Generalizing from Limited Resources in the Open World : Third International Workshop, GLOW 2025, Held in Conjunction with IJCAI 2025, Montreal, Canada, August 16–22, 2025, Proceedings

著者名：Ma, Yuqing (EDT)/Guo, Jinyang (EDT)/Zhao, Xiaowei (EDT)/Gong, Ruihao (EDT)/Liu, Ning (EDT)/Ning, Xuefei (EDT)

Springer（2025/08/14発売）

• 言語：ENG
• ISBN：9789819509874
• eISBN：9789819509881

Description

This book presents the proceedings from the Third International Workshop on Generalizing from Limited Resources in the Open World (GLOW) 2025 held in conjunction with the International Joint Conference on Artificial Intelligence, IJCAI 2025, in Montreal, Canada, during August 16-22, 2025.

The 12 full papers in this book were carefully reviewed and selected from 27 submissions. These papers focus on the academic exploration of efficient methodologies within the realm of artificial intelligence models. We concentrated on both data-efficient strategies, such as zero/few-shot learning and domain adaptation, as well as model-efficient approaches like model sparsification and compact model design.

Table of Contents

.- Evaluating the Behavior of Small Language Models in Answering
Binary Question.

.- Event-Priori-Based Vision-Language Model for Efficient Visual
Understanding.

.- Prompt-Tuning Bandits: Enabling Few-Shot Generalization for Efficient
Multi-Task Offline RL.

.- GateLIP-X:Balancing Adaptation and Generalization in CLIP for
Real-World via a Training-Free Framework.

.- QSE: Mitigating LLM Hallucinations through Query-adaptive
Saliency-localized Activation Editing.

.- Meta-Learning with Heterogeneous Tasks.

.- DIN: Dynamical Interaction Network for Multi-Station Multi-Variable
Weather Prediction.

.- Towards Inclusive NLP: Assessing Compressed Multilingual
Transformers across Diverse Language Benchmarks.

.- Knowledge-Guided Structured Pruning for Multimodal Language Models .

.- Vision Transformers for End-to-End Quark-Gluon Jet Classification
from Calorimeter Images.

.- Special solutions with small volume exist.

.- Adaptive Contextual Embedding for Robust Far-View Borehole
Detection.

.- Class-Aware Sinkhorn-DRO for Few-Shot Domain Adaptation.