This paper studies compressing pre-trained language models, like BERT (Devlin et al.,2019), via teacher-student knowledge distillation. Previous works usually force the student model to strictly mimic the smoothed labels predicted by the teacher BERT. As an alternative, we propose a new method for BERT distillation, i.e., asking the teacher to generate smoothed word ids, rather than labels, for teaching the student model in knowledge distillation. We call this kind of methodTextSmoothing. Practically, we use the softmax prediction of the Masked Language Model(MLM) in BERT to generate word distributions for given texts and smooth those input texts using that predicted soft word ids. We assume that both the smoothed labels and the smoothed texts can implicitly augment the input corpus, while text smoothing is intuitively more efficient since it can generate more instances in one neural network forward step.Experimental results on GLUE and SQuAD demonstrate that our solution can achieve competitive results compared with existing BERT distillation methods.
Multilingual speech data often suffer from long-tailed language distribution, resulting in performance degradation. However, multilingual text data is much easier to obtain, yielding a more useful general language model. Hence, we are motivated to distill the rich knowledge embedded inside a well-trained teacher text model to the student speech model. We propose a novel method called the Distilling a Language model to a Speech model (Distill-L2S), which aligns the latent representations of two different modalities. The subtle differences are handled by the shrinking mechanism, nearest-neighbor…
While pretrained language models (PLMs) have greatly improved text generation, they have also been known to produce unfaithful or inappropriate content. In contrast, classic template-based systems provide strong guarantees of faithfulness at the cost of fluency. We propose TempLM, which achieves the best of both worlds by distilling a PLM into a template-based generator. On the E2E and SynthBio data-to-text datasets, we show that TempLM is more faithful than the original PLM and is more fluent than prior template systems. Notably, on an out-of-domain evaluation, TempLM reduces a finetuned BART…
Mind's Mirror: Distilling Self-Evaluation Capability and Comprehensive Thinking from Large Language Models
Zhejiang University（浙江大学）；School of Public Health, Zhejiang University（浙江大学公共卫生学院）；The Hong Kong University of Science and Technology (Guangzhou)（香港科技大学（广州））；Liangzhu Laboratory and Institute of Wenzhou, Zhejiang University（浙江大学良渚实验室与温州研究院）；University of Illinois Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）
Unveiling the Key Factors for Distilling Chain-of-Thought Reasoning
Department of Computing, The Hong Kong Polytechnic University（香港理工大学计算机系）；Ningbo Key Laboratory of Spatial Intelligence and Digital Derivative（宁波空间智能与数字衍生关键实验室）；Institute of Digital Twin, EIT（数字孪生研究院）；Saarland University（萨尔兰大学）；Meituan Inc.（美团公司）
Peking University（北京大学）；Tencent Research（腾讯研究院）；Beijing Normal University（北京师范大学）
Pre-trained language models like BERT have proven to be highly performant. However, they are often computationally expensive in many practical scenarios, for such heavy models can hardly be readily implemented with limited resources. To improve their efficiency with an assured model performance, we propose a novel speed-tunable FastBERT with adaptive inference time. The speed at inference can be flexibly adjusted under varying demands, while redundant calculation of samples is avoided. Moreover, this model adopts a unique self-distillation mechanism at fine-tuning, further enabling a greater c…
Large pre-trained multilingual models like mBERT, XLM-R achieve state of the art results on language understanding tasks. However, they are not well suited for latency critical applications on both servers and edge devices. It's important to reduce the memory and compute resources required by these models. To this end, we propose pQRNN, a projection-based embedding-free neural encoder that is tiny and effective for natural language processing tasks. Without pre-training, pQRNNs significantly outperform LSTM models with pre-trained embeddings despite being 140x smaller. With the same number of …
Pre-trained language models have been found to capture a surprisingly rich amount of lexical knowledge, ranging from commonsense properties of everyday concepts to detailed factual knowledge about named entities. Among others, this makes it possible to distill high-quality word vectors from pre-trained language models. However, it is currently unclear to what extent it is possible to distill relation embeddings, i.e. vectors that characterize the relationship between two words. Such relation embeddings are appealing because they can, in principle, encode relational knowledge in a more fine-gra…
DRAG: Distilling RAG for SLMs from LLMs to Transfer Knowledge and Mitigate Hallucination via Evidence and Graph-based Distillation
VILA Lab（VILA实验室）；Mohamed bin Zayed University of AI（莫扎伊德·本·泽德人工智能大学）；McGill University（麦吉尔大学）；National University of Science and Technology（国家科学与技术大学）
Pre-trained language models achieve outstanding performance in NLP tasks. Various knowledge distillation methods have been proposed to reduce the heavy computation and storage requirements of pre-trained language models. However, from our observations, student models acquired by knowledge distillation suffer from adversarial attacks, which limits their usage in security sensitive scenarios. In order to overcome these security problems, RoSearch is proposed as a comprehensive framework to search the student models with better adversarial robustness when performing knowledge distillation. A dire…
