PyTorchKR Papers

최신 논문

총 5,082편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2207.00552

Reduce Indonesian Vocabularies with an Indonesian Sub-word Separator

Indonesian is an agglutinative language since it has a compounding process of word-formation. Therefore, the translation model of this language requires a mechanism that is even lower than the word level, referred to as the sub-word level. This compounding process leads to a rare word problem since the number of vocabulary explodes. We propose a strategy to address the unique word problem of the neural machine translation (NMT) system, which uses Indonesian as a pair language. Our approach uses a rule-based method to transform a word into its roots and accompanied affixes to retain its meaning and context. Using a rule-based algorithm has more advantages: it does not require corpus data but only applies the standard Indonesian rules. Our experiments confirm that this method is practical. It reduces the number of vocabulary significantly up to 57\%, and on the English to Indonesian translation, this strategy provides an improvement of up to 5 BLEU points over a similar NMT system that does not use this technique.

arXiv:2206.15407공식 코드 · Papers with Code

Shifts 2.0: Extending The Dataset of Real Distributional Shifts

Distributional shift, or the mismatch between training and deployment data, is a significant obstacle to the usage of machine learning in high-stakes industrial applications, such as autonomous driving and medicine. This creates a need to be able to assess how robustly ML models generalize as well as the quality of their uncertainty estimates. Standard ML baseline datasets do not allow these properties to be assessed, as the training, validation and test data are often identically distributed. Recently, a range of dedicated benchmarks have appeared, featuring both distributionally matched and shifted data. Among these benchmarks, the Shifts dataset stands out in terms of the diversity of tasks as well as the data modalities it features. While most of the benchmarks are heavily dominated by 2D image classification tasks, Shifts contains tabular weather forecasting, machine translation, and vehicle motion prediction tasks. This enables the robustness properties of models to be assessed on a diverse set of industrial-scale tasks and either universal or directly applicable task-specific conclusions to be reached. In this paper, we extend the Shifts Dataset with two datasets sourced from industrial, high-risk applications of high societal importance. Specifically, we consider the tasks of segmentation of white matter Multiple Sclerosis lesions in 3D magnetic resonance brain images and the estimation of power consumption in marine cargo vessels. Both tasks feature ubiquitous distributional shifts and a strict safety requirement due to the high cost of errors. These new datasets will allow researchers to further explore robust generalization and uncertainty estimation in new situations. In this work, we provide a description of the dataset and baseline results for both tasks.

arXiv:2206.14982

Building Multilingual Machine Translation Systems That Serve Arbitrary X-Y Translations

Multilingual Neural Machine Translation (MNMT) enables one system to translate sentences from multiple source languages to multiple target languages, greatly reducing deployment costs compared with conventional bilingual systems. The MNMT training benefit, however, is often limited to many-to-one directions. The model suffers from poor performance in one-to-many and many-to-many with zero-shot setup. To address this issue, this paper discusses how to practically build MNMT systems that serve arbitrary X-Y translation directions while leveraging multilinguality with a two-stage training strategy of pretraining and finetuning. Experimenting with the WMT'21 multilingual translation task, we demonstrate that our systems outperform the conventional baselines of direct bilingual models and pivot translation models for most directions, averagely giving +6.0 and +4.1 BLEU, without the need for architecture change or extra data collection. Moreover, we also examine our proposed approach in an extremely large-scale data setting to accommodate practical deployment scenarios.

arXiv:2206.14532공식 코드 · Papers with Code

Revisiting Label Smoothing and Knowledge Distillation Compatibility: What was Missing?

