Jiaqi Gui

2026

Random Amalgamation of Adapters for Flatter Loss Landscapes: Towards Class-Incremental Learning with Better Stability

AAAI 2026technical

Class-incremental learning (CIL) enables models to continuously learn from streaming data while mitigating catastrophic forgetting of prior knowledge. Our research reveals that the CIL performance of pre-trained models (PTMs) varies significantly across different datasets, a phenomenon underexplored

Cited by 0 Source ↗PDF ↗Scholar ↗

AerialFusion: Co-Motion-Driven Unified Registration and Fusion on Multi-modal Data Streams from Aerial View

Random Amalgamation of Adapters for Flatter Loss Landscapes: Towards Class-Incremental Learning with Better Stability