Machine Unlearning

Ask about this topic

Overview

Machine unlearning asks how to remove a person’s data, a class, or a learned concept from a model after training, often to support privacy, safety, or changing data rights. Approaches range from exact or certified deletion methods, such as data partitioning and noisy or influence-based updates, to approximate parameter edits, representation changes, and inference-time steering; some systems are designed from the outset to make deletion easier. Researchers evaluate methods against retrained models and test forgetting, retained utility, privacy attacks, recovery, and repeated deletion across language, vision, diffusion, graph, and multimodal systems. A central tension is that suppressed outputs do not necessarily mean information is erased, while stronger removal can damage related knowledge; scalable methods with reliable guarantees and low collateral harm remain a challenge.

Latest papers

CardsList

354 papers58 papers in the last four weeks, up 76% on the four weeks before. 0.3% of all new papers.

Jul 2, 2026

LACUNA: A Testbed for Evaluating Localization Precision for LLM Unlearning

LLMs memorize sensitive training data, including personally identifiable information (PII), creating a pressing need for reliable post hoc removal methods. Unlearning has emerged as a promising solution, with state-of-the-art(SOTA) methods often following a localize-first, unlearn-second paradigm that targets specific model parameters. However, existing benchmarks evaluate unlearning solely at the output level, leaving open the question of whether unlearning truly erases knowledge from a model's parameters or merely obfuscates it, a concern reinforced by the success of resurfacing attacks. To bridge this gap, we introduce LACUNA: the first unlearning testbed with ground-truth parameter-level localization. LACUNA injects PII of synthetic individuals into predefined parameters of 1B and 7B OLMo-based models via masked continual pretraining, enabling direct evaluation of whether unlearning targets the weights responsible for knowledge storage. We use LACUNA to benchmark current SOTA unlearning methods and find that, despite strong output-level performance, existing methods are highly imprecise and susceptible to resurfacing attacks. We further show that when localization is successful, even a simple gradient-based unlearning method achieves strong erasure and robustness to resurfacing attacks, highlighting the importance of precise unlearning. We release LACUNA to complement behavioral evaluations and drive further advances in robust, localization-based unlearning.

Matteo Boglioni, Thibault Rousset, Siva Reddy +2

Mila – Quebec Artificial Intelligence Institute · McGill University · McGill-NLP/LACUNA

Privacy Leakage in Language ModelsMachine UnlearningLLM UnlearningMachine Unlearning Evaluation
AskPDF
Jun 30, 2026

DriftScope: Measuring The Hidden Effects of Diffusion Model Adaptation

Adapting pre-trained text-to-image diffusion models, whether to learn new visual concepts or erase unwanted ones, is routinely evaluated on its intended effects alone. We argue this framing is incomplete. Through sparse autoencoder analysis and zero-shot classification, we demonstrate that adaptation systematically damages semantically unrelated concepts in ways that aggregate metrics structurally cannot surface: when damage is severe enough for FID and KID to respond, the model is already nearly unusable; when the model remains functional, FID and KID stay flat while specific classes silently suffer worst-case zero-shot accuracy drops of up to 18.9 points and concept-level distributions shift dramatically. This pattern appears at both ends of the adaptation spectrum (concept customization and concept unlearning), suggesting it is a systematic consequence of weight-level modification rather than an artifact of any particular method. To surface this hidden drift before deployment, we introduce DriftScope, a prompt-level diagnostic tool that takes any two model checkpoints and returns a ranked list of tokens whose visual concepts have shifted most between them. DriftScope optimizes a soft prompt to attribute drift at the token level without requiring access to real data or model internals. The result is an interpretable, concept-level audit that aggregate evaluation cannot provide.

Héctor Laria, Yiping Han, Julian D. Santamaria +4

Computer Vision Center, Barcelona, Spain · Universitat Autonoma de Barcelona, Barcelona, Spain · Program of Computer Science, City University of Hong Kong (Dongguan), China +1

Diffusion Model AdaptationDiffusion Model UnlearningMachine UnlearningT2I Diffusion ModelsDiffusion Models
AskPDF
Jun 30, 2026

Look But Don't Touch with Sparse Autoencoders for Unlearning in Diffusion Models

Sparse autoencoders (SAEs) have recently been proposed as interpretable tools for concept-level manipulation, under the assumption that isolated features can serve as controllable intervention points. In this work, we systematically evaluate this assumption in the context of object erasure and steering in diffusion models. We show that while SAEs reliably detect and localize semantic concepts within diffusion model activations, direct intervention in their latent space frequently induces out-of-distribution activations, resulting in severe visual artifacts. To disentangle detection from intervention, we use SAE activations purely as semantic detectors to identify image regions containing the target object, and replace those patch embeddings with the ones that do not contain it. This detection-based replacement preserves the diffusion model's activation statistics and produces significantly cleaner erasure results than latent steering. Our findings reveal a fundamental gap between concept detection and concept intervention in diffusion models: monosemantic or sparse features are not inherently suitable as control knobs for steering. These results position SAEs as powerful interpretability tools for analyzing generative models, but highlight important limitations when used for direct manipulation, such as unlearning.