This work investigates the compatibility between label smoothing (LS) and knowledge distillation (KD). Contemporary findings addressing this thesis statement take dichotomous standpoints: Muller et al. (2019) and Shen et al. (2021b). Critically, there is no effort to understand and resolve these contradictory findings, leaving the primal question -- to smooth or not to smooth a teacher network? -- unanswered. The main contributions of our work are the discovery, analysis and validation of systematic diffusion as the missing concept which is instrumental in understanding and resolving these contradictory findings. This systematic diffusion essentially curtails the benefits of distilling from an LS-trained teacher, thereby rendering KD at increased temperatures ineffective. Our discovery is comprehensively supported by large-scale experiments, analyses and case studies including image classification, neural machine translation and compact student distillation tasks spanning across multiple datasets and teacher-student architectures. Based on our analysis, we suggest practitioners to use an LS-trained teacher with a low-temperature transfer to achieve high performance students. Code and models are available at https://keshik6.github.io/revisiting-ls-kd-compatibility/

arXiv:2206.14504공식 코드 · Papers with Code

GERNERMED++: Transfer Learning in German Medical NLP

We present a statistical model for German medical natural language processing trained for named entity recognition (NER) as an open, publicly available model. The work serves as a refined successor to our first GERNERMED model which is substantially outperformed by our work. We demonstrate the effectiveness of combining multiple techniques in order to achieve strong results in entity recognition performance by the means of transfer-learning on pretrained deep language models (LM), word-alignment and neural machine translation. Due to the sparse situation on open, public medical entity recognition models for German texts, this work offers benefits to the German research community on medical NLP as a baseline model. Since our model is based on public English data, its weights are provided without legal restrictions on usage and distribution. The sample code and the statistical model is available at: https://github.com/frankkramer-lab/GERNERMED-pp

arXiv:2206.12529

Probing Causes of Hallucinations in Neural Machine Translations

Hallucination, one kind of pathological translations that bothers Neural Machine Translation, has recently drawn much attention. In simple terms, hallucinated translations are fluent sentences but barely related to source inputs. Arguably, it remains an open problem how hallucination occurs. In this paper, we propose to use probing methods to investigate the causes of hallucinations from the perspective of model architecture, aiming to avoid such problems in future architecture designs. By conducting experiments over various NMT datasets, we find that hallucination is often accompanied by the deficient encoder, especially embeddings, and vulnerable cross-attentions, while, interestingly, cross-attention mitigates some errors caused by the encoder.

arXiv:2206.12840

Your Autoregressive Generative Model Can be Better If You Treat It as an Energy-Based One

Autoregressive generative models are commonly used, especially for those tasks involving sequential data. They have, however, been plagued by a slew of inherent flaws due to the intrinsic characteristics of chain-style conditional modeling (e.g., exposure bias or lack of long-range coherence), severely limiting their ability to model distributions properly. In this paper, we propose a unique method termed E-ARM for training autoregressive generative models that takes advantage of a well-designed energy-based learning objective. By leveraging the extra degree of freedom of the softmax operation, we are allowed to make the autoregressive model itself be an energy-based model for measuring the likelihood of input without introducing any extra parameters. Furthermore, we show that E-ARM can be trained efficiently and is capable of alleviating the exposure bias problem and increase temporal coherence for autoregressive generative models. Extensive empirical results, covering benchmarks like language modeling, neural machine translation, and image generation, demonstrate the effectiveness of the proposed approach.

arXiv:2206.12664공식 코드 · Papers with Code

Evaluation of Semantic Answer Similarity Metrics

There are several issues with the existing general machine translation or natural language generation evaluation metrics, and question-answering (QA) systems are indifferent in that context. To build robust QA systems, we need the ability to have equivalently robust evaluation systems to verify whether model predictions to questions are similar to ground-truth annotations. The ability to compare similarity based on semantics as opposed to pure string overlap is important to compare models fairly and to indicate more realistic acceptance criteria in real-life applications. We build upon the first to our knowledge paper that uses transformer-based model metrics to assess semantic answer similarity and achieve higher correlations to human judgement in the case of no lexical overlap. We propose cross-encoder augmented bi-encoder and BERTScore models for semantic answer similarity, trained on a new dataset consisting of name pairs of US-American public figures. As far as we are concerned, we provide the first dataset of co-referent name string pairs along with their similarities, which can be used for training.

arXiv:2206.11860

Exploiting Transliterated Words for Finding Similarity in Inter-Language News Articles using Machine Learning

