This paper presents a pilot study to automatic linguistic preprocessing of Ancient and Byzantine Greek, and morphological analysis more specifically. To this end, a novel subword-based BERT language model was trained on the basis of a varied corpus o
f Modern, Ancient and Post-classical Greek texts. Consequently, the obtained BERT embeddings were incorporated to train a fine-grained Part-of-Speech tagger for Ancient and Byzantine Greek. In addition, a corpus of Greek Epigrams was manually annotated and the resulting gold standard was used to evaluate the performance of the morphological analyser on Byzantine Greek. The experimental results show very good perplexity scores (4.9) for the BERT language model and state-of-the-art performance for the fine-grained Part-of-Speech tagger for in-domain data (treebanks containing a mixture of Classical and Medieval Greek), as well as for the newly created Byzantine Greek gold standard data set. The language models and associated code are made available for use at https://github.com/pranaydeeps/Ancient-Greek-BERT
Knowledge Distillation (KD) offers a natural way to reduce the latency and memory/energy usage of massive pretrained models that have come to dominate Natural Language Processing (NLP) in recent years. While numerous sophisticated variants of KD algo
rithms have been proposed for NLP applications, the key factors underpinning the optimal distillation performance are often confounded and remain unclear. We aim to identify how different components in the KD pipeline affect the resulting performance and how much the optimal KD pipeline varies across different datasets/tasks, such as the data augmentation policy, the loss function, and the intermediate representation for transferring the knowledge between teacher and student. To tease apart their effects, we propose Distiller, a meta KD framework that systematically combines a broad range of techniques across different stages of the KD pipeline, which enables us to quantify each component's contribution. Within Distiller, we unify commonly used objectives for distillation of intermediate representations under a universal mutual information (MI) objective and propose a class of MI-objective functions with better bias/variance trade-off for estimating the MI between the teacher and the student. On a diverse set of NLP datasets, the best Distiller configurations are identified via large-scale hyper-parameter optimization. Our experiments reveal the following: 1) the approach used to distill the intermediate representations is the most important factor in KD performance, 2) among different objectives for intermediate distillation, MI-performs the best, and 3) data augmentation provides a large boost for small training datasets or small student networks. Moreover, we find that different datasets/tasks prefer different KD algorithms, and thus propose a simple AutoDistiller algorithm that can recommend a good KD pipeline for a new dataset.
Despite constant improvements in machine translation quality, automatic poetry translation remains a challenging problem due to the lack of open-sourced parallel poetic corpora, and to the intrinsic complexities involved in preserving the semantics,
style and figurative nature of poetry. We present an empirical investigation for poetry translation along several dimensions: 1) size and style of training data (poetic vs. non-poetic), including a zero-shot setup; 2) bilingual vs. multilingual learning; and 3) language-family-specific models vs. mixed-language-family models. To accomplish this, we contribute a parallel dataset of poetry translations for several language pairs. Our results show that multilingual fine-tuning on poetic text significantly outperforms multilingual fine-tuning on non-poetic text that is 35X larger in size, both in terms of automatic metrics (BLEU, BERTScore, COMET) and human evaluation metrics such as faithfulness (meaning and poetic style). Moreover, multilingual fine-tuning on poetic data outperforms bilingual fine-tuning on poetic data.
Dialogue summarization helps readers capture salient information from long conversations in meetings, interviews, and TV series. However, real-world dialogues pose a great challenge to current summarization models, as the dialogue length typically ex
ceeds the input limits imposed by recent transformer-based pre-trained models, and the interactive nature of dialogues makes relevant information more context-dependent and sparsely distributed than news articles. In this work, we perform a comprehensive study on long dialogue summarization by investigating three strategies to deal with the lengthy input problem and locate relevant information: (1) extended transformer models such as Longformer, (2) retrieve-then-summarize pipeline models with several dialogue utterance retrieval methods, and (3) hierarchical dialogue encoding models such as HMNet. Our experimental results on three long dialogue datasets (QMSum, MediaSum, SummScreen) show that the retrieve-then-summarize pipeline models yield the best performance. We also demonstrate that the summary quality can be further improved with a stronger retrieval model and pretraining on proper external summarization datasets.