Enrico Cassano, Riccardo Renzulli, Rayyan Ahmed +2

University of Turin, Computer Science Department, Italy · LTCI, Télécom Paris, Institut Polytechnique de Paris, France

Sparse AutoencodersDiffusion Model UnlearningConcept ErasureMachine UnlearningDiffusion Models
AskPDF
Jun 29, 2026

IREU: Identity-Related Encoder-Only Unlearning for Customized Portrait Generation

Customized Portrait Generation (CPG) technologies have been widely used to generate high-fidelity person images given an input image indicating the identity and a text prompt indicating the required edits. Yet these methods pose significant privacy risks by spreading fake visual information. Against such risks, each public generator should be able to suppress its generation ability for a particular person when requested. Therefore, in this work we investigate the identity unlearning problem for CPG. Since there are no previous methods in this field, we propose a simple baseline that updates the image encoder by minimizing identity similarity between generated and input images for target identities to be unlearned, while maximizing it for identities to be retained. However, we find such a global perturbation in the feature space harms the fidelity of generated images for other identities to be retained. To solve this problem, we propose a novel method IREU, which first locates identity-related features in an offline manner and then only performs feature perturbations on them. The experimental results show that our proposed method IREU achieves better identity unlearning performance for target identities to be unlearned, and also keeps high fidelity for other identities to be retained. In addition, our unlearned image encoder is generalizable across different generators with the same encoder without fine-tuning, which is friendly for deployment in practice.

Chaoyi Shi, Shanshan Zhang, Jian Yang

PCA Lab, School of Computer Science and Engineering, Nanjing University of Science and Technology · PCA Lab, School of Intelligence Science and Technology, Nanjing University

Subject-Driven Image GenerationMachine Unlearning
AskPDF
Jun 29, 2026

The Forgetting-Retention Dilemma: Certified Unlearning Theory in Continual Learning

Machine unlearning aims to eliminate the influence of specific data from trained models to safeguard privacy. However, this presents a significant challenge in the context of continual learning (CL), where models update sequentially on dynamic datasets. A major limitation is that current certified unlearning algorithms fail to account for the complex, cumulative model evolution inherent to CL framework. In this work, we establish the first theoretical foundation bridging CL and machine unlearning. We formulate the CL's unlearning objective as the minimization of post-unlearning excess risk, which decomposes into CL excess risk and unlearning loss, characterizing the fundamental trade-off between preserving historical knowledge and targeted forgetting. Under mild assumptions, we first establish an upper bound for the CL excess risk in non-convex models. We then adapt two certified unlearning approaches, gradient-based and Hessian-based, to the CL framework. Our analysis reveals that while the gradient-based approach is less effective than the Hessian-based method in minimizing unlearning loss, it offers the distinct advantage of nearly zero storage overhead for enabling unlearning. This insight motivates a hybrid strategy that reduces storage costs while maintaining post-unlearning performance. Experimental results further validate our theoretical findings.

Yiting Hu, Lingjie Duan, Qian Zhang

Singapore University of Technology and Design, Singapore · The Hong Kong University of Science and Technology (Guangzhou), Guangzhou, China.

Continual LearningStatistical Learning TheoryMachine Unlearning
AskPDF
Jun 26, 2026

CBD: API-Only LLM Black-Box Unlearning through Controlled Behavioral Divergence

Edge devices increasingly invoke large language models (LLMs) through API services for context aware edge intelligence, while edge generated data may be collected to improve LLMs and may introduce sensitive, copyrighted, harmful, or outdated information into model behavior. Machine unlearning offers a practical way to remove the influence of undesired data without retraining LLMs. However, existing methods still face two gaps. The first is API only black box access, where target model parameters and internal logits are unavailable. The second is how to preserve retained utility when unlearning target data and retained data share highly similar prompt structures or semantic patterns. To address these challenges, we propose Controlled Behavioral Divergence (CBD), an API only black box unlearning framework. CBD uses two auxiliary models to create controlled behavioral divergence between retained inputs and unlearning target inputs, converts this divergence into an unlearning relevance score, and routes unlearning related prompts away from the target LLM. To improve discrimination accuracy under high similarity between target and retained data, CBD constructs a gradient statistics based discriminative basis by estimating empirical Fisher matrices and solving a regularized generalized eigenvalue problem, guiding the unlearning signal toward target specific information rather than shared prompt structures. Compared with eleven white box and gray box unlearning baselines, CBD achieves a better unlearning utility trade off and its performance varies little across settings. On ToFU forget10, CBD approaches the retrained reference on the forget set while raising model utility to 74.90, about 15% above the second best baseline. On WMDP, it lowers hazardous knowledge accuracy to 25.68, near random guessing, while preserving MMLU accuracy of 52.67. Code is at https://github.com/DGL-codes/CBD.

Zhiqiang Xie, Yijing Lin, Zhipeng Gao +1

State Key Laboratory of Networking and Switching Technology, Beijing University of Posts and Telecommunications · Beijing Advanced Innovation Center for Future Blockchain and Privacy Computing · Department of Electronic and Electrical Engineering, Sungkyunkwan University, Suwon, South Korea

Machine UnlearningLLM Unlearning
AskPDF
Jun 24, 2026

DFMU: Data-Frugal Machine Unlearning

Machine unlearning is an emerging domain that ensures the safe removal of elements (includes concepts, attributes, entity and class) from the trained model along with least drop in model performance. The domain of machine unlearning brings its own indigenous challenges since the removal of pre-trained elements from model will always degrade the model performance on remaining elements. The existing methods basically rely on retraining for removal of elements from the pre-trained model, which is compute extensive. In this work, we propose a machine unlearning method which helps to reduce the computational requirement for faster retain-dataset accuracy convergence which also does not require extensive retraining of the pre-trained model. The proposed method, Data-Frugal Machine Unlearning (DFMU) requires only a single forward and backward pass for computing the importance score of various computational blocks of a model. The importance score computation is based on knowledge preserving pruning which helps to converge faster and requires far less data as compared to the existing methods. Experimentally, it achieves 40% more retain-accuracy with just 13% of data samples in comparison with SOTA method on various public datasets and also averages 88% faster processing time for forgetting a given class.