Yonsei University（延世大学）；Seoul National University（首尔国立大学）
Deep learning for Information Retrieval (IR) requires a large amount of high-quality query-document relevance labels, but such labels are inherently sparse. Label smoothing redistributes some observed probability mass over unobserved instances, often uniformly, uninformed of the true distribution. In contrast, we propose knowledge distillation for informed labeling, without incurring high computation overheads at evaluation time. Our contribution is designing a simple but efficient teacher model which utilizes collective knowledge, to outperform state-of-the-arts distilled from a more complex …
Graduate School of Informatics, Kyoto University（京都大学信息学研究生院）
Connectionist temporal classification (CTC) -based models are attractive because of their fast inference in automatic speech recognition (ASR). Language model (LM) integration approaches such as shallow fusion and rescoring can improve the recognition accuracy of CTC-based ASR by taking advantage of the knowledge in text corpora. However, they significantly slow down the inference of CTC. In this study, we propose to distill the knowledge of BERT for CTC-based ASR, extending our previous study for attention-based ASR. CTC-based ASR learns the knowledge of BERT during training and does not use …
National Taiwan University（台湾大学）；Academia Sinica（中央研究院）
Transformer-based models are widely used in natural language understanding (NLU) tasks, and multimodal transformers have been effective in visual-language tasks. This study explores distilling visual information from pretrained multimodal transformers to pretrained language encoders. Our framework is inspired by cross-modal encoders' success in visual-language tasks while we alter the learning objective to cater to the language-heavy characteristics of NLU. After training with a small number of extra adapting steps and finetuned, the proposed XDBERT (cross-modal distilled BERT) outperforms pre…
Microsoft Dynamics 365 AI Research（微软Dynamics 365人工智能研究院）；Carnegie Mellon University（卡内基梅隆大学）
Large-scale pre-trained language model such as BERT has achieved great success in language understanding tasks. However, it remains an open question how to utilize BERT for language generation. In this paper, we present a novel approach, Conditional Masked Language Modeling (C-MLM), to enable the finetuning of BERT on target generation tasks. The finetuned BERT (teacher) is exploited as extra supervision to improve conventional Seq2Seq models (student) for better text generation performance. By leveraging BERT's idiosyncratic bidirectional nature, distilling knowledge learned in BERT can encou…
Vision-language models (VLMs) exhibit uneven performance across languages, a problem that is often exacerbated when the model size is reduced. While Knowledge distillation (KD) demonstrates promising results in transferring knowledge from larger to smaller VLMs, applying KD in multilingualism is an underexplored area. This paper presents a controlled empirical study of KD behavior across five distillation approaches, isolating their effects on cross-lingual representation consistency and downstream performance stability under model compression. We study five distillation formulations across CL…
We propose a simple yet effective method to compress an RNN-Transducer (RNN-T) through the well-known knowledge distillation paradigm. We show that the transducer's encoder outputs naturally have a high entropy and contain rich information about acoustically similar word-piece confusions. This rich information is suppressed when combined with the lower entropy decoder outputs to produce the joint network logits. Consequently, we introduce an auxiliary loss to distill the encoder logits from a teacher transducer's encoder, and explore training strategies where this encoder distillation works ef…
We present an approach to ranking with dense representations that applies knowledge distillation to improve the recently proposed late-interaction ColBERT model. Specifically, we distill the knowledge from ColBERT's expressive MaxSim operator for computing relevance scores into a simple dot product, thus enabling single-step ANN search. Our key insight is that during distillation, tight coupling between the teacher model and the student model enables more flexible distillation strategies and yields better learned representations. We empirically show that our approach improves query latency and…
Research Center for Social Computing and Information Retrieval（社会计算与信息检索研究中心）；Harbin Institute of Technology, China（哈尔滨工业大学（中国））
Many natural language processing tasks can be modeled into structured prediction and solved as a search problem. In this paper, we distill an ensemble of multiple models trained with different initialization into a single model. In addition to learning to match the ensemble's probability output on the reference states, we also use the ensemble to explore the search space and learn from the encountered states in the exploration. Experimental results on two typical search-based structured prediction tasks -- transition-based dependency parsing and neural machine translation show that distillatio…
