↓Skip to main content

Few-Shot Learning

Breaking Long-Tailed Learning Bottlenecks: A Controllable Paradigm with Hypernetwork-Generated Diverse Experts
·2226 words·11 mins· loading · loading
Few-Shot Learning 🏢 University of Science and Technology of China
Controllable long-tailed learning achieved via hypernetwork-generated diverse experts, adapting to user preferences and distribution shifts.
Be Confident in What You Know: Bayesian Parameter Efficient Fine-Tuning of Vision Foundation Models
·3138 words·15 mins· loading · loading
Computer Vision Few-Shot Learning 🏢 Rochester Institute of Technology
Bayesian-PEFT boosts vision model accuracy and confidence in few-shot learning by integrating Bayesian components into PEFT, solving the underconfidence problem.
Attention Temperature Matters in ViT-Based Cross-Domain Few-Shot Learning
·2330 words·11 mins· loading · loading
Computer Vision Few-Shot Learning 🏢 Huazhong University of Science and Technology
Boosting Vision Transformer’s transferability in cross-domain few-shot learning is achieved by a simple yet effective method: strategically adjusting attention temperature to remedy ineffective target…
ARC: A Generalist Graph Anomaly Detector with In-Context Learning
·2432 words·12 mins· loading · loading
Machine Learning Few-Shot Learning 🏢 Griffith University
ARC: a novel generalist graph anomaly detector leveraging in-context learning for efficient, one-for-all anomaly detection across various datasets without retraining.
An Efficient Memory Module for Graph Few-Shot Class-Incremental Learning
·2278 words·11 mins· loading · loading
Machine Learning Few-Shot Learning 🏢 Tsinghua University
Mecoin: a novel memory module for efficient graph few-shot class-incremental learning, tackles catastrophic forgetting by employing structured memory units and a memory representation adaptation modul…
A Closer Look at the CLS Token for Cross-Domain Few-Shot Learning
·3699 words·18 mins· loading · loading
AI Generated Computer Vision Few-Shot Learning 🏢 Huazhong University of Science and Technology
Leaving the CLS token of a Vision Transformer randomly initialized during cross-domain few-shot learning consistently improves performance; a novel method leveraging this phenomenon achieves state-of-…