Sajith U, Prateek Keserwani

Samsung R&D Institute India-Bangalore, India

Machine Unlearning
AskPDF
Jun 23, 2026

Erased, but Not Gone: Output Forgetting Is Not True Forgetting

Machine unlearning (MU) is commonly judged by output forgetting, such as low forget-set accuracy or reduced membership inference. But if output-level success can coexist with retraining-inconsistent residuals in representation space, what kind of forgetting are current evaluations actually certifying? We study this question through retraining-consistent representation forgetting, using a population of retrained models (i.e., trained from scratch without the forget data) as an operational reference distribution for ideal forgetting. Across multiple unlearning methods, datasets, and models, our theoretical analysis and empirical results show that standard output-level evaluation can systematically overestimate the success of unlearning. Under this stronger lens, current methods often appear forgotten at the output layer while exhibiting a structured mismatch relative to retraining. They partially align with retraining on forget samples, remain more inconsistent on retain samples, and tend to leave residual discrepancy localized along retraining-related directions rather than diffuse in representation space. This structured mismatch is characterized by forget/retain asymmetry, directional mismatch, and concentrated residuals along retraining-related directions. In many cases, the discrepancies exceed the variability of independent retraining and lead to a risk of information recovery during relearning attacks. Overall, the results suggest that current MU is often evaluated for apparent forgetting rather than retraining-consistent forgetting. More broadly, retraining reveals what output forgetting hides.

Teresa Pui Yee Yong, Win Kent Ong, Chee Seng Chan

Universiti Malaya, Kuala Lumpur, Malaysia · VinUniversity, Hanoi, Vietnam

Machine UnlearningMachine Unlearning Evaluation
AskPDF
Jun 23, 2026

Co-occurring associated retained concepts in Diffusion Unlearning

Unlearning has emerged as a key technique to mitigate harmful content generation in diffusion models. However, existing methods often remove not only the target concept, but also benign co-occurring concepts. As illustrated in Fig.1, unlearning nudity can unintentionally suppress the concept of person, preventing a model from generating images with person. We define these undesirably suppressed co-occurring concepts that must be preserved CARE (Co-occurring Associated REtained concepts). Then, we introduce the CARE score, a general metric that directly quantifies their preservation across unlearning tasks. With this foundation, we propose ReCARE (Robust erasure for CARE), a framework that explicitly safeguards CARE while erasing only the target concept. ReCARE automatically constructs the CARE-set, a curated vocabulary of benign co-occurring tokens extracted from target images, and leverages this vocabulary during training for stable unlearning. Extensive experiments across various target concepts (Nudity, Van Gogh style, and Tench object) demonstrate that ReCARE achieves overall state-of-the-art performance in balancing robust concept erasure, overall utility, and CARE preservation.

Miso Kim, Georu Lee, Yunji Kim +3

Dongguk University-Seoul · Chung-Ang University · Korea Institute for Advanced Study

Diffusion Model UnlearningConcept ErasureMachine UnlearningDiffusion Models
AskPDF
Jun 23, 2026

FedUP: One-Shot Federated Unlearning via Centroid-Guided Plug-in Filters

Federated unlearning (FU) is critical for complying with legal mandates like the right to be forgotten in decentralized systems, yet current methods face a persistent dilemma between non-target knowledge loss and high request latency. To resolve these issues, we propose FedUP, a one-shot federated unlearning framework utilizing lightweight pluggable filters that act as a "knowledge funnel" to screen out target data while preserving original model performance. By freezing original model parameters and training filters at the server side using differentially private (DP)-protected class centroid samples, FedUP bypasses the need for multi-round client-server communication and complex retraining, reducing unlearning latency from minutes to mere seconds. Additionally, the framework's pluggable architecture ensures inherent reversibility, enabling the seamless restoration of forgotten knowledge by simply removing the filters. Extensive experiments on diverse image and text tasks demonstrate that FedUP effectively reduces non-target knowledge loss and achieves superior unlearning precision and efficiency across various scenarios. Code is available at: https://github.com/suows/FedUP-code.

Feihong Nan, Zhengyi Zhong, Pan Wang +4

1National Key Laboratory of Big Data and Decision, National University of Defense Technology, China.

Federated UnlearningMachine UnlearningFederated Learning
AskPDF
Jun 23, 2026

Selective Capability Unlearning in End-to-End Spoken Language Understanding

Modern spoken language understanding (SLU) systems are increasingly deployed in real-world settings, where specific functionalities may need to be removed due to policy or safety constraints. In SLU, a functionality corresponds to an intent and its associated slot-generation behavior. However, in autoregressive models, suppressing a target intent does not eliminate the conditional mapping that generates slots conditioned on that intent. When the intent prefix is externally supplied, the model can reconstruct the original intent-slot structure. We identify this structural failure as capability persistence. We propose Binding Subspace (BSU), a representation-level framework that isolates and attenuates intent-conditioned directions underlying this mapping. Across SLU benchmarks, BSU substantially reduces forced-prefix recoverability while preserving retained performance.

