Hwidong Na

Samsung Advanced Institute of Technology · Software Solution Lab.

In this paper, we investigate the problem of string-based molecular generation via variational autoencoders (VAEs) that have served a popular generative approach for various tasks in artificial intelligence. We propose a simple, yet effective idea to improve the performance of VAE for the task. Our main idea is to maintain multiple decoders while s...

SpeechBrain is an open-source and all-in-one speech toolkit. It is designed to facilitate the research and development of neural speech processing technologies by being simple, flexible, user-friendly, and well-documented. This paper describes the core architecture designed to support several tasks of common interest, allowing users to naturally co...

Learning robust speaker embeddings is a crucial step in speaker diarization. Deep neural networks can accurately capture speaker discriminative characteristics and popular deep embeddings such as x-vectors are nowadays a fundamental component of modern diarization systems. Recently, some improvements over the standard TDNN architecture used for x-v...

Model-agnostic meta-learning (MAML) is a popular method for few-shot learning but assumes that we have access to the meta-training set. In practice, training on the meta-training set may not always be an option due to data privacy concerns, intellectual property issues, or merely lack of computing resources. In this paper, we consider the novel pro...

The Inversion Transduction Grammar (ITG) constraints have been widely used for word reordering in machine translation studies. They are, however, so restricted that some types of word reordering cannot be handled properly. We analyze three corpora between SVO and SOV languages: Chinese-Korean, English- Japanese, and English-Korean. In our analysis,...

This short paper presents a system description for intrinsic evaluation of the WMT 14’s medical text translation task. Our systems consist of phrase-based statistical machine translation system and query translation system between German-English language pairs. Our work focuses on the query translation task and we achieved the highest BLEU score am...

Translation requires non-isomorphic transformation from the source to the target. However, non-isomorphism can be reduced by learning multi-word units (MWUs). We present a novel way of representating sentence structure based on MWUs, which are not necessarily continuous word sequences. Our proposed method builds a simpler structure of MWUs than wor...

In this paper, we give a summary of various dependency chart parsing algorithms in terms of the use of parsing histories for a new dependency arc decision. Some parsing histories are closely related to the target dependency arc, and it is necessary for the parsing algorithm to take them into consideration. Each dependency treebank may have some uni...

We introduce a novel translation rule that cap-tures discontinuous, partial constituent, and non-projective phrases from source language. Using the traversal order sequences of the de-pendency tree, our proposed method 1) ex-tracts the synchronous rules in linear time and 2) combines them efficiently using the CYK chart parsing algorithm. We analyt...

We propose a reordering method to improve the fluency of the output of the phrase-based SMT (PBSMT) system. We parse the transla-tion results that follow the source language or-der into non-projective dependency trees, then reorder dependency trees to obtain fluent tar-get sentences. Our method ensures that the translation results are grammatically...