Knowledge Distillation (KD) is a common knowledge transfer algorithm used for model compression across a variety of deep learning based natural language processing (NLP) solutions. In its regular manifestations, KD requires access to the teacher's training data for knowledge transfer to the student network. However, privacy concerns, data regulations and proprietary reasons may prevent access to such data. We present, to the best of our knowledge, the first work on Zero-Shot Knowledge Distillation for NLP, where the student learns from the much larger teacher without any task specific data. Our solution combines out of domain data and adversarial training to learn the teacher's output distribution. We investigate six tasks from the GLUE benchmark and demonstrate that we can achieve between 75% and 92% of the teacher's classification score (accuracy or F1) while compressing the model 30 times.
Exploring the Limits of Model Compression in LLMs: A Knowledge Distillation Study on QA Tasks
Fraunhofer Institute for Intelligent Analysis and Information Systems（弗劳恩霍夫智能分析与信息系统研究所）；University of Southern Denmark（南丹麦大学）；Fraunhofer Institute for Applied Information Technology（弗劳恩霍夫应用信息科技研究所）；University Hospital of Cologne（科隆大学医院）
Zhejiang University（浙江大学）；Stevens Institute of Technology（史蒂文斯理工学院）；Alibaba-Zhejiang University Joint Research Institute of Frontier Technologies（阿里巴巴-浙江大学前沿技术联合研究院）
Knowledge distillation has demonstrated encouraging performances in deep model compression. Most existing approaches, however, require massive labeled data to accomplish the knowledge transfer, making the model compression a cumbersome and costly process. In this paper, we investigate the practical few-shot knowledge distillation scenario, where we assume only a few samples without human annotations are available for each category. To this end, we introduce a principled dual-stage distillation scheme tailored for few-shot data. In the first step, we graft the student blocks one by one onto the…
Indian Institute of Science（印度科学学院）；University of Edinburgh（爱丁堡大学）；University of Lincoln（林肯大学）
Knowledge distillation deals with the problem of training a smaller model (Student) from a high capacity source model (Teacher) so as to retain most of its performance. Existing approaches use either the training data or meta-data extracted from it in order to train the Student. However, accessing the dataset on which the Teacher has been trained may not always be feasible if the dataset is very large or it poses privacy or safety concerns (e.g., bio-metric or medical data). Hence, in this paper, we propose a novel data-free method to train the Student from the Teacher. Without even using any …
The Chinese University of Hong Kong（香港中文大学）；Tencent AI Lab（腾讯人工智能实验室）
Model compression has been widely adopted to obtain light-weighted deep neural networks. Most prevalent methods, however, require fine-tuning with sufficient training data to ensure accuracy, which could be challenged by privacy and security issues. As a compromise between privacy and performance, in this paper we investigate few shot network compression: given few samples per class, how can we effectively compress the network with negligible performance drop? The core challenge of few shot network compression lies in high estimation errors from the original network during inference, since the…
Preference-driven Knowledge Distillation for Few-shot Node Classification
College of Electronic and Information Engineering, Tongji University, China（同济大学电子与信息学院）；Shanghai Research Institute for Intelligent Autonomous Systems, Tongji University, China（上海智能自主系统研究院）；National Key Laboratory of Human-Machine Hybrid Augmented Intelligence, Xi’an Jiaotong University, China（人机混合增强智能国家级重点实验室）；School of Computer Science and Technology, Tongji University, China（同济大学计算机科学与技术学院）；Department of Computer Science, University of Illinois Chicago, USA（伊利诺伊大学芝加哥分校计算机科学系）
iFLYTEK Research（科大讯飞研究院）；Harbin Institute of Technology（哈尔滨工业大学）；iFLYTEK AI Research (Hebei)（科大讯飞人工智能研究院（河北））
In this paper, we introduce TextBrewer, an open-source knowledge distillation toolkit designed for natural language processing. It works with different neural network models and supports various kinds of supervised learning tasks, such as text classification, reading comprehension, sequence labeling. TextBrewer provides a simple and uniform workflow that enables quick setting up of distillation experiments with highly flexible configurations. It offers a set of predefined distillation methods and can be extended with custom code. As a case study, we use TextBrewer to distill BERT on several ty…
Applied Artificial Intelligence Institute (AI), Deakin University（迪肯大学应用人工智能研究所）
Knowledge distillation (KD) is an efficient approach to transfer the knowledge from a large "teacher" network to a smaller "student" network. Traditional KD methods require lots of labeled training samples and a white-box teacher (parameters are accessible) to train a good student. However, these resources are not always available in real-world applications. The distillation process often happens at an external party side where we do not have access to much data, and the teacher does not disclose its parameters due to security and privacy concerns. To overcome these challenges, we propose a bl…
Since deep learning became a key player in natural language processing (NLP), many deep learning models have been showing remarkable performances in a variety of NLP tasks, and in some cases, they are even outperforming humans. Such high performance can be explained by efficient knowledge representation of deep learning models. While many methods have been proposed to learn more efficient representation, knowledge distillation from pretrained deep networks suggest that we can use more information from the soft target probability to train other neural networks. In this paper, we propose a new k…