Akanksha Singh, Vinod Kumar Kurmi

Indian Institute of Science Education and Research Bhopal, India

Spoken Language UnderstandingMachine Unlearning
AskPDF
Jun 17, 2026

GDGU: A Gradient Difference-based Graph Unlearning Method for Cyberattack Localization in Electric Vehicle Charging Networks

Electric vehicle charging stations (EVCSs) can expose distribution feeders to cyberattacks. While machine learning methods, including graph neural networks, can localize which bus is compromised, significant challenges remain in data sharing and model training. For example, privacy regulations grant EVCS owners the right to delete their training data from a deployed model, yet retraining from scratch on every request is computationally prohibitive. To address this, we study graph unlearning (GU) for EVCS cyberattack localization, formulated as a feature-level unlearning problem on a graph-level multi-label classification task. Specifically, we propose gradient difference-based graph unlearning (GDGU), which removes the influence of the requested deletion data through a first-order parameter correction. The correction is computed from the gradient difference between the original training data and a modified dataset in which only the charging power features at the requested EVCS buses are unlearned. Then, a batch-normalization recalibration and a brief recovery fine-tuning step are applied to restore localization utility. We benchmark GDGU against two second-order GU baselines on the IEEE 34-bus, 123-bus, and 8500-node distribution networks across three graph neural network backbones and cumulative unlearning scenarios. GDGU matches the strongest baseline on localization utility and reaches forgetting fidelity close to full-retraining, while unlearning 10 to 12 times faster than retraining from scratch and using far less memory than the second-order GU baselines.

Nanhong Liu, Mucun Sun, Jie Zhang

Department of Mechanical Engineering, The University of Texas at Dallas, Richardson, TX 75080, USA · Idaho National Laboratory, Idaho Falls, ID 83415, USA · Department of Mechanical Engineering and (Affilated) Department of Electrical and Computer Engineering, The University of Texas at Dallas, Richardson, TX 75080, USA

Graph Neural NetworksGraph UnlearningElectric Vehicle ChargingCybersecurityMachine UnlearningPower Systems
AskPDF
Jun 17, 2026

Mechanism-Guided Selective Unlearning for RLVR-Induced Reasoning

We propose MAST (Mechanism-Aligned Selective Targeting), a mechanism-guided method for unlearning RLVR-induced reasoning with substantially lower collateral damage than standard full-parameter updates. In matched SFT/RLVR checkpoints on Qwen2.5-Math-1.5B and Qwen3-1.7B-Base, the SFT-to-RLVR increment differs sharply from the SFT update in token-level delta-log-probability, and full-parameter gradient ascent forgets only by damaging retain MATH and GSM8K. MAST ranks attention-projection tensors by off-principal energy, update magnitude, and forget-gradient coupling magnitude, then updates only the top-ranked subset. On the primary model, MAST induces statistically significant target forgetting (MATH forget 45/150 to 37/150; McNemar p=0.0078) while preserving GSM8K (+0.8 pp) and MATH retain (-0.5 pp). The advantage reproduces across seeds, NPO/SimNPO objectives, and Qwen3, where MAST preserves GSM8K while full-parameter unlearning collapses it.

Chenyu Zhou, Qiliang Jiang, Shuning Wu +1

School of Engineering, Institute of Science Tokyo, Japan · College of Control Science and Engineering, Zhejiang University, China · Department of Electrical and Computer Engineering, National University of Singapore, Singapore

Machine UnlearningLLM Unlearning
AskPDF
Jun 17, 2026

Machine Unlearning for the XGBoost Model with Network Intrusion Datasets

Machine Unlearning (MU) has emerged as an important technique for removing specific data points from trained models without requiring full retraining. However, most existing MU research focuses on deep learning and image data, leaving a gap in the domain of network intrusion detection, which relies heavily on tabular data. This work introduces XGBoost-Forget, an unlearning approach for the XGBoost model, to address this gap. The approach is evaluated on two tabular Network Intrusion (NI) datasets, IoT-23 and GeNIS, using multiple metrics to assess model performance, unlearning efficiency, and forgetting quality. The results show that XGBoost-Forget maintains predictive performance close to the original model while providing significantly faster unlearning, demonstrating its potential for MU in tabular NI settings.

Diana Magalhães, Eva Maia, João Vitorino +1

GECAD, ISEP, Polytechnic of Porto, Rua Dr. António Bernardino de Almeida, 4249-015 Porto, Portugal

Machine Unlearning
AskPDF
Jun 16, 2026

PreUnlearn: Auditing Collateral Knowledge Damage Before Large Language Model Unlearning

Machine unlearning for large language models (LLMs) aims to remove specified knowledge while preserving the rest of the model's capabilities. However, the boundary between knowledge to forget and knowledge to retain is often unclear, since related and even distant information may be entangled in the model. In this paper, we study LLM unlearning from a data-centric perspective and measure how unlearning effects propagate from the forget set to same-domain and distant-domain knowledge not after but before unlearning. We find a consistent decay pattern: collateral damage is strongest near the forget set, weakens with semantic distance, but does not disappear at domain boundaries. We further ask whether such damage can be audited before unlearning is executed. We formulate forget-set auditing as a pre-unlearning prediction task and analyze which data features are most predictive of downstream damage. Our results show that interaction features between the forget set and evaluation set provide the strongest signals, suggesting that collateral damage is partly reflected in data geometry before model updates occur. These findings position forget-set auditing as an early warning tool for identifying risky unlearning runs and designing more reliable unlearning procedures. Code and data are available at https://github.com/BartSu/PreUnlearn.

