United States

Existing pre-trained language models (PLMs) are often computationally expensive in inference, making them impractical in various resource-limited real-world applications. To address this issue, we propose a dynamic token reduction approach to accelerate PLMs&#39; inference, named TR-BERT, which could flexibly adapt the layer number of each token in inference to avoid redundant calculation. Specially, TR-BERT formulates the token reduction process as a multi-step token selection problem and automatically learns the selection strategy via reinforcement learning. The experimental results on several downstream NLP tasks show that TR-BERT is able to speed up BERT by 2-5 times to satisfy various performance demands. Moreover, TR-BERT can also achieve better performance with less computation in a suite of long-text tasks since its token-level layer number adaption greatly accelerates the self-attention operation in PLMs.

NAACL 2021

TR-BERT: Dynamic Token Reduction for Accelerating BERT Inference

2021 Annual Conference of the North American Chapter of the Association for Computational Linguistics.

**Whova App** 
Stay in touch with your fellow conference attendees via the [Whova App](https://whova.com/portal/webapp/nacon_202106/)

**Conference Structure**
https://2021.naacl.org/blog/conference-structure/

**Walkthrough video of how to NAACL 2021** 

Please take a moment to view this video explaining how to navigate the platform, attend sessions network with other attendees. 


<figure class="video_container">
  <iframe src="https://screencast-o-matic.com/watch/crhwbGVh3vx?v=6&ff=1&title=0&controls=1" width=640  height=350  frameborder="0" allowfullscreen="true"> </iframe>
</figure>

<br>
<br>

Existing pre-trained language models (PLMs) are often computationally expensive in inference, making them impractical in various resource-limited real-world applications. To address this issue, we propose a dynamic token reduction approach to accelerate PLMs' inference, named TR-BERT, which could flexibly adapt the layer number of each token in inference to avoid redundant calculation. Specially, TR-BERT formulates the token reduction process as a multi-step token selection problem and automatically learns the selection strategy via reinforcement learning. The experimental results on several downstream NLP tasks show that TR-BERT is able to speed up BERT by 2-5 times to satisfy various performance demands. Moreover, TR-BERT can also achieve better performance with less computation in a suite of long-text tasks since its token-level layer number adaption greatly accelerates the self-attention operation in PLMs.

technical paper

Translated texts have been used for malicious purposes, i.e., plagiarism or fake reviews. Existing detectors have been built around a specific translator (e.g., Google) but fail to detect a translated text from a strange translator. If we use the same translator, the translated text is similar to its round-trip translation, which is when text is translated into another language and translated back into the original language. However, a round-trip translated text is significantly different from the original text or a translated text using a strange translator. Hence, we propose a detector using text similarity with round-trip translation (TSRT). TSRT achieves 86.9% accuracy in detecting a translated text from a strange translator. It outperforms existing detectors (77.9%) and human recognition (53.3%).

Machine Translated Text Detection Through Text Similarity with Round-Trip Translation

There are common semantics shared across text and images. Given a sentence in a source language, whether depicting the visual scene helps translation into a target language? Existing multimodal neural machine translation methods (MNMT) require triplets of bilingual sentence - image for training and tuples of source sentence - image for inference. In this paper, we propose imagiT, a novel machine translation method via visual imagination. ImagiT first learns to generate visual representation from the source sentence, and then utilizes both source sentence and the ``imagined representation'' to produce a target translation. Unlike previous methods, it only needs the source sentence at the inference time. Experiments demonstrate that imagiT benefits from visual imagination and significantly outperforms the text-only neural machine translation baselines. Further analysis reveals that the imagination process in imagiT helps fill in missing information when performing the degradation strategy.

Generative Imagination Elevates Machine Translation

Machine translation of user-generated code-mixed inputs to English is of crucial importance in applications like web search and targeted advertising. We address the scarcity of parallel training data for training such models by designing a strategy of converting existing non-code-mixed parallel data sources to code-mixed parallel data. We present an m-BERT based procedure whose core learnable component is a ternary sequence labeling model, that can be trained with a limited code-mixed corpus alone. We show a 5.8 point increase in BLEU on heavily code-mixed sentences by training a translation model using our data augmentation strategy on an Hindi-English code-mixed translation task.

Training Data Augmentation for Code-Mixed Translation

Non-autoregressive Transformer is a promising text generation model.

