English
Related papers

Related papers: How to Backdoor the Knowledge Distillation

200 papers

Knowledge distillation (KD) has proved to be an effective approach for deep neural network compression, which learns a compact network (student) by transferring the knowledge from a pre-trained, over-parameterized network (teacher). In…

Machine Learning · Computer Science 2021-04-13 Zi Wang

Post-training of language models (LMs) increasingly relies on the following two stages: (i) knowledge distillation, where the LM is trained to imitate a larger teacher LM, and (ii) reinforcement learning from human feedback (RLHF), where…

Machine Learning · Computer Science 2025-02-06 Daniil Tiapkin , Daniele Calandriello , Johan Ferret , Sarah Perrin , Nino Vieillard , Alexandre Ramé , Mathieu Blondel

Detecting semantic backdoors in classification models--where some classes can be activated by certain natural, but out-of-distribution inputs--is an important problem that has received relatively little attention. Semantic backdoors are…

Machine Learning · Computer Science 2026-01-08 Arpad Berta , Gabor Danner , Istvan Hegedus , Mark Jelasity

Knowledge Distillation (KD) aims to distill the knowledge of a cumbersome teacher model into a lightweight student model. Its success is generally attributed to the privileged information on similarities among categories provided by the…

Computer Vision and Pattern Recognition · Computer Science 2021-03-05 Li Yuan , Francis E. H. Tay , Guilin Li , Tao Wang , Jiashi Feng

The concept of knowledge distillation (KD) describes the training of a student model from a teacher model and is a widely adopted technique in deep learning. However, it is still not clear how and why distillation works. Previous studies…

Machine Learning · Computer Science 2025-10-20 Giulia Lanzillotta , Felix Sarnthein , Gil Kur , Thomas Hofmann , Bobby He

As backdoor attacks become more stealthy and robust, they reveal critical weaknesses in current defense strategies: detection methods often rely on coarse-grained feature statistics, and purification methods typically require full…

Cryptography and Security · Computer Science 2025-08-05 Man Hu , Yahui Ding , Yatao Yang , Liangyu Chen , Yanhao Jia , Shuai Zhao

In this paper, we show that knowledge distillation can be subverted to manipulate language model benchmark scores, revealing a critical vulnerability in current evaluation practices. We introduce "Data Laundering," a process that enables…

Computation and Language · Computer Science 2025-06-05 Jonibek Mansurov , Akhmed Sakip , Alham Fikri Aji

Dataset distillation aims to compress training data into fewer examples via a teacher, from which a student can learn effectively. While its success is often attributed to structure in the data, modern neural networks also memorize specific…

Machine Learning · Computer Science 2026-02-23 Freya Behrens , Lenka Zdeborová

Knowledge distillation has been used to transfer knowledge learned by a sophisticated model (teacher) to a simpler model (student). This technique is widely used to compress model complexity. However, in most applications the compressed…

Machine Learning · Computer Science 2020-11-24 Hadi Pouransari , Mojan Javaheripi , Vinay Sharma , Oncel Tuzel

Knowledge Distillation (KD) methods are capable of transferring the knowledge encoded in a large and complex teacher into a smaller and faster student. Early methods were usually limited to transferring the knowledge only between the last…

Computer Vision and Pattern Recognition · Computer Science 2020-05-05 Nikolaos Passalis , Maria Tzelepi , Anastasios Tefas

The proliferation of diffusion models trained on web-scale, provenance-uncertain image collections has made it essential, yet technically unresolved, to determine whether a model has learned from specific copyrighted data without…

Machine Learning · Computer Science 2026-04-06 Muxing Li , Zesheng Ye , Sharon Li , Andy Song , Guangquan Zhang , Feng Liu

Backdoor attacks are a kind of emergent security threat in deep learning. After being injected with a backdoor, a deep neural model will behave normally on standard inputs but give adversary-specified predictions once the input contains…

Cryptography and Security · Computer Science 2022-10-20 Yangyi Chen , Fanchao Qi , Hongcheng Gao , Zhiyuan Liu , Maosong Sun

Prompt-based approaches offer a cutting-edge solution to data privacy issues in continual learning, particularly in scenarios involving multiple data suppliers where long-term storage of private user data is prohibited. Despite delivering…

Machine Learning · Computer Science 2024-12-18 Trang Nguyen , Anh Tran , Nhat Ho

Knowledge distillation is widely used as a means of improving the performance of a relatively simple student model using the predictions from a complex teacher model. Several works have shown that distillation significantly boosts the…

Machine Learning · Computer Science 2021-07-09 Michal Lukasik , Srinadh Bhojanapalli , Aditya Krishna Menon , Sanjiv Kumar

Policy distillation in deep reinforcement learning provides an effective way to transfer control policies from a larger network to a smaller untrained network without a significant degradation in performance. However, policy distillation is…

Machine Learning · Computer Science 2020-01-01 Yuxiang Sun , Pooyan Fazli

Knowledge distillation extracts general knowledge from a pre-trained teacher network and provides guidance to a target student network. Most studies manually tie intermediate features of the teacher and student, and transfer knowledge…

Machine Learning · Computer Science 2021-02-08 Mingi Ji , Byeongho Heo , Sungrae Park

Backdoor data poisoning is an emerging form of adversarial attack usually against deep neural network image classifiers. The attacker poisons the training set with a relatively small set of images from one (or several) source class(es),…

Machine Learning · Computer Science 2020-10-16 Zhen Xiang , David J. Miller , George Kesidis

Recently, deep neural networks have been shown to be vulnerable to backdoor attacks. A backdoor is inserted into neural networks via this attack paradigm, thus compromising the integrity of the network. As soon as an attacker presents a…

Cryptography and Security · Computer Science 2024-06-18 Zonghao Ying , Bin Wu

The promise of LLM watermarking rests on a core assumption that a specific watermark proves authorship by a specific model. We demonstrate that this assumption is dangerously flawed. We introduce the threat of watermark spoofing, a…

Cryptography and Security · Computer Science 2026-02-24 Hyeseon An , Shinwoo Park , Suyeon Woo , Yo-Sub Han

Backdoor attack is a major threat to deep learning systems in safety-critical scenarios, which aims to trigger misbehavior of neural network models under attacker-controlled conditions. However, most backdoor attacks have to modify the…

Machine Learning · Computer Science 2023-08-24 Yizhen Yuan , Rui Kong , Shenghao Xie , Yuanchun Li , Yunxin Liu
‹ Prev 1 8 9 10 Next ›