Bo Su, Ankit Shah, Thai Le

Indiana University Bloomington, IN, USA

LLM AuditingMachine UnlearningLLM UnlearningMachine Unlearning Evaluation
AskPDF
Jun 16, 2026

SAGE: Retain-Aware Post-Hoc Sanitization of Final Unlearning Vector

Large Language Model (LLM) unlearning aims to remove undesirable knowledge or behaviors while preserving retained capabilities. Current unlearning methods all involve a trade-off between unlearning and retention. We have found that the retention activation bias can also be used to quantify the damage an unlearning method inflicts on retention, without considering the specific implementation of the unlearning process. This allows us to restore retention performance for any unlearning method using a post-hoc approach. Therefore, we propose a complementary post-hoc setting to sanitize the final update vector without rerunning the original unlearning pipeline. In this setting, we design SAGE, Spectral Activation-GEometry Sanitization, a source-agnostic correction for final unlearning updates. SAGE collects real module inputs from a small retain proxy, extracts their dominant activation geometry, and solves a source-anchored optimization objective in closed form, which suppresses update components aligned with high-energy retained directions while preserving the source method's forgetting carrier. Across multiple unlearning methods, model scales, and benchmarks, SAGE consistently relieves the retain-forget trade-off, identifying post-hoc sanitization of final vectors as a practical and underexplored axis for machine unlearning.

Jingyuan Zhang, Yucheng Bai, Peixi Wen +6

Institute of Image Processing and Pattern Recognition, Shanghai Jiao Tong University

Machine UnlearningLLM Unlearning
AskPDF
Jun 15, 2026

RepSelect: Robust LLM Unlearning via Representation Selectivity

Making large language models (LLMs) deeply forget specific knowledge and values without sacrificing general capabilities remains a central challenge in unlearning. Current methods are easily reversed by fine-tuning or few-shot prompting, suggesting their forgetting is only shallow. We identify the root cause. Existing methods target representations shared with both the retain set and the subspace recovered by a fine-tuning attacker, making unlearning both disruptive to general capabilities and easy to reverse. We propose RepSelect (Representation Selectivity), which isolates forget-set-specific representations by collapsing top principal components of weight gradients before each update, leaving general capabilities intact while limiting what fine-tuning can recover. We evaluate across two forget categories, biohazardous knowledge and abusive tendencies, and four model families spanning dense and Mixture-of-Experts architectures (Llama 3, Qwen 3.5, Gemma 4 E4B, DeepSeek V2 Lite). Compared to five popular baselines (GradDiff, NPO, SimNPO, RMU, UNDIAL), RepSelect achieves a 4-50x larger reduction in post-relearning answer accuracy than the strongest baseline, and is near-perfectly robust to few-shot prompting attacks. Targeting selective representations is thus an important step towards deep and robust LLM forgetting.

Filip Sondej, Yushi Yang, Adam Mahdi

Independent · University of Oxford

LLM SafetyRepresentation Geometry in Language ModelsMachine UnlearningLLM Unlearning
AskPDF
Jun 15, 2026

TrustErase: Auditable Instant Machine Unlearning with Passport-Embedded Representations

The demand for privacy-compliant AI has amplified the need for machine unlearning; yet, existing retraining or distillation-based methods remain unverifiable and computationally costly. We introduce TrustErase, a verifiable, data-free unlearning framework leveraging passport-embedded representations for instant, modular, and auditable forgetting. By treating passports as cryptographic keys within parameter-efficient adaptation layers, TrustErase enables the removal of specific classes or datasets through simple deactivation, without retraining, fine-tuning, or access to the original data. A singular value based decomposition conceals passports within model weights, ensuring that unlearning actions remain transparent and provably compliant. Evaluations on MNIST, CIFAR10 and CIFAR100 show that TrustErase matches or exceeds state-of-the-art benchmarks such as DELETE, L2UL, and Boundary Shrink, while operating in a strictly data-free regime. Ultimately, TrustErase establishes a new paradigm for trustworthy, accountable, and instantly forgettable AI systems.

Rutger Hendrix, Leonardo G. Russo, Concetto Spampinato +2

University of Catania Catania, Italy

Privacy-Preserving MLMachine Unlearning
AskPDF
Jun 15, 2026

To forget is to preserve: Machine Unlearning for 3D medical image segmentation

With new data privacy laws such as the General Data Protection Regulation (GDPR) [1] that allow individuals to ask that any of their personal information be erased from trained machine learning models, there has been a push to investigate the unlearning of data from models as a way to comply with these laws. In this regard, based on four mechanics, we consider several approximate unlearning strategies applied to the MRBrainS18 dataset [2]. We use a 3D ResNet-50 [3] as a backbone architecture for segmentation that has been pre-trained with the Med3D framework [4]. Considering the pre-trained model as a baseline, we evaluate respective retention accuracy on 2 types of subjects, i.e., retain and forget. We assess these approaches through their Dice similarity coefficient and mean absolute error (MAE) values using two separate training horizons 20 and 50 epochs. The results show that the Noisy Label strategy had the best overall trade-off with a decrease of 93% in the forget set while maintaining 84% accuracy for the retained set after 50 epochs. All other strategies showed extreme levels of forgetting at higher epoch numbers while also demonstrating catastrophic degradation of their retain set performance. The results of this study provide a strict baseline of performance metrics for unlearning on a subject-specific level and provide practitioners with clear criteria for selecting the proper strategies.

Nitesh Kumar Singh, Akhilesh Singh, Arjun Arora

School of Computer Science UPES, University of Tomorrow Dehradun, India.