Non-Autoregressive Translation by Learning Target Categorical Codes

We often use perturbations to regularize neural models. For neural encoder-decoders, previous studies applied the scheduled sampling (Bengio et al., 2015) and adversarial perturbations (Sato et al., 2019) as perturbations but these methods require considerable computational time. Thus, this study addresses the question of whether these approaches are efficient enough for training time. We compare several perturbations in sequence-to-sequence problems with respect to computational time. Experimental results show that the simple techniques such as word dropout (Gal and Ghahramani, 2016) and random replacement of input tokens achieve comparable (or better) scores to the recently proposed perturbations, even though these simple methods are faster. Our code is publicly available at https://github.com/takase/rethink_perturbations.

Rethinking Perturbations in Encoder-Decoders for Fast Training

In open-domain question answering, dense passage retrieval has become a new paradigm to retrieve relevant passages for finding answers. Typically, the dual-encoder architecture is adopted to learn dense representations of questions and passages for semantic matching. However, it is difficult to effectively train a dual-encoder due to the challenges including the discrepancy between training and inference, the existence of unlabeled positives and limited training data. To address these challenges, we propose an optimized training approach, called RocketQA, to improving dense passage retrieval. We make three major technical contributions in RocketQA, namely cross-batch negatives, denoised hard negatives and data augmentation. The experiment results show that RocketQA significantly outperforms previous state-of-the-art models on both MSMARCO and Natural Questions. We also conduct extensive experiments to examine the effectiveness of the three strategies in RocketQA. Besides, we demonstrate that the performance of end-to-end QA can be improved based on our RocketQA retriever.

RocketQA: An Optimized Training Approach to Dense Passage Retrieval for Open-Domain Question Answering

Improving Zero-Shot Cross-lingual Transfer for Multilingual Question Answering over Knowledge Graph

Recent QA with logical reasoning questions re- quires passage-level relations among the sentences. However, current approaches still focus on sentence-level relations interacting among tokens. In this work, we explore aggregating passage-level clues for solving logical reasoning QA by using discourse-based information. We propose a discourse-aware graph network (DAGN) that reasons relying on the discourse structure of the texts. The model encodes discourse information as a graph with elementary discourse units (EDUs) and dis- course relations and learns the discourse-aware features via a graph network for downstream QA tasks. Experiments are conducted on two logical reasoning QA datasets, Re- Clor and LogiQA, and our proposed DAGN achieves competitive results. The source code is available at https://github.com/Eleanor- H/DAGN.

DAGN: Discourse-Aware Graph Network for Logical Reasoning

Recently Graph Neural Network (GNN) has been used as a promising tool in multi-hop question answering task. However, the unnecessary updations and simple edge constructions prevent an accurate answer span extraction in a more direct and interpretable way. In this paper, we propose a novel model of Breadth First Reasoning Graph (BFR-Graph), which presents a new message passing way that better conforms to the reasoning process. In BFRGraph, the reasoning message is required to start from the question node and pass to the next sentences node hop by hop until all the edges have been passed, which can effectively prevent each node from over-smoothing or being updated multiple times unnecessarily. To introduce more semantics, we also define the reasoning graph as a weighted graph with considering the number of co-occurrence entities and the distance between sentences. Then we present a more direct and interpretable way to aggregate scores from different levels of granularity based on the GNN. On HotpotQA leaderboard, the proposed BFR-Graph achieves state-of-the-art on answer span prediction.

Breadth First Reasoning Graph for Multi-hop Question Answering

Obtaining training data for multi-hop question answering (QA) is time-consuming and resource-intensive. We explore the possibility to train a well-performed multi-hop QA model without referencing any human-labeled multi-hop question-answer pairs, i.e., unsupervised}multi-hop QA. We propose MQA-QG, an unsupervised framework that can generate human-like multi-hop training data from both homogeneous and heterogeneous data sources. MQA-QG generates questions by first selecting/generating relevant information from each data source and then integrating the multiple information to form a multi-hop question. Using only generated training data, we can train a competent multi-hop QA which achieves 61% and 83% of the supervised learning performance for the HybridQA and the HotpotQA dataset, respectively. We also show that pretraining the QA system with the generated data would greatly reduce the demand for human-annotated training data. Our codes are publicly available at https://github.com/teacherpeterpan/Unsupervised-Multi-hop-QA.

TR-BERT: Dynamic Token Reduction for Accelerating BERT Inference

Downloads

Next from NAACL 2021

Machine Translated Text Detection Through Text Similarity with Round-Trip Translation

Similar lecture

Extracting a Knowledge Base of Mechanisms from COVID-19 Papers

TR-BERT: Dynamic Token Reduction for Accelerating BERT Inference

.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 NAACL 2021

Machine Translated Text Detection Through Text Similarity with Round-Trip Translation

Similar lecture

Extracting a Knowledge Base of Mechanisms from COVID-19 Papers

Downloads