and ease of interpretation. Despite their popularity and power, these models have been found to produce unexpected or discriminatory outcomes. Given their overwhelming success for most tasks, it is of interest to identify sources of their unexpected and discriminatory behavior. However, there has not been much work on understanding and debugging tree-based classifiers in the context of fairness. We introduce FairDebugger, a system that utilizes recent advances in machine unlearning research to identify training data subsets responsible for instances of fairness violations in the outcomes of a random forest classifier. FairDebugger generates top-$k$ explanations (in the form of coherent training data subsets) for model unfairness. Toward this goal, FairDebugger first utilizes machine unlearning to estimate the change in the tree structures of the random forest when parts of the underlying training data are removed, and then leverages the Apriori algorithm from frequent itemset mining to reduce the subset search space. We empirically evaluate our approach on three real-world datasets, and demonstrate that the explanations generated by FairDebugger are consistent with insights from prior studies on these datasets.
To comply with AI and data regulations, the need to forget private or copyrighted information from trained machine learning models is increasingly important. The key challenge in unlearning is forgetting the necessary data in a timely manner, while preserving model performance. In this work, we address the zero-shot unlearning scenario, whereby an unlearning algorithm must be able to remove data given only a trained model and the data to be forgotten. We explore unlearning from an information theoretic perspective, connecting the influence of a sample to the information gain a model receives by observing it. From this, we derive a simple but principled zero-shot unlearning method based on the geometry of the model. Our approach takes the form of minimising the gradient of a learned function with respect to a small neighbourhood around a target forget point. This induces a smoothing effect, causing forgetting by moving the boundary of the classifier. We explore the intuition behind why this approach can jointly unlearn forget samples while preserving general model performance through a series of low-dimensional experiments. We perform extensive empirical evaluation of our method over a range of contemporary benchmarks, verifying that our method is competitive with state-of-the-art performance under the strict constraints of zero-shot unlearning. Code for the project can be found at https://github.com/jwf40/Information-Theoretic-Unlearning