Unsupervised Data Augmentation (UDA) is a semisupervised technique that applies a consistency loss to penalize differences between a model's predictions on (a) observed (unlabeled) examples; and (b) corresponding noised' examples produced via data au
gmentation. While UDA has gained popularity for text classification, open questions linger over which design decisions are necessary and how to extend the method to sequence labeling tasks. In this paper, we re-examine UDA and demonstrate its efficacy on several sequential tasks. Our main contribution is an empirical study of UDA to establish which components of the algorithm confer benefits in NLP. Notably, although prior work has emphasized the use of clever augmentation techniques including back-translation, we find that enforcing consistency between predictions assigned to observed and randomly substituted words often yields comparable (or greater) benefits compared to these more complex perturbation models. Furthermore, we find that applying UDA's consistency loss affords meaningful gains without any unlabeled data at all, i.e., in a standard supervised setting. In short, UDA need not be unsupervised to realize much of its noted benefits, and does not require complex data augmentation to be effective.
Back-translation (BT) of target monolingual corpora is a widely used data augmentation strategy for neural machine translation (NMT), especially for low-resource language pairs. To improve effectiveness of the available BT data, we introduce HintedBT
---a family of techniques which provides hints (through tags) to the encoder and decoder. First, we propose a novel method of using both high and low quality BT data by providing hints (as source tags on the encoder) to the model about the quality of each source-target pair. We don't filter out low quality data but instead show that these hints enable the model to learn effectively from noisy data. Second, we address the problem of predicting whether a source token needs to be translated or transliterated to the target language, which is common in cross-script translation tasks (i.e., where source and target do not share the written script). For such cases, we propose training the model with additional hints (as target tags on the decoder) that provide information about the operation required on the source (translation or both translation and transliteration). We conduct experiments and detailed analyses on standard WMT benchmarks for three cross-script low/medium-resource language pairs: Hindi,Gujarati,Tamil-to-English. Our methods compare favorably with five strong and well established baselines. We show that using these hints, both separately and together, significantly improves translation quality and leads to state-of-the-art performance in all three language pairs in corresponding bilingual settings.
Despite showing increasingly human-like conversational abilities, state-of-the-art dialogue models often suffer from factual incorrectness and hallucination of knowledge (Roller et al., 2020). In this work we explore the use of neural-retrieval-in-th
e-loop architectures - recently shown to be effective in open-domain QA (Lewis et al., 2020b; Izacard and Grave, 2020) - for knowledge-grounded dialogue, a task that is arguably more challenging as it requires querying based on complex multi-turn dialogue context and generating conversationally coherent responses. We study various types of architectures with multiple components - retrievers, rankers, and encoder-decoders - with the goal of maximizing knowledgeability while retaining conversational ability. We demonstrate that our best models obtain state-of-the-art performance on two knowledge-grounded conversational tasks. The models exhibit open-domain conversational capabilities, generalize effectively to scenarios not within the training data, and, as verified by human evaluations, substantially reduce the well-known problem of knowledge hallucination in state-of-the-art chatbots.
Word embeddings capture semantic meaning of individual words. How to bridge word-level linguistic knowledge with sentence-level language representation is an open problem. This paper examines whether sentence-level representations can be achieved by
building a custom sentence database focusing on one aspect of a sentence's meaning. Our three separate semantic aspects are whether the sentence: (1) communicates a causal relationship, (2) indicates that two things are correlated with each other, and (3) expresses information or knowledge. The three classifiers provide epistemic information about a sentence's content.
Neural language models typically tokenise input text into sub-word units to achieve an open vocabulary. The standard approach is to use a single canonical tokenisation at both train and test time. We suggest that this approach is unsatisfactory and m
ay bottleneck our evaluation of language model performance. Using only the one-best tokenisation ignores tokeniser uncertainty over alternative tokenisations, which may hurt model out-of-domain performance. In this paper, we argue that instead, language models should be evaluated on their marginal likelihood over tokenisations. We compare different estimators for the marginal likelihood based on sampling, and show that it is feasible to estimate the marginal likelihood with a manageable number of samples. We then evaluate a pretrained language model on both the one-best-tokenisation and marginal perplexities, and show that the marginal perplexity can be significantly better than the one best, especially on out-of-domain data. We link this difference in perplexity to the tokeniser uncertainty as measured by tokeniser entropy. We discuss some implications of our results for language model training and evaluation, particularly with regard to tokenisation robustness.
