• Home
  • Search
  • Backdoor in deep learning: new threats and opportunities
  • https://doi.org/10.32657/10356/182221Copy DOI Icon

Backdoor in deep learning: new threats and opportunities

  • Jan 1, 2025
  • Kangjie Chen
Show More
  • Abstract
  • Literature Map
  • Similar Papers
Abstract

Deep learning has become increasingly popular due to its remarkable ability to learn high-dimensional feature representations. Numerous algorithms and models have been developed to enhance the application of deep learning across various real-world tasks, including image classification, natural language processing, and autonomous driving. However, deep learning models are susceptible to backdoor threats, where an attacker manipulates the training process or data to cause incorrect predictions on malicious samples containing specific triggers, while maintaining normal performance on benign samples. With the advancement of deep learning, including evolving training schemes and the need for large-scale training data, new threats in the backdoor domain continue to emerge. Conversely, backdoors can also be leveraged to protect deep learning models, such as through watermarking techniques. In this thesis, we conduct an in-depth investigation into backdoor techniques from three novel perspectives. In the first part of this thesis, we demonstrate that emerging deep learning training schemes can introduce new backdoor risks. Specifically, pre-trained Natural Language Processing (NLP) models can be easily adapted to a variety of downstream language tasks, significantly accelerating the development of language models. However, the pre-trained model becomes a single point of failure for these downstream models. We propose a novel task-agnostic backdoor attack against pre-trained NLP models, wherein the adversary does not need prior information about the downstream tasks when implanting the backdoor into the pre-trained model. Any downstream models transferred from this malicious model will inherit the backdoor, even after extensive transfer learning, revealing the severe vulnerability of pre-trained foundation models to backdoor attacks. In the second part of this thesis, we develop novel backdoor attack methods suited to new threat scenarios. The rapid expansion of deep learning models necessitates large-scale training data, much of which is unlabeled and outsourced to third parties for annotation. To ensure data security, most datasets are read-only for training samples, preventing the addition of input triggers. Consequently, attackers can only achieve data poisoning by uploading malicious annotations. In this practical scenario, all existing data poisoning methods that add triggers to the input are infeasible. Therefore, we propose new backdoor attack methods that involve poisoning only the labels without modifying any input samples. In the third part of this thesis, we utilize the backdoor technique to proactively protect our deep learning models, specifically for intellectual property protection. Considering the complexity of deep learning tasks, generating a well-trained deep learning model requires substantial computational resources, training data, and expertise. Therefore, it is essential to protect these assets and prevent copyright infringement. Inspired by backdoor attacks that can induce specific behaviors in target models through carefully designed samples, several watermarking methods have been proposed to protect the intellectual property of deep learning models. Model owners can train their models to produce unique outputs for certain crafted samples and use these samples for ownership verification. While various extraction techniques have been designed for supervised deep learning models, challenges arise when applying them to deep reinforcement learning models due to differences in model features and scenarios. Therefore, we propose a novel watermarking scheme to protect deep reinforcement learning models from unauthorized distribution. Instead of using spatial watermarks as in conventional deep learning models, we design temporal watermarks that minimize potential impact and damage to the protected deep reinforcement learning model while achieving high-fidelity ownership verification. In summary, this thesis investigates the evolving landscape of backdoor threats during the development of deep learning techniques and the use of backdoors for beneficial purposes in intellectual property protection.

Similar Papers
  • Research Article

Abstract 184: The utility of deep metric learning for breast cancer identification on mammographic images

  • Jul 01, 2021
  • Cancer Research
  • Justin Du +8
  • Conference Article
  • Citations9

Temporal Watermarks for Deep Reinforcement Learning Models

  • May 03, 2021
  • Kangjie Chen +4
  • Dissertation
  • Citations3

Automated stock trading system using deep reinforcement learning and price and sentiment prediction modules.

  • Jun 15, 2021
  • Roberto Fray Da Silva
  • Dissertation

Temporal analysis of healthcare trajectories using deep learning

  • Jan 01, 2023
  • Riazat Ryan
  • Research Article
  • Citations105

Pretrained domain-specific language model for natural language processing tasks in the AEC domain

  • Jun 21, 2022
  • Computers in Industry
  • Zhe Zheng +4
  • Book Chapter

Watermarks for Deep Reinforcement Learning

  • Nov 28, 2022
  • Kangjie Chen
  • Research Article

P–260 Towards better explainable deep learning models for embryo selection in ART

  • Aug 06, 2021
  • Human Reproduction
  • Ashu Sharma +4
  • Research Article
  • Citations1

Deep multi-instance learning model based on gadoxetic acid-enhanced MRI for predicting microvascular invasion of hepatocellular carcinoma: a multicenter, retrospective study

  • Oct 22, 2025
  • BMC Cancer
  • Yi Luo +7
  • Research Article
  • Citations1

Detecting Empathy in Roman Urdu Using Transfer Learning

  • Feb 10, 2025
  • ACM Transactions on Asian and Low-Resource Language Information Processing
  • Hafsa Sattar +3
  • Research Article
  • Citations9

Evaluating deep transfer learning for whole-brain cognitive decoding

  • Jul 13, 2023
  • Journal of the Franklin Institute
  • Armin W Thomas +3
  • Conference Article
  • Citations1

MobileDLSearch: Ontology-based Mobile Platform for Effective Sharing and Reuse of Deep Learning Models

  • Dec 01, 2021
  • Zhangcheng Qiang +5
  • PDF
  • Research Article
  • Citations23

Explainable artificial intelligence (XAI) for predicting the need for intubation in methanol-poisoned patients: a study comparing deep and machine learning models

  • Jul 08, 2024
  • Scientific Reports
  • Khadijeh Moulaei +14
  • PDF
  • Research Article
  • Citations107

Intrusion Detection System for Industrial Internet of Things Based on Deep Reinforcement Learning

  • Jan 01, 2022
  • Wireless Communications and Mobile Computing
  • Sumegh Tharewal +5
  • Research Article
  • Citations3

Weighted Suspiciousness and Balanced Aggregation to Boost Spectrum-based Fault Localization of Deep Learning Models

  • Feb 12, 2025
  • ACM Transactions on Software Engineering and Methodology
  • Wenjie Xu +4
  • Research Article

Increasing trust in complex machine learning systems

  • Jun 01, 2021
  • ACM SIGIR Forum
  • Jaehun Kim
Cactus Communications logo

Copyright 2026 Cactus Communications. All rights reserved.