Персона: Сорока, Артем Александрович
Email Address
Birth Date
Результаты поиска
Cross-Modal Transfer Learning for Image and Sound
© 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.Recently the research on transfer learning between similar domains has become increasingly common. However, the fields of cross-domain and cross-modal knowledge transfers are more complicated and have been studied less. We propose the new transfer learning strategy between tasks on essentially different domains called as cross-modal transfer learning and consider its ideas and the algorithm. The key element of cross-modal transfer pipeline is cross-modal adapter, i.e. a neural network that transforms the target domain features to the source domain features that can be efficiently processed by a pre-trained neural network. In the experiments the dataset ImageNet and audio dataset ESC-50 are chosen as source domain and target domain respectively. It is shown that a fairly simple neural cross-modal adapter makes it possible to achieve high classification accuracy on target domain using the knowledge obtained by pre-trained neural network on the source domain. Our experiments also show that cross-modal transfer learning noticeably reduces the training time in comparison with the building target model “from scratch”.
Estimating the Transfer Learning Ability of a Deep Neural Networks by Means of Representations
Explaining the Transfer Learning Ability of a Deep Neural Networks by Means of Representations
Russian Language Speech Generation from Facial Video Recordings Using Variational Autoencoder
Pre-Pandemic Cross-Reactive Immunity against SARS-CoV-2 among Siberian Populations
In December 2019, a new coronavirus, SARS-CoV-2, was found to in Wuhan, China. Cases of infection were subsequently detected in other countries in a short period of time, resulting in the declaration of the COVID-19 pandemic by the World Health Organization (WHO) on 11 March 2020. Questions about the impact of herd immunity of pre-existing immune reactivity to SARS-CoV-2 on COVID-19 severity, associated with the immunity to seasonal manifestation, are still to be resolved and may be useful for understanding some processes that precede the emergence of a pandemic virus. Perhaps this will contribute to understanding some of the processes that precede the emergence of a pandemic virus. We assessed the specificity and virus-neutralizing capacity of antibodies reacting with the nucleocapsid and spike proteins of SARS-CoV-2 in a set of serum samples collected in October and November 2019, before the first COVID-19 cases were documented in this region. Blood serum samples from 799 residents of several regions of Siberia, Russia, (the Altai Territory, Irkutsk, Kemerovo and Novosibirsk regions, the Republic of Altai, Buryatia, and Khakassia) were analyzed. Sera of non-infected donors were collected within a study of seasonal influenza in the Russian Federation. The sample collection sites were located near the flyways and breeding grounds of wild waterfowl. The performance of enzyme-linked immunosorbent assay (ELISA) for the collected sera included the usage of recombinant SARS-CoV-2 protein antigens: full-length nucleocapsid protein (CoVN), receptor binding domain (RBD) of S-protein and infection fragment of the S protein (S5-6). There were 183 (22.9%) sera reactive to the S5-6, 270 (33.8%) sera corresponding to the full-length N protein and 128 (16.2%) sera simultaneously reactive to both these proteins. Only 5 out of 799 sera had IgG antibodies reactive to the RBD. None of the sera exhibited neutralizing activity against the nCoV/Victoria/1/2020 SARS-CoV-2 strain in Vero E6 cell culture. The data obtained in this study suggest that some of the population of the analyzed regions of Russia had cross-reactive humoral immunity against SARS-CoV-2 before the COVID-19 pandemic started. Moreover, among individuals from relatively isolated regions, there were significantly fewer reliably cross-reactive sera. The possible significance of these data and impact of cross-immunity to SARS-CoV-2 on the prevalence and mortality of COVID-19 needs further assessment.