paper-with-me

홈 › Papers

Bi-Drop: Enhancing Fine-tuning Generalization via Synchronous sub-net Estimation and Optimization

2023-05-24 · Shoujie Tong, Heming Xia, Damai Dai, Runxin Xu, Tianyu Liu, Binghuai Lin, Yunbo Cao, Zhifang Sui

Pretrained language models have achieved remarkable success in natural language understanding. However, fine-tuning pretrained models on limited training data tends to overfit and thus diminish performance. This paper presents Bi-Drop, a fine-tuning strategy that selectively updates model parameters using gradients from various sub-nets dynamically generated by dropout. The sub-net estimation of Bi-Drop is performed in an in-batch manner, so it overcomes the problem of hysteresis in sub-net updating, which is possessed by previous methods that perform asynchronous sub-net estimation. Also, Bi-Drop needs only one mini-batch to estimate the sub-net so it achieves higher utility of training data. Experiments on the GLUE benchmark demonstrate that Bi-Drop consistently outperforms previous fine-tuning methods. Furthermore, empirical results also show that Bi-Drop exhibits excellent generalization ability and robustness for domain transfer, data imbalance, and low-resource scenarios.

📄 PDF Abstract BibTeX arXiv:2305.14760

Code (0)

등록된 구현이 없습니다.

Tasks

Natural Language Understanding

Methods 이 논문이 사용한 방법론

Dropout Dropout is a regularization technique for neural networks that drops a unit (along with connections) at training time with a specified probability $p$ (a common value is…

Similar Papers 제목 키워드 기반

FedLoDrop: Federated LoRA with Dropout for Generalized LLM Fine-tuning

2025-10-14 · Sijing Xie, Dingzhu Wen, Changsheng You, Qimei Chen 외 arxiv

Fine-tuning (FT) large language models (LLMs) is crucial for adapting general-purpose models to specific tasks, enhancing accuracy and relevance with minimal resources. To further enhance generalization ability while red…

EventDrop: data augmentation for event-based learning

2021-06-07 · Fuqiang Gu, Weicong Sng, Xuke Hu, Fangwen Yu

The advantages of event-sensing over conventional sensors (e.g., higher dynamic range, lower time latency, and lower power consumption) have spurred research into machine learning for event data. Unsurprisingly, deep lea…

Data AugmentationDiversity

Why Multi-Step Tool-Use Reinforcement Learning Collapses and How Supervisory Signals Fix It

2026-06-24 · Yupu Hao, Zhuoran Jin, Huanxuan Liao, Kang Liu 외 arxiv

Tool use enables large language models (LLMs) to perform complex tasks, and recent agentic reinforcement learning (RL) methods show promise for enhancing model capabilities. However, RL alone often leads to instability o…

Reinforcement Learning

Selective Self-to-Supervised Fine-Tuning for Generalization in Large Language Models

2025-02-12 · Sonam Gupta, Yatin Nandwani, Asaf Yehudai, Dinesh Khandelwal 외

Fine-tuning Large Language Models (LLMs) on specific datasets is a common practice to improve performance on target tasks. However, this performance gain often leads to overfitting, where the model becomes too specialize…

Mathematical ReasoningMMLUReading ComprehensionTruthfulQA

Asynchronous Sharpness-Aware Minimization For Fast and Accurate Deep Learning

2025-03-14 · Junhyuk Jo, Jihyun Lim, Sunwoo Lee

Sharpness-Aware Minimization (SAM) is an optimization method that improves generalization performance of machine learning models. Despite its superior generalization, SAM has not been actively used in real-world applicat…

Deep Learning