Thailand

Pre-trained language models (PLMs) achieve great success in NLP. However, their huge model sizes hinder their applications in many practical systems. Knowledge distillation is a popular technique to compress PLMs, which learns a small student model from a large teacher PLM. However, the knowledge learned from a single teacher may be limited and even biased, resulting in low-quality student model. In this paper, we propose a multi-teacher knowledge distillation framework named MT-BERT for pre-trained language model compression, which can train high-quality student model from multiple teacher PLMs. In MT-BERT we design a multi-teacher co-finetuning method to jointly finetune multiple teacher PLMs in downstream tasks with shared pooling and prediction layers to align their output space for better collaborative teaching. In addition, we propose a multi-teacher hidden loss and a multi-teacher distillation loss to transfer the useful knowledge in both hidden states and soft labels from multiple teacher PLMs to the student model. Experiments on three benchmark datasets validate the effectiveness of MT-BERT in compressing PLMs.

IJCNLP-AACL 2021

One Teacher is Enough? Pre-trained Language Model Distillation from Multiple Teachers

multi-teacher knowledge distillation

pre-trained language model

knowledge distillation

**Welcome to ACL-IJCNLP 2021!**

The great event is jointly organized by the Association for Computational Linguistics (ACL) and Asian Federation of Natural Language Processing (AFNLP). 

As in previous years, the program of the conference includes a poster session, tutorials, workshops and demonstrations in addition to the main conference.


We were able to keep the registration fees similar to those charged for the virtual ACL 2020. The one fee allows attendance at the main conference and any/all tutorials and workshops. These fees would be $125 Regular Early and $175 Regular Late; $50 Student Early and $75 Student Late. Early registration closes at midnight July 11, 2021 (Eastern Daylight Time).

**Reminder:** It is ACL’s policy that at least one author of each accepted paper (including ACL Finding papers) must register for the conference.

**Reminder2:** Underline site will open closer to the event. If you already registered you will receive access detail

Registration is now open

The great event is jointly organized by the Association for Computational Linguistics (ACL) and Asian Federation of Natural Language Processing (AFNLP).

technical paper

Transformer-based pre-trained language models like BERT, though powerful in many tasks, are expensive in both memory and computation, due to their large number of parameters. Previous works show that some parameters in these models can be pruned away without severe accuracy drop. However, these redundant features contribute to a comprehensive understanding of the training data and removing them weakens the model's representation ability. In this paper, we propose GhostBERT, which generates more features with very cheap operations from the remaining features. In this way, GhostBERT has similar memory and computational cost as the pruned model, but enjoys much larger representation power. The proposed ghost module can also be applied to unpruned BERT models to enhance their performance with negligible additional parameters and computation. Empirical results on the GLUE benchmark on three backbone models (i.e., BERT, RoBERTa and ELECTRA) verify the efficacy of our proposed method.

GhostBERT: Generate More Features with Cheap Operations for BERT

The Lottery Ticket Hypothesis suggests that an over-parametrized network consists of ''lottery tickets'', and training a certain collection of them (i.e., a subnetwork) can match the performance of the full model. In this paper, we study such a collection of tickets, which is referred to as ''winning tickets'', in extremely over-parametrized models, e.g., pre-trained language models. We observe that at certain compression ratios, the generalization performance of the winning tickets can not only match but also exceed that of the full model. In particular, we observe a phase transition phenomenon: As the compression ratio increases, generalization performance of the winning tickets first improves then deteriorates after a certain threshold. We refer to the tickets on the threshold as ''super tickets''. We further show that the phase transition is task and model dependent --- as the model size becomes larger and the training data set becomes smaller, the transition becomes more pronounced. Our experiments on the GLUE benchmark show that the super tickets improve single task fine-tuning by 0.9 points on BERT-base and 1.0 points on BERT-large, in terms of task-average score. We also demonstrate that adaptively sharing the super tickets across tasks benefits multi-task learning.