On Distilling the Displacement Knowledge for Few-Shot Class-Incremental Learning
School of Computer Science and Engineering, Southeast University（东南大学计算机科学与工程学院）；Key Laboratory of New Generation Artificial Intelligence Technology and Its Interdisciplinary Applications (Southeast University), Ministry of Education（教育部新一代人工智能技术及交叉应用重点实验室（东南大学））
Knowledge distillation (KD) is a successful approach for deep neural network acceleration, with which a compact network (student) is trained by mimicking the softmax output of a pre-trained high-capacity network (teacher). In tradition, KD usually relies on access to the training samples and the parameters of the white-box teacher to acquire the transferred knowledge. However, these prerequisites are not always realistic due to storage costs or privacy issues in real-world applications. Here we propose the concept of decision-based black-box (DB3) knowledge distillation, with which the student…
Qualcomm AI Research（高通人工智能研究院）；University of Amsterdam（阿姆斯特丹大学）
This paper aims to accelerate video stream processing, such as object detection and semantic segmentation, by leveraging the temporal redundancies that exist between video frames. Instead of propagating and warping features using motion alignment, such as optical flow, we propose a novel knowledge distillation schema coined as Delta Distillation. In our proposal, the student learns the variations in the teacher's intermediate features over time. We demonstrate that these temporal variations can be effectively distilled due to the temporal redundancies within video frames. During inference, bot…
Hong Kong University of Science and Technology（香港科技大学）；Huawei Noah’s Ark Lab（华为诺亚方舟实验室）
The recent large-scale vision-language pre-training (VLP) of dual-stream architectures (e.g., CLIP) with a tremendous amount of image-text pair data, has shown its superiority on various multimodal alignment tasks. Despite its success, the resulting models are not capable of multimodal generative tasks due to the weak text encoder. To tackle this problem, we propose to augment the dual-stream VLP model with a textual pre-trained language model (PLM) via vision-language knowledge distillation (VLKD), enabling the capability for multimodal generation. VLKD is pretty data- and computation-efficie…
Department of Electrical and Electronic Engineering（电子与电气工程系）；Imperial College London（帝国理工学院）；Imperial Centre for Quantum Engineering, Science and Technology（帝国量子工程、科学与技术中心）；Blackett Laboratory, Department of Physics（布莱克特实验室，物理系）；Graduate Institute of Applied Physics（应用物理研究院）；Brookhaven National Laboratory（布鲁赫斯国家实验室）；Quantum Information Center（量子信息中心）；Master Program in Intelligent Computing and Big Data（智能计算与大数据硕士项目）；Chung Yuan Christian University；JIJ
Department of Electrical Engineering, POSTECH（POSTECH电气工程系）；Graduate School of Artificial Intelligence, POSTECH（POSTECH人工智能研究生院）；POSTECH（POSTECH（浦项科技大学））
Deep models are susceptible to learning spurious correlations, even during the post-processing. We take a closer look at the knowledge distillation -- a popular post-processing technique for model compression -- and find that distilling with biased training data gives rise to a biased student, even when the teacher is debiased. To address this issue, we propose a simple knowledge distillation algorithm, coined DeTT (Debiasing by Teacher Transplanting). Inspired by a recent observation that the last neural net layer plays an overwhelmingly important role in debiasing, DeTT directly transplants …
Aptiv Technical Center（安波福技术中心）；Aptiv Advance Research Center（安波福高级研究中心）
Many radar signal processing methodologies are being developed for critical road safety perception tasks. Unfortunately, these signal processing algorithms are often poorly suited to run on embedded hardware accelerators used in automobiles. Conversely, end-to-end machine learning (ML) approaches better exploit the performance gains brought by specialized accelerators. In this paper, we propose a teacher-student knowledge distillation approach for low-level radar perception tasks. We utilize a hybrid model for stationary object detection as a teacher to train an end-to-end ML student model. Th…
Given the success with in-context learning of large pre-trained language models, we introduce in-context learning distillation to transfer in-context few-shot learning ability from large models to smaller models. We propose to combine in-context learning objectives with language modeling objectives to distill both the ability to read in-context examples and task knowledge to the smaller models. We perform in-context learning distillation under two different few-shot learning paradigms: Meta In-context Tuning (Meta-ICT) and Multitask In-context Tuning (Multitask-ICT). Multitask-ICT performs bet…
Summarize-Exemplify-Reflect: Data-driven Insight Distillation Empowers LLMs for Few-shot Tabular Classification
ETH Zürich（苏黎世联邦理工学院）；University of Copenhagen（哥本哈根大学）；University of Science and Technology of China（中国科学技术大学）；University of Amsterdam（阿姆斯特丹大学）；East China Normal University（东华大学）
iFLYTEK Research（科大讯飞研究院）；Harbin Institute of Technology（哈尔滨工业大学）
