This paper presents the participation of DBMS-KU Interpolation system in WMT19 shared task, namely, Kazakh-English language pair. We examine the use of interpolation method using a different language model order. Our Interpolation system combines a direct translation with Russian as a pivot language. We use 3-gram and 5-gram language model orders to perform the language translation in this work. To reduce noise in the pivot translation process, we prune the phrase table of source-pivot and pivot-target. Our experimental results show that our Interpolation system outperforms the Baseline in terms of BLEU-cased score by +0.5 and +0.1 points in Kazakh-English and English-Kazakh, respectively. In particular, using the 5-gram language model order in our system could obtain better BLEU-cased score than utilizing the 3-gram one. Interestingly, we found that by employing the Interpolation system could reduce the perplexity score of English-Kazakh when using 3-gram language model order.
Hyperparameters and architecture greatly influence the performance of convolutional neural networks (CNNs); therefore, their optimization is important to obtain the desired results. One of the state-ofthe-art methods to achieve this is the use of neuroevolution that utilizes a genetic algorithm (GA) to optimize a CNN. However, the GA is often trapped into a local optimum resulting in premature convergence. In this study, we propose an approach called the "diversity-guided genetic algorithm-convolutional neural network (DGGA-CNN)" that uses adaptive parameter control and random injection to facilitate the search process by exploration and exploitation while preserving the population diversity. The alternation between exploration and exploitation is guided by using an average pairwise Hamming distance. Moreover, the DGGA fully handles the architecture of the CNN by using a novel finite state machine (FSM) combined with three novel mutation mechanisms that are specifically created for architecture chromosomes. Tests conducted on suggestion mining and twitter airline datasets reveal that the DGGA-CNN performs well with valid architectures and a comparison with other methods demonstrates its capability and efficiency.
This paper describes Kata.ai's submission for the Social Media Mining for Health (SMM4H) 2021 shared task. We participated in three tasks: classifying adverse drug effect, COVID-19 self-report, and COVID-19 symptoms. Our system is based on BERT model pre-trained on the domain-specific text. In addition, we perform data cleaning and augmentation, as well as hyperparameter optimization and model ensemble to further boost the BERT performance. We achieved the first rank in both classifying adverse drug effects and COVID-19 selfreport tasks.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.