Baidu AI Security Research
Baidu Security focuses on the challenges faced by AI systems and applications in reality, discloses security risks, resists multi-dimensional attacks, and prevents AI abuse. Baidu Security takes the lead in AI security research, with research achievements covering international academic conferences and releasing AI The most influential open source tool in terms of confrontation, and provide protection scheme for AI core assets
Scope of study
AI model security
AI data security
AI capability abuse
AI hardware system security
AI software realizes security
AI model security
Model robustness
Model Robustness
Each AI model has a range of actual application scenarios. Out of range,
The model does not ensure security. Baidu Security Releases AI Model
Robust platform.
View details
AI model security
Real World Counterattack
Physical World
Adversarial Attack
View details
AI model security
Open source toolkit for countering attacks
Top 3 ranking of confrontation tools on Github supports image, voice and text
Generation and testing of multiple countermeasures
View details
AI model security
AI Hardware Security
Model Anti theft
Model Anti-Theft
AI model is the key intellectual property of AI services, and Baidu Security provides
For hardware based model anti stealing scheme.
View details
AI software realizes security
Automatic vulnerability mining tool
Baidu DeepExplorer
Baidu Security is in the process of automatic program analysis, fuzzing, and formal test
Research results have been published on all sides.
View details
AI data security
Secure computing platform for privacy protection
Privacy-protected ML
Break the data island, let the data carry out machine learning operation under the premise of protecting privacy,
Maximize value.
View details
AI capability abuse
Deep forgery detection
Deep forgery makes it easy to change face, voice and mouth shape
It is difficult to distinguish with the naked eye.
Baidu DeepFake detection provides intelligent identification of deep forgery
View details
logo Business consulting: 400-805-4999
Safety products
DDoS attack protection
Solution
IDC Smart Yundun Cloud advanced defense xSRC gSRC
Security Services
Penetration test Safety training
Open service
Website security detection SMS SMS content security
other
Safe Community Privacy Agreement
Follow us
bar-code

© 2024  Baidu Must read before using Baidu Feedback    Jing ICP Certificate No. 030173    Jinggong Network Anbei No. 1100000200001

Link
system
I
We

* Identity type
️ enterprise
️ personal
* Enterprise name
Enterprise name cannot be empty
* Real name
The real name cannot be empty
* phone number
The phone number cannot be empty
* mailbox
The mailbox cannot be empty
* Application service
* Consultation content
The consultation content cannot be empty
* Verification Code
The verification code cannot be empty
Submit
Submitted successfully
Your application has been submitted, and our business team will contact you later. Thank you!
close
￥
zero
cash coupon
zero
Exchange Certificate
Collect immediately
Receive successfully