Recent state-of-the-art (SOTA) effective neural network methods and fine-tuning methods based on pre-trained models (PTM) have been used in Chinese word segmentation (CWS), and they achieve great results. However, previous works focus on training the
models with the fixed corpus at every iteration. The intermediate generated information is also valuable. Besides, the robustness of the previous neural methods is limited by the large-scale annotated data. There are a few noises in the annotated corpus. Limited efforts have been made by previous studies to deal with such problems. In this work, we propose a self-supervised CWS approach with a straightforward and effective architecture. First, we train a word segmentation model and use it to generate the segmentation results. Then, we use a revised masked language model (MLM) to evaluate the quality of the segmentation results based on the predictions of the MLM. Finally, we leverage the evaluations to aid the training of the segmenter by improved minimum risk training. Experimental results show that our approach outperforms previous methods on 9 different CWS datasets with single criterion training and multiple criteria training and achieves better robustness.
Current language models are usually trained using a self-supervised scheme, where the main focus is learning representations at the word or sentence level. However, there has been limited progress in generating useful discourse-level representations.
In this work, we propose to use ideas from predictive coding theory to augment BERT-style language models with a mechanism that allows them to learn suitable discourse-level representations. As a result, our proposed approach is able to predict future sentences using explicit top-down connections that operate at the intermediate layers of the network. By experimenting with benchmarks designed to evaluate discourse-related knowledge using pre-trained sentence representations, we demonstrate that our approach improves performance in 6 out of 11 tasks by excelling in discourse relationship detection.
Executing natural language instructions in a physically grounded domain requires a model that understands both spatial concepts such as left of'' and above'', and the compositional language used to identify landmarks and articulate instructions relat
ive to them. In this paper, we study instruction understanding in the blocks world domain. Given an initial arrangement of blocks and a natural language instruction, the system executes the instruction by manipulating selected blocks. The highly compositional instructions are composed of atomic components and understanding these components is a necessary step to executing the instruction. We show that while end-to-end training (supervised only by the correct block location) fails to address the challenges of this task and performs poorly on instructions involving a single atomic component, knowledge-free auxiliary signals can be used to significantly improve performance by providing supervision for the instruction's components. Specifically, we generate signals that aim at helping the model gradually understand components of the compositional instructions, as well as those that help it better understand spatial concepts, and show their benefit to the overall task for two datasets and two state-of-the-art (SOTA) models, especially when the training data is limited---which is usual in such tasks.
The aim of this study is to evaluate Jordanian Women's opinions, beliefs, and practices towards using different medicinal plants for postpartum health problems care.
Method: A cross-sectional observational study was conducted on 300 mothers aged 18
years and above. A structured valid and reliable questionnaire was used for collecting personal, medical and nutritional related data including: gestational weight gain characteristics, the effect of delivery and breast feeding on postpartum weight gain and herbal tea consumption for the management of different postpartum health problems such as postpartum colic, flatulence, spasm, maternal bleeding, lactation and weight gain. The above data where collected through a personal interview by the trained investigators.
Results: Around 45% of participants were overweight or obese with average post-pregnancy BMI of 25.1±4.94 kg/m2. Majority of participants (84%) used one or more medicinal plants after delivery to control their postpartum health problems. The participants may seek herbal help mainly for maternal purposes such as decreasing post-delivery colic, flatulence and spasm (52.9%), treating maternal postpartum bleeding (41.7%) and lactation enhancement (41.0%). Conversely, only 9.0% of participants used herbals for weight control. The most commonly used herbals were cinnamon (49.0%), sage (42.0%), and anise (38.0%).
Conclusions: The potential risk of medicinal plant self-medication is high for managing postpartum complications that need a professional evidence-based practice recommendation
Recently, pre-trained language representation models such as BERT and RoBERTa have achieved significant results in a wide range of natural language processing (NLP) tasks, however, it requires extremely high computational cost. Curriculum Learning (C
L) is one of the potential solutions to alleviate this problem. CL is a training strategy where training samples are given to models in a meaningful order instead of random sampling. In this work, we propose a new CL method which gradually increases the block-size of input text for training the self-attention mechanism of BERT and its variants using the maximum available batch-size. Experiments in low-resource settings show that our approach outperforms the baseline in terms of convergence speed and final performance on downstream tasks.