With the recent breakthrough of deep learning technologies, research on machine reading comprehension (MRC) has attracted much attention and found its versatile applications in many use cases. MRC is an important natural language processing (NLP) tas
k aiming to assess the ability of a machine to understand natural language expressions, which is typically operationalized by first asking questions based on a given text paragraph and then receiving machine-generated answers in accordance with the given context paragraph and questions. In this paper, we leverage two novel pretrained language models built on top of Bidirectional Encoder Representations from Transformers (BERT), namely BERT-wwm and MacBERT, to develop effective MRC methods. In addition, we also seek to investigate whether additional incorporation of the categorical information about a context paragraph can benefit MRC or not, which is achieved based on performing context paragraph clustering on the training dataset. On the other hand, an ensemble learning approach is proposed to harness the synergistic power of the aforementioned two BERT-based models so as to further promote MRC performance.
Recent advances in Unsupervised Neural Machine Translation (UNMT) has minimized the gap between supervised and unsupervised machine translation performance for closely related language-pairs. However and the situation is very different for distant la
nguage pairs. Lack of overlap in lexicon and low syntactic similarity such as between English and IndoAryan languages leads to poor translation quality in existing UNMT systems. In this paper and we show that initialising the embedding layer of UNMT models with cross-lingual embeddings leads to significant BLEU score improvements over existing UNMT models where the embedding layer weights are randomly initialized. Further and freezing the embedding layer weights leads to better gains compared to updating the embedding layer weights during training. We experimented using Masked Sequence to Sequence (MASS) and Denoising Autoencoder (DAE) UNMT approaches for three distant language pairs. The proposed cross-lingual embedding initialization yields BLEU score improvement of as much as ten times over the baseline for English-Hindi and English-Bengali and English-Gujarati. Our analysis shows that initialising embedding layer with static cross-lingual embedding mapping is essential for training of UNMT models for distant language-pairs.
Recent work has adopted models of pragmatic reasoning for the generation of informative language in, e.g., image captioning. We propose a simple but highly effective relaxation of fully rational decoding, based on an existing incremental and characte
r-level approach to pragmatically informative neural image captioning. We implement a mixed, fast' and slow', speaker that applies pragmatic reasoning occasionally (only word-initially), while unrolling the language model. In our evaluation, we find that increased informativeness through pragmatic decoding generally lowers quality and, somewhat counter-intuitively, increases repetitiveness in captions. Our mixed speaker, however, achieves a good balance between quality and informativeness.
The paper reports on a corpus study of German light verb constructions (LVCs). LVCs come in families which exemplify systematic interpretation patterns. The paper's aim is to account for the properties determining these patterns on the basis of a corpus study on German LVCs of the type stehen unter' NP' (stand under NP').
The use of automatic methods for the study of lexical semantic change (LSC) has led to the creation of evaluation benchmarks. Benchmark datasets, however, are intimately tied to the corpus used for their creation questioning their reliability as well
as the robustness of automatic methods. This contribution investigates these aspects showing the impact of unforeseen social and cultural dimensions. We also identify a set of additional issues (OCR quality, named entities) that impact the performance of the automatic methods, especially when used to discover LSC.
Named Entity Recognition is an essential task in natural language processing to detect entities and classify them into predetermined categories. An entity is a meaningful word, or phrase that refers to proper nouns. Named Entities play an important r
ole in different NLP tasks such as Information Extraction, Question Answering and Machine Translation. In Machine Translation, named entities often cause translation failures regardless of local context, affecting the output quality of translation. Annotating named entities is a time-consuming and expensive process especially for low-resource languages. One solution for this problem is to use word alignment methods in bilingual parallel corpora in which just one side has been annotated. The goal is to extract named entities in the target language by using the annotated corpus of the source language. In this paper, we compare the performance of two alignment methods, Grow-diag-final-and and Intersect Symmetrisation heuristics, to exploit the annotation projection of English-Brazilian Portuguese bilingual corpus to detect named entities in Brazilian Portuguese. A NER model that is trained on annotated data extracted from the alignment methods, is used to evaluate the performance of aligners. Experimental results show the Intersect Symmetrisation is able to achieve superior performance scores compared to the Grow-diag-final-and heuristic in Brazilian Portuguese.
