539 research outputs found
Sort by
ELITR Demo at ECSPM Symposium
No full textI introduced the EU project ELITR, its goals and the current stage, and demoed it in practice
CUNI Non-Autoregressive System for the WMT 22 Efficient Translation Shared Task
No full textWe present a non-autoregressive system submission to the WMT 22 Efficient Translation Shared Task. Our system was used by Helcl et al. (2022)
in an attempt to provide fair comparison between non-autoregressive and autoregressive models. This submission is an effort to
establish solid baselines along with sound evaluation methodology,
particularly in terms of measuring the decoding speed. The model itself
is a 12-layer Transformer model trained with connectionist temporal classification on knowledge-distilled dataset by a strong autoregressive teacher model
Improving Both Domain Robustness and Domain Adaptability in Machine Translation
No full textWe consider two problems of NMT domain adaptation using meta-learning. First, we want to reach domain robustness, i.e., we want to reach high quality on both domains seen in the training data and unseen domains. Second, we want our systems to be adaptive, i.e., making it possible to finetune systems with just hundreds of in-domain parallel sentences. We study the domain adaptability of meta-learning when improving the domain robustness of the model. In this paper, we propose a novel approach, RMLNMT (Robust Meta-Learning Framework for Neural Machine Translation Domain Adaptation), which improves the robustness of existing meta-learning models. More specifically, we show how to use a domain classifier in curriculum learning and we integrate the word-level domain mixing model into the meta-learning framework with a balanced sampling strategy. Experiments on English-German and English-Chinese translation show that RMLNMT improves in terms of both domain robustness and domain adaptability in seen and unseen domains
Findings of the 2022 Conference on Machine Translation (WMT22)
No full textThe article presents the results of WMT22 translation shared tasks
Two Reproductions of a Human-Assessed Comparative Evaluation of a Semantic Error Detection System
No full textIn this paper, we present the results of two reproduction studies for the human evaluation originally reported by Dušek and Kasner (2020) in which the authors comparatively evaluated outputs produced by a semantic error detection system for data-to-text generation against reference outputs. In the first reproduction, the original evaluators repeat the evaluation, in a test of the repeatability of the original evaluation. In the second study, two new evaluators carry out the evaluation task, in a test of the reproducibility of the original evaluation under otherwise identical conditions. We describe our approach to reproduction, and present and analyse results, finding different degrees of reproducibility depending on result type, data and labelling task. Our resources are available and open-sourced
Making a Semantic Event-type Ontology Multilingual
No full textWe present an extension of the SynSemClass event-type ontology, originally conceived as a bilingual Czech-English resource. We added German entries to the classes representing the concepts of the ontology. Having a different starting point than the original work (unannotated parallel corpus without links to a valency lexicon and, of course, different existing lexical resources), it was a challenge to adapt the annotation guidelines, the data model and the tools used for the original version. We describe the process and results of working in such a setup. We also show the next steps to adapt the annotation process, data structures and formats and tools necessary to make the addition of a new language in the future more smooth and efficient, and possibly to allow for various teams to work on SynSemClass extensions to many languages concurrently. We also present the latest release which contains the results of adding German, freely available for download as well as for online access
Short-Term Word-Learning in a Dynamically Changing Environment
No full textNeural sequence-to-sequence automatic speech recognition (ASR) systems are in principle open vocabulary systems, when using appropriate modeling units. In practice, however, they often fail to recognize words not seen during training, e.g., named entities, numbers or technical terms. To alleviate this problem, Huber et al. proposed to supplement an end-to-end ASR system with a word/phrase memory and a mechanism to access this memory to recognize the words and phrases correctly. In this paper we study, a) methods to acquire important words for this memory dynamically and, b) the trade-off between improvement in recognition accuracy of new words and the potential danger of false alarms for those added words. We demonstrate significant improvements in the detection rate of new words with only a minor increase in false alarms (F1 score 0.30 → 0.80), when using an appropriate number of new words. In addition, we show that important keywords can be extracted from supporting documents and used effectively
AI Technologies for Machine Supervision and Help in a Rehabilitation Scenario
No full textWe consider, evaluate, and develop methods for home rehabilitation scenarios. We show the required modules for this scenario. Due to the large number of modules, the framework falls into the category of Composite AI. Our work is based on collected videos with high-quality execution and samples of typical errors. They are augmented by sample dialogues about the exercise to be executed and the assumed errors. We study and discuss body pose estimation technology, dialogue systems of different kinds and the emerging constraints of verbal communication. We demonstrate that the optimization of the camera and the body pose allows high-precision recording and requires the following components: (1) optimization needs a 3D representation of the environment, (2) a navigation dialogue to guide the patient to the optimal pose, (3) semantic and instance maps are necessary for verbal instructions about the navigation. We put forth different communication methods, from video-based presentation to chit-chat-like dialogues through rule-based methods. We discuss the methods for different aspects of the challenges that can improve the performance of the individual components. Due to the emerging solutions, we claim that the range of applications will drastically grow in the very near future
Neural Pipeline for Zero-Shot Data-to-Text Generation
No full textIn data-to-text (D2T) generation, training on in-domain data leads to overfitting to the data representation and repeating training data noise. We examine how to avoid finetuning pretrained language models (PLMs) on D2T generation datasets while still taking advantage of surface realization capabilities of PLMs. Inspired by pipeline approaches, we propose to generate text by transforming single-item descriptions with a sequence of modules trained on general-domain text-based operations: ordering, aggregation, and paragraph compression. We train PLMs for performing these operations on a synthetic corpus WikiFluent which we build from English Wikipedia. Our experiments on two major triple-to-text datasets — WebNLG and E2E — show that our approach enables D2T generation from RDF triples in zero-shot settings
Large Neural Language Models for Data-to-text Generation
No full textCurrent research state-of-the-art in automatic data-to-text generation, a major task in natural language generation, is dominated by large language models based on the Transformer neural network architecture. These models are capable of producing lifelike, natural texts; however, they are hard to control and often do not adhere to the input data, omitting important content or producing "hallucinated" text which is not grounded in the input data. In this talk, I will first show the basic operation principles of the large language models. I will then detail our experiments aiming at higher accuracy of generated text in two ways: (1) improving accuracy of the generating language models themselves, (2) automatically detecting errors in generated texts