MicroPatch: Directed Backdoor Erasing via Victim Parameter Decoupling
Summary
Deep neural networks (large AI models inspired by how brains work) can be attacked through data poisoning, where attackers secretly add harmful examples to training data to make the model behave badly. Existing fixes reduce the attack's success but often make the model worse at normal tasks. Researchers propose MicroPatch, which identifies which parts of the model were corrupted by poisoned data and repairs just those parts by using reverse engineering (reconstructing the hidden attack pattern) and influence functions (mathematical tools that show how each piece of training data affected the final model).
Solution / Mitigation
The source describes MicroPatch as the approach: (1) use reverse engineering to reconstruct backdoor trigger patterns, (2) apply influence functions to quantify the impact of individual data points on model parameters, (3) decouple victim components of model parameters by comparing parameter influences of clean and poisoned data, and (4) patch these victim components to purify the model.
Classification
Related Issues
Original source: http://ieeexplore.ieee.org/document/11570851
First tracked: July 6, 2026 at 08:03 PM
Classified by LLM (prompt v3) · confidence: 92%