ترغب بنشر مسار تعليمي؟ اضغط هنا

Nystromformer: A Nystrom-Based Algorithm for Approximating Self-Attention

120   0   0.0 ( 0 )
 نشر من قبل Yunyang Xiong
 تاريخ النشر 2021
  مجال البحث الهندسة المعلوماتية
والبحث باللغة English




اسأل ChatGPT حول البحث

Transformers have emerged as a powerful tool for a broad range of natural language processing tasks. A key component that drives the impressive performance of Transformers is the self-attention mechanism that encodes the influence or dependence of other tokens on each specific token. While beneficial, the quadratic complexity of self-attention on the input sequence length has limited its application to longer sequences -- a topic being actively studied in the community. To address this limitation, we propose Nystr{o}mformer -- a model that exhibits favorable scalability as a function of sequence length. Our idea is based on adapting the Nystr{o}m method to approximate standard self-attention with $O(n)$ complexity. The scalability of Nystr{o}mformer enables application to longer sequences with thousands of tokens. We perform evaluations on multiple downstream tasks on the GLUE benchmark and IMDB reviews with standard sequence length, and find that our Nystr{o}mformer performs comparably, or in a few cases, even slightly better, than standard self-attention. On longer sequence tasks in the Long Range Arena (LRA) benchmark, Nystr{o}mformer performs favorably relative to other efficient self-attention methods. Our code is available at https://github.com/mlpen/Nystromformer.



قيم البحث

اقرأ أيضاً

In this paper, we introduce the prior knowledge, multi-scale structure, into self-attention modules. We propose a Multi-Scale Transformer which uses multi-scale multi-head self-attention to capture features from different scales. Based on the linguis tic perspective and the analysis of pre-trained Transformer (BERT) on a huge corpus, we further design a strategy to control the scale distribution for each layer. Results of three different kinds of tasks (21 datasets) show our Multi-Scale Transformer outperforms the standard Transformer consistently and significantly on small and moderate size datasets.
In longitudinal electronic health records (EHRs), the event records of a patient are distributed over a long period of time and the temporal relations between the events reflect sufficient domain knowledge to benefit prediction tasks such as the rate of inpatient mortality. Medical concept embedding as a feature extraction method that transforms a set of medical concepts with a specific time stamp into a vector, which will be fed into a supervised learning algorithm. The quality of the embedding significantly determines the learning performance over the medical data. In this paper, we propose a medical concept embedding method based on applying a self-attention mechanism to represent each medical concept. We propose a novel attention mechanism which captures the contextual information and temporal relationships between medical concepts. A light-weight neural net, Temporal Self-Attention Network (TeSAN), is then proposed to learn medical concept embedding based solely on the proposed attention mechanism. To test the effectiveness of our proposed methods, we have conducted clustering and prediction tasks on two public EHRs datasets comparing TeSAN against five state-of-the-art embedding methods. The experimental results demonstrate that the proposed TeSAN model is superior to all the compared methods. To the best of our knowledge, this work is the first to exploit temporal self-attentive relations between medical events.
One of the most popular paradigms of applying large, pre-trained NLP models such as BERT is to fine-tune it on a smaller dataset. However, one challenge remains as the fine-tuned model often overfits on smaller datasets. A symptom of this phenomenon is that irrelevant words in the sentences, even when they are obvious to humans, can substantially degrade the performance of these fine-tuned BERT models. In this paper, we propose a novel technique, called Self-Supervised Attention (SSA) to help facilitate this generalization challenge. Specifically, SSA automatically generates weak, token-level attention labels iteratively by probing the fine-tuned model from the previous iteration. We investigate two different ways of integrating SSA into BERT and propose a hybrid approach to combine their benefits. Empirically, on a variety of public datasets, we illustrate significant performance improvement using our SSA-enhanced BERT model.
Recently, end-to-end sequence-to-sequence models for speech recognition have gained significant interest in the research community. While previous architecture choices revolve around time-delay neural networks (TDNN) and long short-term memory (LSTM) recurrent neural networks, we propose to use self-attention via the Transformer architecture as an alternative. Our analysis shows that deep Transformer networks with high learning capacity are able to exceed performance from previous end-to-end approaches and even match the conventional hybrid systems. Moreover, we trained very deep models with up to 48 Transformer layers for both encoder and decoders combined with stochastic residual connections, which greatly improve generalizability and training efficiency. The resulting models outperform all previous end-to-end ASR approaches on the Switchboard benchmark. An ensemble of these models achieve 9.9% and 17.7% WER on Switchboard and CallHome test sets respectively. This finding brings our end-to-end models to competitive levels with previous hybrid systems. Further, with model ensembling the Transformers can outperform certain hybrid systems, which are more complicated in terms of both structure and training procedure.
Review rating prediction of text reviews is a rapidly growing technology with a wide range of applications in natural language processing. However, most existing methods either use hand-crafted features or learn features using deep learning with simp le text corpus as input for review rating prediction, ignoring the hierarchies among data. In this paper, we propose a Hierarchical bi-directional self-attention Network framework (HabNet) for paper review rating prediction and recommendation, which can serve as an effective decision-making tool for the academic paper review process. Specifically, we leverage the hierarchical structure of the paper reviews with three levels of encoders: sentence encoder (level one), intra-review encoder (level two) and inter-review encoder (level three). Each encoder first derives contextual representation of each level, then generates a higher-level representation, and after the learning process, we are able to identify useful predictors to make the final acceptance decision, as well as to help discover the inconsistency between numerical review ratings and text sentiment conveyed by reviewers. Furthermore, we introduce two new metrics to evaluate models in data imbalance situations. Extensive experiments on a publicly available dataset (PeerRead) and our own collected dataset (OpenReview) demonstrate the superiority of the proposed approach compared with state-of-the-art methods.

الأسئلة المقترحة

التعليقات
جاري جلب التعليقات جاري جلب التعليقات
سجل دخول لتتمكن من متابعة معايير البحث التي قمت باختيارها
mircosoft-partner

هل ترغب بارسال اشعارات عن اخر التحديثات في شمرا-اكاديميا