Super Tickets in Pre-Trained Language Models: From Model Compression to Improving Generalization

Learning disentangled representations of textual data is essential for many natural language tasks such as fair classification, style transfer and sentence generation, among others. The existent dominant approaches in the context of text data {either rely} on training an adversary (discriminator) that aims at making attribute values difficult to be inferred from the latent code {or rely on minimising variational bounds of the mutual information between latent code and the value attribute}. {However, the available methods suffer of the impossibility to provide
a fine-grained control of the degree (or force) of disentanglement.} {In contrast to} {adversarial methods}, which are remarkably simple, although the adversary seems to be performing perfectly well during the training phase, after it is completed a fair amount of information about the undesired attribute still remains. This paper introduces a novel variational upper bound to the mutual information between an attribute and the latent code of an encoder. Our bound aims at controlling the approximation error via the Renyi’s divergence, leading to both better disentangled representations and in particular, a precise control of the desirable degree of disentanglement {than state-of-the-art methods proposed for textual data}. Furthermore, it does not suffer from the degeneracy of other losses in multi-class scenarios. We show the superiority of this method on fair classification and on textual style transfer tasks. Additionally, we provide new insights illustrating various trade-offs in style transfer when attempting to learn disentangled representations and quality of the generated sentence.

A Novel Estimator of Mutual Information for Learning to Disentangle Textual Representations

Beam search is a go-to strategy for decoding neural sequence models. The algorithm can naturally be viewed as a subset optimization problem, albeit one where the corresponding set function does not reflect interactions between candidates. Empirically, this leads to sets often exhibiting high overlap, e.g., strings may differ by only a single word. Yet in use-cases that call for multiple solutions, a diverse or representative set is often desired. To address this issue, we propose a reformulation of beam search, which we call determinantal beam search. Determinantal beam search has a natural relationship to determinantal point processes (DPPs), models over sets that inherently encode intra-set interactions. By posing iterations in beam search as a series of subdeterminant maximization problems, we can turn the algorithm into a diverse subset selection process. In a case study, we use the string subsequence kernel to explicitly encourage n-gram coverage in text generated from a sequence model. We observe that our algorithm offers competitive performance against other diverse set generation strategies in the context of language generation, while providing a more general approach to optimizing for diversity.

Determinantal Beam Search

Pre-trained language models (e.g., BERT) significantly alleviate two traditional challenging problems for Chinese word segmentation (CWS): segmentation ambiguity and out-of-vocabulary (OOV) words. However, such improvements are usually achieved on traditional benchmark datasets and not close to an important goal of CWS: practicability (i.e., low complexity as a standalone task and high beneficiality to downstream tasks). To make a trade-off between traditional evaluation and practicability for CWS, we propose a semi-supervised neural method via pseudo labels. The neural method consists of a teacher model and a student model, which distills knowledge from unlabeled data to the student model so as to improve both in-domain and out-of-domain CWS. Experiments show that our proposed method can not only keep the practicability of the lightweight student model but also improve the performance of segmentation effectively. We also evaluate a range of heterogeneous neural architectures of CWS on downstream Chinese NLP tasks. Results of further experiments demonstrate that our proposed segmenter is reliable and practical as a preprocessing step of the downstream NLP tasks at the minimum cost.

Enhancing Chinese Word Segmentation via Pseudo Labels for Practicability

Code generation is the task of generating code snippets from input user specifications in natural language. Leveraging the linguistically-motivated hierarchical structure of the input can benefit code generation, especially since the specifications are complex sentences containing multiple variables and operations over various data structures. Moreover, recent advances in Transformer architectures have led to improved performance with tree-to-tree style generation for other seq2seq tasks e.g., machine translation. Hence, we present an empirical analysis of the significance of input parse trees for code generation. We run text-to-tree, linearized tree-to-tree, and structured tree-to-tree models, using constituency-based parse trees as input, where the target is Abstract Syntax Tree (AST) of the code. We evaluate our models on the Python-based code generation dataset CoNaLa and a semantic parsing dataset ATIS. We find that constituency trees encoded using a structure-aware model improve performance for both datasets. We also provide an analysis of those aspects of the input parse trees which are most impactful. For instance, we find that structure-aware encodings are better at modelling inputs with multiple variables and capturing long-range dependencies for code generation.