Finding similarities between two inter-language news articles is a challenging problem of Natural Language Processing (NLP). It is difficult to find similar news articles in a different language other than the native language of user, there is a need for a Machine Learning based automatic system to find the similarity between two inter-language news articles. In this article, we propose a Machine Learning model with the combination of English Urdu word transliteration which will show whether the English news article is similar to the Urdu news article or not. The existing approaches to find similarities has a major drawback when the archives contain articles of low-resourced languages like Urdu along with English news article. The existing approaches to find similarities has drawback when the archives contain low-resourced languages like Urdu along with English news articles. We used lexicon to link Urdu and English news articles. As Urdu language processing applications like machine translation, text to speech, etc are unable to handle English text at the same time so this research proposed technique to find similarities in English and Urdu news articles based on transliteration.

arXiv:2206.11062

Answer Fast: Accelerating BERT on the Tensor Streaming Processor

Transformers have become a predominant machine learning workload, they are not only the de-facto standard for natural language processing tasks, but they are also being deployed in other domains such as vision and speech recognition. Many of the transformer-based applications are real-time systems such as machine translation and web search. These real time systems often come with strict end-to-end inference latency requirements. Unfortunately, while the majority of the transformer computation comes from matrix multiplications, transformers also include several non-linear components that tend to become the bottleneck during an inference. In this work, we accelerate the inference of BERT models on the tensor streaming processor. By carefully fusing all the nonlinear components with the matrix multiplication components, we are able to efficiently utilize the on-chip matrix multiplication units resulting in a deterministic tail latency of 130 $\mu$s for a batch-1 inference through BERT-base, which is 6X faster than the current state-of-the-art.

arXiv:2206.10919

Comparing Formulaic Language in Human and Machine Translation: Insight from a Parliamentary Corpus

A recent study has shown that, compared to human translations, neural machine translations contain more strongly-associated formulaic sequences made of relatively high-frequency words, but far less strongly-associated formulaic sequences made of relatively rare words. These results were obtained on the basis of translations of quality newspaper articles in which human translations can be thought to be not very literal. The present study attempts to replicate this research using a parliamentary corpus. The text were translated from French to English by three well-known neural machine translation systems: DeepL, Google Translate and Microsoft Translator. The results confirm the observations on the news corpus, but the differences are less strong. They suggest that the use of text genres that usually result in more literal translations, such as parliamentary corpora, might be preferable when comparing human and machine translations. Regarding the differences between the three neural machine systems, it appears that Google translations contain fewer highly collocational bigrams, identified by the CollGram technique, than Deepl and Microsoft translations.

arXiv:2206.10789

Scaling Autoregressive Models for Content-Rich Text-to-Image Generation

We present the Pathways Autoregressive Text-to-Image (Parti) model, which generates high-fidelity photorealistic images and supports content-rich synthesis involving complex compositions and world knowledge. Parti treats text-to-image generation as a sequence-to-sequence modeling problem, akin to machine translation, with sequences of image tokens as the target outputs rather than text tokens in another language. This strategy can naturally tap into the rich body of prior work on large language models, which have seen continued advances in capabilities and performance through scaling data and model sizes. Our approach is simple: First, Parti uses a Transformer-based image tokenizer, ViT-VQGAN, to encode images as sequences of discrete tokens. Second, we achieve consistent quality improvements by scaling the encoder-decoder Transformer model up to 20B parameters, with a new state-of-the-art zero-shot FID score of 7.23 and finetuned FID score of 3.22 on MS-COCO. Our detailed analysis on Localized Narratives as well as PartiPrompts (P2), a new holistic benchmark of over 1600 English prompts, demonstrate the effectiveness of Parti across a wide variety of categories and difficulty aspects. We also explore and highlight limitations of our models in order to define and exemplify key areas of focus for further improvements. See https://parti.research.google/ for high-resolution images.

arXiv:2206.11184공식 코드 · Papers with Code

Towards Unsupervised Content Disentanglement in Sentence Representations via Syntactic Roles