Multilingual pre-trained language models (MPLMs) not only can handle tasks in different languages but also exhibit surprising zero-shot cross-lingual transferability. However, MPLMs usually are not able to achieve comparable supervised performance on rich-resource languages compared to the state-of-the-art monolingual pre-trained models. In this paper, we aim to improve the multilingual model's supervised and zero-shot performance simultaneously only with the resources from supervised languages. Our approach is based on transferring knowledge from high-performance monolingual models with a tea…
Institute of Computing Technology, Chinese Academy of Sciences（中国科学院计算技术研究所）；University of Chinese Academy of Sciences（中国科学院大学）
Stance detection aims to identify whether the author of a text is in favor of, against, or neutral to a given target. The main challenge of this task comes two-fold: few-shot learning resulting from the varying targets and the lack of contextual information of the targets. Existing works mainly focus on solving the second issue by designing attention-based models or introducing noisy external knowledge, while the first issue remains under-explored. In this paper, inspired by the potential capability of pre-trained language models (PLMs) serving as knowledge bases and few-shot learners, we prop…
Adversarial examples are artificially modified input samples which lead to misclassifications, while not being detectable by humans. These adversarial examples are a challenge for many tasks such as image and text classification, especially as research shows that many adversarial examples are transferable between different classifiers. In this work, we evaluate the performance of a popular defensive strategy for adversarial examples called defensive distillation, which can be successful in hardening neural networks against adversarial examples in the image domain. However, instead of applying …
Recent work has shown that small distilled language models are strong competitors to models that are orders of magnitude larger and slower in a wide range of information retrieval tasks. This has made distilled and dense models, due to latency constraints, the go-to choice for deployment in real-world retrieval applications. In this work, we question this practice by showing that the number of parameters and early query-document interaction play a significant role in the generalization ability of retrieval models. Our experiments show that increasing model size results in marginal gains on in-…
CycleDistill: Bootstrapping Machine Translation using LLMs with Cyclical Distillation
Nilekani Centre at AI4Bharat（AI4Bharat的Nilekani中心）；Indian Institute of Technology（印度理工学院）；Indian Institute of Technology Bombay（印度理工学院孟买分校）；Indian Institute of Engineering Science and Technology Shibpur（Shibpur的印度工程科学与技术学院）
Generative models are increasingly able to produce remarkably high quality images and text. The community has developed numerous evaluation metrics for comparing generative models. However, these metrics do not effectively quantify data diversity. We develop a new diversity metric that can readily be applied to data, both synthetic and natural, of any type. Our method employs random network distillation, a technique introduced in reinforcement learning. We validate and deploy this metric on both images and text. We further explore diversity in few-shot image generation, a setting which was pre…
Columbia University（哥伦比亚大学）；Microsoft Research（微软研究院）
Contrastive language-image pretraining (CLIP) links vision and language modalities into a unified embedding space, yielding the tremendous potential for vision-language (VL) tasks. While early concurrent works have begun to study this potential on a subset of tasks, important questions remain: 1) What is the benefit of CLIP on unstudied VL tasks? 2) Does CLIP provide benefit in low-shot or domain-shifted scenarios? 3) Can CLIP improve existing approaches without impacting inference or pretraining complexity? In this work, we seek to answer these questions through two key contributions. First, …
Northeastern University（东北大学）；Amazon Web Services（亚马逊网络服务）
We aim to identify how different components in the KD pipeline affect the resulting performance and how much the optimal KD pipeline varies across different datasets/tasks, such as the data augmentation policy, the loss function, and the intermediate representation for transferring the knowledge between teacher and student. To tease apart their effects, we propose Distiller, a meta KD framework that systematically combines a broad range of techniques across different stages of the KD pipeline, which enables us to quantify each component's contribution. Within Distiller, we unify commonly used …
On-Policy Distillation for Vision-Language Model Adaptation, an Effective Paradigm on Low-Quality Multimodal Data
The University of Hong Kong（香港大学）；Great Wall Motor（长城汽车）；Wuhan University（武汉大学）；China University of Petroleum (East China)（中国石油大学（华东））；University of Science and Technology of China（中国科学技术大学）；Northwestern Polytechnical University（西北工业大学）
University of Science and Technology of China（中国科学技术大学）；iFLYTEK Research（讯飞研究院）；Jilin Kexun Information Technology Co., Ltd.（吉林科讯信息技术有限公司）
Zero-shot cross-lingual named entity recognition (NER) aims at transferring knowledge from annotated and rich-resource data in source languages to unlabeled and lean-resource data in target languages. Existing mainstream methods based on the teacher-student distillation framework ignore the rich and complementary information lying in the intermediate layers of pre-trained language models, and domain-invariant information is easily lost during transfer. In this study, a mixture of short-channel distillers (MSD) method is proposed to fully interact the rich hierarchical information in the teache…