Analysis of Tree-Structured Architectures for Code Generation

While non-autoregressive (NAR) models are showing great promise for machine translation, their use is limited by their dependence on knowledge distillation from autoregressive models. To address this issue, we seek to understand why distillation is so effective. Prior work suggests that distilled training data is less complex than manual translations. Based on experiments with the Levenshtein Transformer and the Mask-Predict NAR models on the WMT14 German-English task, this paper shows that different types of complexity have different impacts: while reducing lexical diversity and decreasing reordering complexity both help NAR learn better alignment between source and target, and thus improve translation quality, lexical diversity is the main reason why distillation increases model confidence, which affects the calibration of different NAR models differently.

How Does Distilled Data Complexity Impact the Quality and Confidence of Non-Autoregressive Machine Translation?

Current pre-trained language models have lots of knowledge, but a more limited ability to use that knowledge. Bloom's Taxonomy helps educators teach children how to use knowledge
by categorizing comprehension skills, so we use it to analyze and improve the comprehension skills of large pre-trained language models. Our experiments focus on zero-shot question answering, using the taxonomy to provide proximal context that helps the model answer questions by being relevant to those questions. We show targeting context in this manner improves performance across 4 popular common sense question answer datasets.


Comprehension Based Question Answering using Bloom's Taxonomy

Recent work has demonstrated the effectiveness of cross-lingual language model pretraining for cross-lingual understanding. In this study, we present the results of two larger multilingual masked language models, with 3.5B and 10.7B parameters. Our two new models dubbed XLMR-XL and XLMR-XXL outperform XLM-R by 1.8% and 2.4% average accuracy on XNLI. Our model also outperforms the RoBERTa-Large model on several English tasks of the GLUE benchmark by 0.3% on average while handling 99 more languages. This suggests larger capacity models for language understanding may obtain strong performance on both high- and low-resource languages. We make our code and models publicly available.

Larger-Scale Transformers for Multilingual Masked Language Modeling


Learning effective language representations from crowdsourced labels is crucial for many real-world machine learning tasks. A challenging aspect of this problem is that the quality of crowdsourced labels suffer high intra- and inter-observer variability. Since the high-capacity deep neural networks can easily memorize all disagreements among crowdsourced labels, directly applying existing supervised language representation learning algorithms may yield suboptimal solutions. In this paper, we propose \emph{TACMA}, a \underline{t}emporal-\underline{a}ware language representation learning heuristic for \underline{c}rowdsourced labels with \underline{m}ultiple \underline{a}nnotators. The proposed approach (1) explicitly models the intra-observer variability with attention mechanism; (2) computes and aggregates per-sample confidence scores from multiple workers to address the inter-observer disagreements. The proposed heuristic is extremely easy to implement in around 5 lines of code. The proposed heuristic is evaluated on four synthetic and four real-world data sets. The results show that our approach outperforms a wide range of state-of-the-art baselines in terms of prediction accuracy and AUC. To encourage the reproducible results, we make our code publicly available at \url{https://github.com/CrowdsourcingMining/TACMA}.


One Teacher is Enough? Pre-trained Language Model Distillation from Multiple Teachers

Downloads

Next from IJCNLP-AACL 2021

GhostBERT: Generate More Features with Cheap Operations for BERT

Similar lecture

One Teacher is Enough? Pre-trained Language Model Distillation from Multiple Teachers

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES

One Teacher is Enough? Pre-trained Language Model Distillation from Multiple Teachers

.css-70qvj9{display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}Downloads

Next from IJCNLP-AACL 2021

GhostBERT: Generate More Features with Cheap Operations for BERT

Similar lecture

One Teacher is Enough? Pre-trained Language Model Distillation from Multiple Teachers

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES

Downloads