paper-with-me

Papers

Domain-specific knowledge distillation yields smaller and better models for conversational commerce

2022-05-01 · ECNLP (ACL) 2022 5 · Kristen Howell, Jian Wang, Akshay Hazare, Joseph Bradley, Chris Brew, Xi Chen, Matthew Dunn, Beth Hockey, Andrew Maurer, Dominic Widdows

We demonstrate that knowledge distillation can be used not only to reduce model size, but to simultaneously adapt a contextual language model to a specific domain. We use Multilingual BERT (mBERT; Devlin et al., 2019) as a starting point and follow the knowledge distillation approach of (Sahn et al., 2019) to train a smaller multilingual BERT model that is adapted to the domain at hand. We show that for in-domain tasks, the domain-specific model shows on average 2.3% improvement in F1 score, relative to a model distilled on domain-general data. Whereas much previous work with BERT has fine-tuned the encoder weights during task training, we show that the model improvements from distillation on in-domain data persist even when the encoder weights are frozen during task training, allowing a single encoder to support classifiers for multiple tasks and languages.

📄 PDF Abstract BibTeX

Code (0)

등록된 구현이 없습니다.

Tasks

Knowledge DistillationLanguage ModelingLanguage Modelling

Similar Papers 제목 키워드 기반

Distillation versus Contrastive Learning: How to Train Your Rerankers

2025-07-11 · Zhichao Xu, Zhiqi Huang, Shengyao Zhuang, Vivek Srikumar arxiv

Training effective text rerankers is crucial for information retrieval. Two strategies are widely used: contrastive learning (optimizing directly on ground-truth labels) and knowledge distillation (transferring knowledge…

Knowledge DistillationInformation RetrievalContrastive Learning

Enhancing Generalization in Chain of Thought Reasoning for Smaller Models

2025-01-16 · Maxwell J. Yin, Dingyi Jiang, Yongbing Chen, Boyu Wang 외

Chain-of-Thought (CoT) reasoning in smaller language models is a challenging natural language process problem yet highly desirable in many real-life applications. Existing CoT knowledge distillation methods often suffer …

Knowledge DistillationMemorizationPrompt Engineering

Explicit Knowledge Transfer for Weakly-Supervised Code Generation

2022-11-30 · Zhangir Azerbayev, Ansong Ni, Hailey Schoelkopf, Dragomir Radev

Large language models (LLMs) can acquire strong code-generation capabilities through few-shot learning. In contrast, supervised fine-tuning is still needed for smaller models to achieve good performance. Such fine-tuning…

Code GenerationFew-Shot LearningGSM8KKnowledge Distillation+2

A Study on Knowledge Distillation from Weak Teacher for Scaling Up Pre-trained Language Models

2023-05-26 · Hayeon Lee, Rui Hou, Jongpil Kim, Davis Liang 외

Distillation from Weak Teacher (DWT) is a method of transferring knowledge from a smaller, weaker teacher model to a larger student model to improve its performance. Previous studies have shown that DWT can be effective …

Knowledge Distillation

ERNIE 3.0 Tiny: Frustratingly Simple Method to Improve Task-Agnostic Distillation Generalization

2023-01-09 · Weixin Liu, Xuyi Chen, Jiaxiang Liu, Shikun Feng 외

Task-agnostic knowledge distillation attempts to address the problem of deploying large pretrained language model in resource-constrained scenarios by compressing a large pretrained model called teacher into a smaller on…

Knowledge DistillationLanguage ModellingMulti-Task Learning