Linking neural representations to linguistic factors is crucial in order to build and analyze NLP models interpretable by humans. Among these factors, syntactic roles (e.g. subjects, direct objects,$\dots$) and their realizations are essential markers since they can be understood as a decomposition of predicative structures and thus the meaning of sentences. Starting from a deep probabilistic generative model with attention, we measure the interaction between latent variables and realizations of syntactic roles and show that it is possible to obtain, without supervision, representations of sentences where different syntactic roles correspond to clearly identified different latent variables. The probabilistic model we propose is an Attention-Driven Variational Autoencoder (ADVAE). Drawing inspiration from Transformer-based machine translation models, ADVAEs enable the analysis of the interactions between latent variables and input tokens through attention. We also develop an evaluation protocol to measure disentanglement with regard to the realizations of syntactic roles. This protocol is based on attention maxima for the encoder and on latent variable perturbations for the decoder. Our experiments on raw English text from the SNLI dataset show that $\textit{i)}$ disentanglement of syntactic roles can be induced without supervision, $\textit{ii)}$ ADVAE separates syntactic roles better than classical sequence VAEs and Transformer VAEs, $\textit{iii)}$ realizations of syntactic roles can be separately modified in sentences by mere intervention on the associated latent variables. Our work constitutes a first step towards unsupervised controllable content generation. The code for our work is publicly available.

arXiv:2206.09379

0/1 Deep Neural Networks via Block Coordinate Descent

The step function is one of the simplest and most natural activation functions for deep neural networks (DNNs). As it counts 1 for positive variables and 0 for others, its intrinsic characteristics (e.g., discontinuity and no viable information of subgradients) impede its development for several decades. Even if there is an impressive body of work on designing DNNs with continuous activation functions that can be deemed as surrogates of the step function, it is still in the possession of some advantageous properties, such as complete robustness to outliers and being capable of attaining the best learning-theoretic guarantee of predictive accuracy. Hence, in this paper, we aim to train DNNs with the step function used as an activation function (dubbed as 0/1 DNNs). We first reformulate 0/1 DNNs as an unconstrained optimization problem and then solve it by a block coordinate descend (BCD) method. Moreover, we acquire closed-form solutions for sub-problems of BCD as well as its convergence properties. Furthermore, we also integrate $\ell_{2,0}$-regularization into 0/1 DNN to accelerate the training process and compress the network scale. As a result, the proposed algorithm has a high performance on classifying MNIST and Fashion-MNIST datasets. As a result, the proposed algorithm has a desirable performance on classifying MNIST, FashionMNIST, Cifar10, and Cifar100 datasets.

arXiv:2206.08727

The ITU Faroese Pairs Dataset

This article documents a dataset of sentence pairs between Faroese and Danish, produced at ITU Copenhagen. The data covers tranlsation from both source languages, and is intended for use as training data for machine translation systems in this language pair.

arXiv:2206.08709

Statistical and Neural Methods for Cross-lingual Entity Label Mapping in Knowledge Graphs

Knowledge bases such as Wikidata amass vast amounts of named entity information, such as multilingual labels, which can be extremely useful for various multilingual and cross-lingual applications. However, such labels are not guaranteed to match across languages from an information consistency standpoint, greatly compromising their usefulness for fields such as machine translation. In this work, we investigate the application of word and sentence alignment techniques coupled with a matching algorithm to align cross-lingual entity labels extracted from Wikidata in 10 languages. Our results indicate that mapping between Wikidata's main labels stands to be considerably improved (up to $20$ points in F1-score) by any of the employed methods. We show how methods relying on sentence embeddings outperform all others, even across different scripts. We believe the application of such techniques to measure the similarity of label pairs, coupled with a knowledge base rich in high-quality entity labels, to be an excellent asset to machine translation.