Privacy-Preserving ML3D Medical Image SegmentationMachine Unlearning
AskPDF
Jun 15, 2026

Auditing Machine Unlearning: A Systematic Research on Whether Models Truly Forget

Machine unlearning has been extensively studied in response to growing privacy concerns and regulatory requirements. However, auditing whether unlearning algorithms have truly erased the influence of specific data remains an open challenge. The lack of reliable and practical auditing mechanisms can lead to critical privacy risks, such as residual information leakage. This paper initiates a systematic investigation into whether existing unlearning algorithms can truly forget the designated data. We propose the first practical and general-purpose auditing framework for machine unlearning, inspired by the concept of proof of ignorance. Our framework addresses the key practicality limitations of existing methods by eliminating the need for retraining-from-scratch baselines, avoiding the training of large numbers of shadow models, and requiring no intrusive intervention in the original training process. To evaluate the effectiveness of our framework, we first conduct validation experiments to verify its soundness and completeness. We then perform comprehensive experiments across six datasets and ten representative unlearning methods. The results demonstrate that our framework reliably distinguishes between successful and failed unlearning. In particular, we observe that retraining-based and fine-tuning-based methods can achieve effective unlearning, even when the target data remain in the original dataset. In contrast, de-optimization-based methods fail to achieve true unlearning and instead degrade the model's performance. Fisher/Hessian-based methods also fail to unlearn requested data, even formal certification is provided. Moreover, we show that our framework is robust against fake unlearning attempts and generalizes well to large language models.

Dayong Ye, Tianqing Zhu, Ruiding Huang +5

Machine UnlearningLLM UnlearningMachine Unlearning Evaluation
AskPDF
Jun 14, 2026

InstantForget: Update-Free Backdoor Unlearning with Inference-Time Feature Reset

Backdoor unlearning aims to remove a malicious trigger behavior from a deployed model while preserving clean utility. We study the update-free inference-time setting, where model parameters remain frozen. First, we audit a common projection assumption under oracle paired clean and triggered features. Projection succeeds mainly on BadNets and leaves WaNet, Blended, and SIG at 0.683, 0.888, and 0.941 ASR on CIFAR-10 ResNet-18. This failure is not explained by spectral compactness, spatial locality, or subspace misalignment. It is predicted by a logit-triplet gap involving the target margin, target-logit drop, and non-target logit rise. We then introduce InstantForget, a clean-calibrated gated reset that flags anomalous features with a Mahalanobis score and moves only flagged features toward a neutral non-target representation. With one fixed operating point selected on held-out triggered validation, InstantForget reduces average ASR to 0.071 across four non-adaptive CIFAR-10 triggers without triggered samples or parameter updates at deployment. It also reaches 0.981 detection AUROC and transfers to six of eight tested backbones. Reported failures under WaNet, ModelNet10 point blend, two backbone geometries, and adaptive feature-compactness attacks define the method's scope.

Zhenyu Yu

College of Computer Science and Artificial Intelligence Fudan University

Backdoor DefenseMachine Unlearning
AskPDF
Jun 13, 2026

Replay What Matters: Off-Policy Replay for Efficient LLM Reinforcement Unlearning

LLM unlearning has emerged as a cost-effective alternative to full retraining for removing hazardous knowledge from pretrained models while preserving general utility. Recent RL-based methods such as RULE reformulate unlearning as learning a refusal behavior, but their on-policy optimization repeatedly samples from the same forget and retain/boundary prompts throughout training. We identify a critical inefficiency in this process: easy cases quickly converge and provide little useful gradient signal, while hard cases near the forget/retain boundary continue to produce low-reward rollouts that are discarded after a single use. To address this issue, we propose ReRULE, an off-policy replay enhancement for reinforcement unlearning. ReRULE stores low-reward hard-case rollout groups in a replay buffer during early GRPO training and reuses them in later stages through importance-sampled off-policy updates, redirecting computation toward boundary cases that still require learning. Theoretically, we show that ReRULE yields a tighter hard-case convergence bound than pure on-policy RULE. Empirically, ReRULE improves MUSE-Books Retain Quality from 46.3 to 56.2 while adding only 5--11% training time across benchmarks. Its limited improvement on the simpler TOFU setting further supports the intended conditional behavior: replay is most beneficial when the hard/easy disparity is pronounced.

Zirui Pang, Chenlong Zhang, Haosheng Tan +3

The Hong Kong University of Science and Technology (Guangzhou) · Institute of Automation, Chinese Academy of Sciences · University of Glasgow

Experience ReplayMachine UnlearningPolicy LearningLLM UnlearningOff-Policy RL
AskPDF
Jun 12, 2026

Behavioral Audit of Machine Unlearning Has a Privacy Cost

The removal of learned data from Machine Learning models through Machine Unlearning (MU) has been widely studied; however, there has yet to be an agreed-upon scheme for auditing MU. Existing work has shown that a dishonest model owner can falsify evidence to avoid executing MU, while curious auditors (and adversaries) can infer the privacy-sensitive properties of the model and its training data even with limited access. Yet auditing of MU under mutual distrust between the model owner and the auditor remains unexplored. We provide an information-theoretic proof for this scenario: for convex ML models, a generic audit scheme that relies solely on querying the model for behavioral signals cannot identify insufficiently unlearned models without revealing membership information of the retained set. Therefore, auditing MU under the assumption of a dishonest model owner and an honest-but-curious auditor faces an inherent privacy-audit tradeoff. Our empirical results on convex models strongly supports this result, while further experiments demonstrate that this privacy-audit tension persists in non-convex models. Our results call for a more careful consideration of the privacy-audit tension under a realistic auditor threat model, and serve as a foundation for more scrutiny of designs of privacy-preserving audit schemes for the MU pipeline. We also release our code implementation at https://github.com/LiouTang/Behavioral-Unlearn-Audit.

