Picture for Haixu Tang

Haixu Tang

DPAdapter: Improving Differentially Private Deep Learning through Noise Tolerance Pre-training

Add code
Mar 05, 2024
Viaarxiv icon

The Janus Interface: How Fine-Tuning in Large Language Models Amplifies the Privacy Risks

Add code
Oct 24, 2023
Figure 1 for The Janus Interface: How Fine-Tuning in Large Language Models Amplifies the Privacy Risks
Figure 2 for The Janus Interface: How Fine-Tuning in Large Language Models Amplifies the Privacy Risks
Figure 3 for The Janus Interface: How Fine-Tuning in Large Language Models Amplifies the Privacy Risks
Figure 4 for The Janus Interface: How Fine-Tuning in Large Language Models Amplifies the Privacy Risks
Viaarxiv icon

Large Language Model Soft Ideologization via AI-Self-Consciousness

Add code
Sep 28, 2023
Viaarxiv icon

Gradient Shaping: Enhancing Backdoor Attack Against Reverse Engineering

Add code
Jan 29, 2023
Figure 1 for Gradient Shaping: Enhancing Backdoor Attack Against Reverse Engineering
Figure 2 for Gradient Shaping: Enhancing Backdoor Attack Against Reverse Engineering
Figure 3 for Gradient Shaping: Enhancing Backdoor Attack Against Reverse Engineering
Figure 4 for Gradient Shaping: Enhancing Backdoor Attack Against Reverse Engineering
Viaarxiv icon

Selective Amnesia: On Efficient, High-Fidelity and Blind Suppression of Backdoor Effects in Trojaned Machine Learning Models

Add code
Dec 09, 2022
Viaarxiv icon

Understanding Impacts of Task Similarity on Backdoor Attack and Detection

Add code
Oct 12, 2022
Figure 1 for Understanding Impacts of Task Similarity on Backdoor Attack and Detection
Figure 2 for Understanding Impacts of Task Similarity on Backdoor Attack and Detection
Figure 3 for Understanding Impacts of Task Similarity on Backdoor Attack and Detection
Figure 4 for Understanding Impacts of Task Similarity on Backdoor Attack and Detection
Viaarxiv icon

Bounding The Number of Linear Regions in Local Area for Neural Networks with ReLU Activations

Add code
Jul 14, 2020
Figure 1 for Bounding The Number of Linear Regions in Local Area for Neural Networks with ReLU Activations
Figure 2 for Bounding The Number of Linear Regions in Local Area for Neural Networks with ReLU Activations
Figure 3 for Bounding The Number of Linear Regions in Local Area for Neural Networks with ReLU Activations
Viaarxiv icon

Towards Fair Cross-Domain Adaptation via Generative Learning

Add code
Mar 04, 2020
Figure 1 for Towards Fair Cross-Domain Adaptation via Generative Learning
Figure 2 for Towards Fair Cross-Domain Adaptation via Generative Learning
Figure 3 for Towards Fair Cross-Domain Adaptation via Generative Learning
Figure 4 for Towards Fair Cross-Domain Adaptation via Generative Learning
Viaarxiv icon

Understanding Membership Inferences on Well-Generalized Learning Models

Add code
Feb 13, 2018
Figure 1 for Understanding Membership Inferences on Well-Generalized Learning Models
Figure 2 for Understanding Membership Inferences on Well-Generalized Learning Models
Figure 3 for Understanding Membership Inferences on Well-Generalized Learning Models
Figure 4 for Understanding Membership Inferences on Well-Generalized Learning Models
Viaarxiv icon