Naver Corporation

About: Naver Corporation is a company organization based out in Seongnam-si, South Korea. It is known for research contribution in the topics: Terminal (electronics) & Computer science. The organization has 4038 authors who have published 4294 publications receiving 35045 citations. The organization is also known as: NAVER Corporation & NAVER.

Anti-allergic activity of a platycodon root ethanol extract.

Abstract: Platycodon grandiflorum (Campanulaceae) is used as traditional medicine in Asian countries. In Korean traditional medicine, Platycodon root has been widely used since ancient times as a traditional drug to treat cold, cough and asthma. However, its effects on bone marrow-derived mast cell (BMMC)-mediated allergy and inflammation mechanisms remain unknown. In this study, the biological effect of Platycodon root ethanol extract (PE) was evaluated in BMMC after induction of allergic mediators by phorbol 12-myristate 13-acetate (PMA) plus calcium ionophore A23187 (A23187) stimulation. The effect of PE on the production of several allergic mediators, such as interleukin-6 (IL-6), prostaglandin D2 (PGD2), leukotriene C4 (LTC4), β-Hexosaminidase (β-Hex) and cyclooxygenase-2 (COX-2) protein, was investigated. The results demonstrate that PE inhibits PMA + A23187 induced production of IL-6, PGD2, LTC4, β-Hexosaminidase and COX-2 protein. Taken together, these results indicate that PE has the potential for use in the treatment of allergy.

SuperLoss: A Generic Loss for Robust Curriculum Learning

Abstract: Curriculum learning is a technique to improve a model performance and generalization based on the idea that easy samples should be presented before difficult ones during training. While it is generally complex to estimate a priori the difficulty of a given sample, recent works have shown that curriculum learning can be formulated dynamically in a self-supervised manner. The key idea is to somehow estimate the importance (or weight) of each sample directly during training based on the observation that easy and hard samples behave differently and can therefore be separated. However, these approaches are usually limited to a specific task (e.g., classification) and require extra data annotations, layers or parameters as well as a dedicated training procedure. We propose instead a simple and generic method that can be applied to a variety of losses and tasks without any change in the learning procedure. It consists in appending a novel loss function on top of any existing task loss, hence its name: the SuperLoss. Its main effect is to automatically downweight the contribution of samples with a large loss, i.e. hard samples, effectively mimicking the core principle of curriculum learning. As a side effect, we show that our loss prevents the memorization of noisy samples, making it possible to train from noisy data even with non-robust loss functions. Experimental results on image classification, regression, object detection and image retrieval demonstrate consistent gain, particularly in the presence of noise.

Look who's not talking.

Abstract: The objective of this work is speaker diarisation of speech recordings 'in the wild'. The ability to determine speech segments is a crucial part of diarisation systems, accounting for a large proportion of errors. In this paper, we present a simple but effective solution for speech activity detection based on the speaker embeddings. In particular, we discover that the norm of the speaker embedding is an extremely effective indicator of speech activity. The method does not require an independent model for speech activity detection, therefore allows speaker diarisation to be performed using a unified representation for both speaker modelling and speech activity detection. We perform a number of experiments on in-house and public datasets, in which our method outperforms popular baselines.

User authentication system using ip address and method thereof

Abstract: A method for authenticating a user by IP address check includes: receiving a URL and a session cookie from a client; determining whether or not an IP address of the client has been changed based on the session cookie; resetting the session cookie, if the IP address has been changed, by adding the changed IP address as a temporary IP address thereto; determining whether or not the URL is required to perform IP address check; requesting a re-login to the client if it is determined that the URL is required to perform IP address check; and adding the temporary IP address to a valid IP address list for the user if the re-login is successful.

System and Method for Filtering Message

Abstract: A system and method for message filtering are disclosed. A conversion target letter table is generated which stores information on letters that may be used to elude filtering, the table storing the hex code information of each letter that may be used for eluding filtering and the conversion letter information regarding the conversion letter to which each letter is to be converted. A hex code of a message text inputted by a user is compared with the hex codes of the table, and when a conversion target letter is included, it is converted to the conversion letter stored in the conversion target table associated with the letter. When the conversion is completed, it is determined whether or not pre-configured prohibited words are included in the converted message; and message filtering is performed.