Liou Tang, James Joshi, Ashish Kundu

University of Pittsburgh, Pittsburgh, PA, USA · Cisco Research, San Jose, CA, USA

Privacy AuditingMachine UnlearningMachine Unlearning Evaluation
AskPDF
Jun 12, 2026

Rethinking Backdoor Adversarial Unlearning through the Lens of Catastrophic Forgetting in Continual Learning

Existing studies reveal that current backdoor defenses exhibit limited robustness and often fail against specific types of attacks. More concerningly, prevailing safety tuning strategies tend to provide only superficial safety protection, as they fall short of completely eliminating the backdoor effects. In this work, we present a novel formulation of backdoor learning and unlearning as a sequential, three-stage process from a continual learning perspective. Within this framework, we formally define complete backdoor unlearning and further derive the necessary conditions for achieving it based on the mechanism of catastrophic forgetting. Guided by these insights, we propose Blind Inversion-Backdoor Adversarial Unlearning (BI-BAU), which formulates the generation of adversarial examples satisfying the unlearning conditions as a blind inversion problem. We solve this by integrating the bi-level optimization process of adversarial training into an Expectation-Maximization (EM) algorithm framework to optimize the maximum a posteriori (MAP) objective. Furthermore, BI-BAU is extended to untargeted adversarial scenarios with unknown target classes, as well as to multi-modal contrastive learning tasks, enhancing its applicability to real-world deployment scenarios where pre-trained models may be compromised. Extensive experiments demonstrate that our method exhibits general applicability across a wide spectrum of backdoor attacks and can effectively and thoroughly eliminate the backdoor effects from a backdoor model.

Zhenqian Zhu, Yamin Hu, Yujiang Liu +5

Harbin Institute of Technology, Shenzhen Shenzhen, Guangdong, China · Shenzhen Key Laboratory of Media Security, Shenzhen University Shenzhen, Guangdong, China

Backdoor DefenseMachine Unlearning
AskPDF
Jun 11, 2026

Natively Unlearnable Large Language Models

Unlearning aims to remove the influence of specific training data sources, but this has proved challenging because the contributions of different sources are entangled within the model. Isolating source contributions to disjoint parameters makes removal easier, though it obstructs joint learning across sources. We propose NULLs (Natively Unlearnable LLMs), a model class that satisfies the two opposing goals of isolating source-specific contributions and learning jointly across sources, by training a set of shared backbone neurons alongside a pool of sparsely activated sinks. During training, information specific to a source naturally concentrates in its sinks while information shared across sources accumulates in the backbone. A source is then unlearned at deployment by disabling its corresponding sinks, with no gradient updates and no access to the retained data. We show that NULLs scales to Wikipedia's ~6M articles, isolating each as an independent source. Unlearning a single article removes knowledge specific to it while preserving facts shared with semantically related articles, closely matching retraining from scratch. We note that unlearning with NULLs is also robust: in a case study of unlearning the Harry Potter books, NULLs resists both adversarial extraction and relearning that reverses post-hoc unlearning. Finally, NULLs preserves general language capabilities, matching a standard transformer on downstream benchmarks. Together, these results suggest that source-level unlearning need not be an afterthought. It can be built natively into LLM training while retaining the benefits of shared representation learning.

Gaurav R. Ghosal, Pratyush Maini, Aditi Raghunathan

Carnegie Mellon University · DatologyAI

Activation SparsityMachine UnlearningLLM UnlearningLLM Training
AskPDF
Jun 11, 2026

Reinforcement Learning for Neural Model Editing

Editing pretrained neural networks requires specialized algorithms tailored to specific objectives. Designing such algorithms is often time-consuming and demands significant effort. We present an exploratory framework that formulates neural model editing as a reinforcement learning problem, where agents modify models using reward feedback. We introduce two environments: MaskWorld, where agents scale weights multiplicatively, and ShiftWorld, where agents apply additive weight updates. The reward function combines a utility-preservation objective with a task-specific editing objective, enabling agents to learn targeted modifications while maintaining overall model performance. We evaluate the framework on bias mitigation in text classification and machine unlearning in image classification, both of which traditionally rely on specialized algorithms. Our results show that the learned policies reduce forget set accuracy to nearly 0% while preserving over 90% retain set accuracy on the unlearning task. In the bias mitigation setting, the learned policies improve bias-related performance by more than 5% while maintaining general classification utility. Our findings show that neural model editing can be cast as a reinforcement learning problem, allowing editing policies to be learned from reward feedback rather than manually engineered for each task.

Shaivi Malik

Reinforcement LearningModel EditingMachine Unlearning
AskPDF
Jun 11, 2026

MLUBench: A Benchmark for Lifelong Unlearning Evaluation in MLLMs

Multimodal large language models (MLLMs) are trained on massive multimodal data, making data unlearning increasingly important as data owners may request the removal of specific content. In practice, these requests often arrive sequentially over time, giving rise to the challenging problem of MLLM Lifelong Unlearning. However, most existing benchmarks are limited in scale and scope, failing to capture the complexities of MLLM lifelong unlearning. To fill this gap, we introduce the MLUBench, a large-scale and comprehensive benchmark featuring 127 entities across 9 classes under lifelong unlearning requests. We perform extensive experiments using MLUBench and reveal that existing unlearning methods suffer from severe, cumulative degradation. More critically, we further identify the unique challenge of this problem: unlike in unimodal models, MLLM lifelong unlearning is constrained by the need to preserve multimodal alignment. Continually unlearning from one modality could degrade the entire model. To alleviate this challenge, we propose LUMoE, an effective method. Experiments demonstrate that LUMoE significantly mitigates the degradation problem faced by baselines. The source code and the MLUBench dataset are open-sourced in https://github.com/lihe-maxsize/Lifelong_Unlearning_main.