In the natural language processing literature, neural networks are becoming increasingly deeper and complex. The recent poster child of this trend is the deep language representation model, which includes BERT, ELMo, and GPT. These developments have led to the conviction that previous-generation, shallower neural networks for language understanding are obsolete. In this paper, however, we demonstrate that rudimentary, lightweight neural networks can still be made competitive without architecture changes, external training data, or additional input features. We propose to distill knowledge from…
In this paper, we investigate improvements to the GEC sequence tagging architecture with a focus on ensembling of recent cutting-edge Transformer-based encoders in Large configurations. We encourage ensembling models by majority votes on span-level edits because this approach is tolerant to the model architecture and vocabulary size. Our best ensemble achieves a new SOTA result with an $F_{0.5}$ score of 76.05 on BEA-2019 (test), even without pre-training on synthetic datasets. In addition, we perform knowledge distillation with a trained ensemble to generate new synthetic training datasets, "…
Response retrieval is a subset of neural ranking in which a model selects a suitable response from a set of candidates given a conversation history. Retrieval-based chat-bots are typically employed in information seeking conversational systems such as customer support agents. In order to make pairwise comparisons between a conversation history and a candidate response, two approaches are common: cross-encoders performing full self-attention over the pair and bi-encoders encoding the pair separately. The former gives better prediction quality but is too slow for practical use. In this paper, we…
A computationally expensive and memory intensive neural network lies behind the recent success of language representation learning. Knowledge distillation, a major technique for deploying such a vast language model in resource-scarce environments, transfers the knowledge on individual word representations learned without restrictions. In this paper, inspired by the recent observations that language representations are relatively positioned and have more semantic knowledge as a whole, we present a new knowledge distillation objective for language representation learning that transfers the conte…
Research Institute for Artificial Intelligence, Romanian Academy（罗马尼亚科学院人工智能研究所）
Running large-scale pre-trained language models in computationally constrained environments remains a challenging problem yet to be addressed, while transfer learning from these models has become prevalent in Natural Language Processing tasks. Several solutions, including knowledge distillation, network quantization, or network pruning have been previously proposed; however, these approaches focus mostly on the English language, thus widening the gap when considering low-resource languages. In this work, we introduce three light and fast versions of distilled BERT models for the Romanian langu…
Zhejiang University（浙江大学）；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）；Alibaba Group（阿里巴巴集团）
Logical rules, both transferable and explainable, are widely used as weakly supervised signals for many downstream tasks such as named entity tagging. To reduce the human effort of writing rules, previous researchers adopt an iterative approach to automatically learn logical rules from several seed rules. However, obtaining more seed rules can only be accomplished by extra human annotation with heavy costs. Limited by the size and quality of the seed rules, the model performance of previous systems is bounded. In this paper, we develop a novel framework STREAM to distill task-specific logical …
Peking University（北京大学）；Tencent Medical AI Lab（腾讯医疗人工智能实验室）；Peng Cheng Laboratory（鹏城实验室）
Automatically generating radiology reports can improve current clinical practice in diagnostic radiology. On one hand, it can relieve radiologists from the heavy burden of report writing; On the other hand, it can remind radiologists of abnormalities and avoid the misdiagnosis and missed diagnosis. Yet, this task remains a challenging job for data-driven neural networks, due to the serious visual and textual data biases. To this end, we propose a Posterior-and-Prior Knowledge Exploring-and-Distilling approach (PPKED) to imitate the working patterns of radiologists, who will first examine the a…
Recent advances in pre-training huge models on large amounts of text through self supervision have obtained state-of-the-art results in various natural language processing tasks. However, these huge and expensive models are difficult to use in practise for downstream tasks. Some recent efforts use knowledge distillation to compress these models. However, we see a gap between the performance of the smaller student models as compared to that of the large teacher. In this work, we leverage large amounts of in-domain unlabeled transfer data in addition to a limited amount of labeled training insta…
ISTI-CNR（意大利国家研究委员会信息科学与技术研究所）；Univ. of Modena and Reggio Emilia（摩德纳和雷焦艾米利亚大学）
Image-text matching is gaining a leading role among tasks involving the joint understanding of vision and language. In literature, this task is often used as a pre-training objective to forge architectures able to jointly deal with images and texts. Nonetheless, it has a direct downstream application: cross-modal retrieval, which consists in finding images related to a given query text or vice-versa. Solving this task is of critical importance in cross-modal search engines. Many recent methods proposed effective solutions to the image-text matching problem, mostly using recent large vision-lan…
