Yasheng Wang

4 papers Huawei Technologies Ltd.

• Benchmarking Retrieval-Augmented Multimomal Generation for Document Question Answering
• DeepDiver: Adaptive Web-Search Intensity Scaling via Reinforcement Learning
• QFFT, Question-Free Fine-Tuning for Adaptive Reasoning
• RidgeLoRA: Matrix Ridge Enhanced Low-Rank Adaptation of Large Language Models