He Li, Haoang Chi, Qizhou Wang +6

College of Computer Science and Technology, National University of Defense Technology, Changsha, China · TMLR Group, Hong Kong Baptist University, Hong Kong, China · School of Statistics and Data Science, Shanghai University of Finance and Economics, Shanghai, China +2

VLM UnlearningMultimodal Large Language ModelsMLLM UnlearningMachine UnlearningLLM UnlearningMachine Unlearning Evaluation
AskPDF
Jun 9, 2026

Routing-Aware Expert Calibration for Machine Unlearning in Mixture-of-Experts Language Models

Machine unlearning is increasingly important for large language models, yet unlearning in Mixture-of-Experts (MoE) architectures remains underexplored. Unlike dense models, MoE architectures employ a router at each layer to assign each token to a sparse subset of experts. In this work, we observe that forget data often activates a small subset of experts disproportionately, while these experts may receive much weaker activation from retain data. This forget--retain routing mismatch can leave forget-critical experts under-regularized during unlearning. To address this, we propose TRACE, Targeted Routing-Aware Calibration of Experts, for MoE unlearning. TRACE first detects forget-critical experts from offline activation statistics, and then calibrates retain regularization by reweighting token-level retain losses so that each selected expert's retain-side activation frequency better matches its forget-side counterpart. Experiments on WMDP and MUSE-BOOKS across multiple MoE LLMs show that TRACE consistently improves the forget-utility trade-off, yielding a 9% relative utility improvement over the strongest baseline under comparable forgetting quality and the best performance on three out of four MUSE-BOOKS metrics.

Jingyi Xie, Yijun Lin, Yinjiang Xiong +2

Renmin University of China · Lightstandard · Zhejiang University +1

Mixture-of-Experts Language ModelsMachine UnlearningLLM Unlearning
AskPDF
Jun 5, 2026

REMEDI: A Benchmark for Retention and Unlearning Evaluation in Multi-label Clinical Disease Inference

Language models trained for clinical disease inference are trained on patient data, which may include sensitive and private information, and data owners may request the removal of their data from a trained model due to privacy or copyright concerns. However, exactly unlearning patient-specific data is intractable, and retraining with minor data removal is resource-intensive. While there exists several machine unlearning methods that can be used, their utility is generally restricted to non-medical domains. Moreover, the existing benchmarks for evaluating such unlearning methods primarily utilize synthetically curated datasets, which are not truly representative of real-world systems. Hence, the effectiveness of these unlearning methods in the medical domain is largely unclear. To this end, we introduce REMEDI, an extensive benchmark for machine unlearning tailored to multi-label and multiclass clinical disease inference, where label correlations, longitudinal structure, and safety constraints make unlearning particularly challenging. Unlike the existing benchmarks, REMEDI considers: (1) a relevant application domain (medical), (2) comprehensive unlearning setups involving diverse sets of forget instances, (3) challenging unlearning scenarios including multi-label and multi-class classification tasks, and (4) evaluation metrics involving performance both in terms of utility and extent of unlearning achieved. REMEDI is developed using the MIMIC-III clinical database that contains comprehensive clinical data of patients. Experiments with existing unlearning methods indicate that there exists a trade-off between utility and unlearning performance. They are also largely unsuited to multi-label classification tasks. To facilitate reproducibility, we make our benchmark publicly available.

Anurag Sharma, Sai Teja Chunchu, Prasenjit Mitra +2

IIT Kharagpur · India · Carnegie Mellon University +4

Multi-Label ClassificationMachine UnlearningMachine Unlearning Evaluation
AskPDF
Jun 5, 2026

TRACER: Token ReAssignment for Concept ERasure in Generative Recommendation

Generative recommendation formulates next-item prediction as autoregressive generation over semantic ID (SID) sequences derived from users' historical interactions, making modern recommender systems structurally similar to large language models (LLMs). As privacy and safety concerns grow, these systems increasingly require concept unlearning to remove sensitive or harmful concepts associated with items. However, existing LLM unlearning methods cannot be directly applied to generative recommendation. Unlike word tokens with explicit semantics, SIDs are abstract identifiers that are often shared by both forget and retain items, leading to severe conflicts between concept removal and recommendation utility preservation. To address this challenge, we propose TRACER, an end-to-end concept unlearning framework based on token reassignment. Rather than directly suppressing shared SIDs, TRACER reassigns concept-related items to alternative tokens that better facilitate forgetting while minimizing side effects on retained items. We further introduce a coherence regularizer to preserve semantic consistency among retain items during unlearning. Experiments on real-world recommendation datasets demonstrate that TRACER effectively removes target concepts while substantially better preserving recommendation utility than existing unlearning baselines.

Ziheng Chen, Jiali Cheng, Zezhong Fan +4

Stony Brook University · University of Massachusetts Lowell · Columbia University +3

Sequential RecommendationConcept ErasureMachine UnlearningGenerative Recommendation
AskPDF