Since hardware resources are limited, the objective of training deep learning models is typically to maximize accuracy subject to the time and memory constraints of training and inference. We study the impact of model size in this setting, focusing on Transformer models for NLP tasks that are limited by compute: self-supervised pretraining and high-resource machine translation. We first show that even though smaller Transformer models execute faster per iteration, wider and deeper models converge in significantly fewer steps. Moreover, this acceleration in convergence typically outpaces the additional computational overhead of using larger models. Therefore, the most compute-efficient training strategy is to counterintuitively train extremely large models but stop after a small number of iterations.
This leads to an apparent trade-off between the training efficiency of large Transformer models and the inference efficiency of small Transformer models. However, we show that large models are more robust to compression techniques such as quantization and pruning than small models. Consequently, one can get the best of both worlds: heavily compressed, large models achieve higher accuracy than lightly compressed, small models.
Rutgers University（罗格斯大学）；Alibaba DAMO Academy（阿里巴巴达摩院）；University of Science and Technology of China（中国科学技术大学）
Zero-shot translation, directly translating between language pairs unseen in training, is a promising capability of multilingual neural machine translation (NMT). However, it usually suffers from capturing spurious correlations between the output language and language invariant semantics due to the maximum likelihood training objective, leading to poor transfer performance on zero-shot translation. In this paper, we introduce a denoising autoencoder objective based on pivot language into traditional training objective to improve the translation accuracy on zero-shot directions. The theoretical…
Tsinghua University（清华大学）；Beijing MoreHealth Technology Group Co. Ltd（北京妙健康科技集团有限公司）
Extractive Reading Comprehension (ERC) has made tremendous advances enabled by the availability of large-scale high-quality ERC training data. Despite of such rapid progress and widespread application, the datasets in languages other than high-resource languages such as English remain scarce. To address this issue, we propose a Cross-Lingual Transposition ReThinking (XLTT) model by modelling existing high-quality extractive reading comprehension datasets in a multilingual environment. To be specific, we present multilingual adaptive attention (MAA) to combine intra-attention and inter-attentio…
DeferKV: Rethinking Eviction Timing for One-Shot KV Cache Compression
University of Electronic Science and Technology of China（电子科技大学）；Ubiquitous Intelligence and Trusted Services Key Laboratory of Sichuan Province（四川省泛在智能与可信服务重点实验室）
Tokyo Institute of Technology（东京工业大学）；RIKEN（理化学研究所）；Tohoku University（东北大学）
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 rand…
Exposure bias describes the phenomenon that a language model trained under the teacher forcing schema may perform poorly at the inference stage when its predictions are conditioned on its previous predictions unseen from the training corpus. Recently, several generative adversarial networks (GANs) and reinforcement learning (RL) methods have been introduced to alleviate this problem. Nonetheless, a common issue in RL and GANs training is the sparsity of reward signals. In this paper, we adopt two simple strategies, multi-range reinforcing, and multi-entropy sampling, to amplify and denoise the…
Pretraining Neural Language Models (NLMs) over a large corpus involves chunking the text into training examples, which are contiguous text segments of sizes processable by the neural architecture. We highlight a bias introduced by this common practice: we prove that the pretrained NLM can model much stronger dependencies between text segments that appeared in the same training example, than it can between text segments that appeared in different training examples. This intuitive result has a twofold role. First, it formalizes the motivation behind a broad line of recent successful NLM training…
The Pennsylvania State University（宾夕法尼亚州立大学）；Moffett AI（墨菲特人工智能公司）
Transformer-based pre-trained language models have significantly improved the performance of various natural language processing (NLP) tasks in the recent years. While effective and prevalent, these models are usually prohibitively large for resource-limited deployment scenarios. A thread of research has thus been working on applying network pruning techniques under the pretrain-then-finetune paradigm widely adopted in NLP. However, the existing pruning results on benchmark transformers, such as BERT, are not as remarkable as the pruning results in the literature of convolutional neural networ…
The Chinese University of Hong Kong（香港中文大学）；Tencent AI Lab（腾讯AI实验室）
Transformer becomes the state-of-the-art translation model, while it is not well studied how each intermediate component contributes to the model performance, which poses significant challenges for designing optimal architectures. In this work, we bridge this gap by evaluating the impact of individual component (sub-layer) in trained Transformer models from different perspectives. Experimental results across language pairs, training strategies, and model capacities show that certain components are consistently more important than the others. We also report a number of interesting findings that…
Extreme Multi-label Text Classification (XMC) involves learning a classifier that can assign an input with a subset of most relevant labels from millions of label choices. Recent approaches, such as XR-Transformer and LightXML, leverage a transformer instance to achieve state-of-the-art performance. However, in this process, these approaches need to make various trade-offs between performance and computational requirements. A major shortcoming, as compared to the Bi-LSTM based AttentionXML, is that they fail to keep separate feature representations for each resolution in a label tree. We thus …
Rethinking Length-Based Training: Batch Composition and Loss Normalization in Speech Token Language Models
Beijing Institute of Technology, Zhuhai（北京理工大学珠海校区）；Institute of Science Tokyo（东京科学大学）；Northeastern University（东北大学）；Sichuan University（四川大学）；Wuhan University（武汉大学）；Ant Group（蚂蚁集团）
Shuo Wen Jie Zi: Rethinking Dictionaries and Glyphs for Chinese Language Pre-training
Peking University（北京大学）；Beijing Institute for General Artificial Intelligence (BIGAI)（北京通用人工智能研究院）；National Key Laboratory of General Artificial Intelligence（通用人工智能全国重点实验室）
University of Rochester（罗切斯特大学）；Tencent AI Lab（腾讯人工智能实验室）；University of Pennsylvania（宾夕法尼亚大学）
Despite the success of large language models (LLMs) in various natural language processing (NLP) tasks, the stored knowledge in these models may inevitably be incomplete, out-of-date, or incorrect. This motivates the need to utilize external knowledge to assist LLMs. Unfortunately, current methods for incorporating external knowledge often require additional training or fine-tuning, which can be costly and may not be feasible for LLMs. To address this issue, we propose a novel post-processing approach, rethinking with retrieval (RR), which retrieves relevant external knowledge based on the dec…