arXiv:2206.07245공식 코드 · Papers with Code

An Extractive-and-Abstractive Framework for Source Code Summarization

(Source) Code summarization aims to automatically generate summaries/comments for a given code snippet in the form of natural language. Such summaries play a key role in helping developers understand and maintain source code. Existing code summarization techniques can be categorized into extractive methods and abstractive methods. The extractive methods extract a subset of important statements and keywords from the code snippet using retrieval techniques, and generate a summary that preserves factual details in important statements and keywords. However, such a subset may miss identifier or entity naming, and consequently, the naturalness of generated summary is usually poor. The abstractive methods can generate human-written-like summaries leveraging encoder-decoder models from the neural machine translation domain. The generated summaries however often miss important factual details. To generate human-written-like summaries with preserved factual details, we propose a novel extractive-and-abstractive framework. The extractive module in the framework performs a task of extractive code summarization, which takes in the code snippet and predicts important statements containing key factual details. The abstractive module in the framework performs a task of abstractive code summarization, which takes in the entire code snippet and important statements in parallel and generates a succinct and human-written-like natural language summary. We evaluate the effectiveness of our technique, called EACS, by conducting extensive experiments on three datasets involving six programming languages. Experimental results show that EACS significantly outperforms state-of-the-art techniques in terms of all three widely used metrics, including BLEU, METEOR, and ROUGH-L.

arXiv:2206.06586

FreeTransfer-X: Safe and Label-Free Cross-Lingual Transfer from Off-the-Shelf Models

Cross-lingual transfer (CLT) is of various applications. However, labeled cross-lingual corpus is expensive or even inaccessible, especially in the fields where labels are private, such as diagnostic results of symptoms in medicine and user profiles in business. Nevertheless, there are off-the-shelf models in these sensitive fields. Instead of pursuing the original labels, a workaround for CLT is to transfer knowledge from the off-the-shelf models without labels. To this end, we define a novel CLT problem named FreeTransfer-X that aims to achieve knowledge transfer from the off-the-shelf models in rich-resource languages. To address the problem, we propose a 2-step knowledge distillation (KD, Hinton et al., 2015) framework based on multilingual pre-trained language models (mPLM). The significant improvement over strong neural machine translation (NMT) baselines demonstrates the effectiveness of the proposed method. In addition to reducing annotation cost and protecting private labels, the proposed method is compatible with different networks and easy to be deployed. Finally, a range of analyses indicate the great potential of the proposed method.

arXiv:2206.06383

An Exploration of Post-Editing Effectiveness in Text Summarization

Automatic summarization methods are efficient but can suffer from low quality. In comparison, manual summarization is expensive but produces higher quality. Can humans and AI collaborate to improve summarization performance? In similar text generation tasks (e.g., machine translation), human-AI collaboration in the form of "post-editing" AI-generated text reduces human workload and improves the quality of AI output. Therefore, we explored whether post-editing offers advantages in text summarization. Specifically, we conducted an experiment with 72 participants, comparing post-editing provided summaries with manual summarization for summary quality, human efficiency, and user experience on formal (XSum news) and informal (Reddit posts) text. This study sheds valuable insights on when post-editing is useful for text summarization: it helped in some cases (e.g., when participants lacked domain knowledge) but not in others (e.g., when provided summaries include inaccurate information). Participants' different editing strategies and needs for assistance offer implications for future human-AI summarization systems.

arXiv:2206.06323

Visual Transformer for Object Detection

Convolutional Neural networks (CNN) have been the first choice of paradigm in many computer vision applications. The convolution operation however has a significant weakness which is it only operates on a local neighborhood of pixels, thus it misses global information of the surrounding neighbors. Transformers, or Self-attention networks to be more specific, on the other hand, have emerged as a recent advance to capture long range interactions of the input, but they have mostly been applied to sequence modeling tasks such as Neural Machine Translation, Image captioning and other Natural Language Processing tasks. Transformers has been applied to natural language related tasks and achieved promising results. However, its applications in visual related tasks are far from being satisfying. Taking into consideration of both the weaknesses of Convolutional Neural Networks and those of the Transformers, in this paper, we consider the use of self-attention for discriminative visual tasks, object detection, as an alternative to convolutions. In this paper, we propose our model: DetTransNet. Extensive experiments show that our model leads to consistent improvements in object detection on COCO across many different models and scales, including ResNets, while keeping the number of parameters similar. In particular, our method achieves a 1.2% Average Precision improvement on COCO object detection task over other baseline models.