Abstract: A method for preventing forgery of data according to an embodiment includes determining a noise level based on metadata of original data, generating noise by applying the determined noise level to a preset noise pattern, generating transformed data of the original data by adding the generated noise to the original data, and transmitting the transformed data and the metadata to a server.

Abstract: There is provided an encryption processing method performed by an encryption processing apparatus. The encryption processing method comprises compressing data to obtain compressed data, determining, within the compressed data, a section to be encrypted and encrypting the section to obtain partially encrypted data.

Abstract: An apparatus for encryption according to an embodiment of the present disclosure comprises a classifier configured to classify each data included in an original data set into one of encryption target data and non-encryption target data on the basis of at least one of determination on whether or not an operation to be applied to each data included in the original data set for an analysis of the original data set is a preset operation and determination on whether or not each data included in the original data set is sensitive information; and an encryptor configured to encrypt the encryption target data among the data included in the original data set using a homomorphic encryption algorithm.

Abstract: Provided are an optimization system and method of a deep learning model. According to example embodiments, by optimizing the structure of the deep learning model appropriately for a target dataset without fixing the structure of the deep learning model, it is possible to generate a model structure capable of having high performance on the target dataset and also saving resources.

Abstract: A method of data encryption according to one embodiment includes generating a structured query language (SQL) syntax that includes manipulation target data and encryption policy identification information in the form of an annotation regarding the manipulation target data, receiving encryption policy information that corresponds to the encryption policy identification information in the form of an annotation from an encryption policy server, and converting the manipulation target data in the SQL syntax into cipher text on the basis of the encryption policy information.

Abstract: A content joint editing method according to an embodiment of the inventive concept may preserve the real-time property of update of edits as much as possible even if the number of editors increases when multiple editors simultaneously edit content. A content editing method can share and edit various contents without calling a separate external application in an online collaboration service. A content sharing method can manage each user's access rights to a content in individual content units in collaboration service that displays a plurality of contents at the same time in the collaboration space for sharing contents among multiple users.

Abstract: A few-shot learning method according to an embodiment may be performed in a computing device including one or more processors and a memory storing one or more programs executed by the one or more processors. The method may include performing a first task for subjecting a first model to few-shot learning (FSL) based on one or more meta-training data and performing a second task for subjecting a second model to supervised learning based on one or more derived data modified from the one or more meta-training data.

Abstract: A method for performing an operation according to an embodiment includes assigning an error value to encrypted data; and performing a homomorphic operation for an approximation function, which is obtained by approximating a target function, by using, as an input value, the encrypted data to which the error value is assigned.

Abstract: A content scheduling apparatus includes a content storage storing contents, a screen division information storage storing screen division information including a screen division form having division regions and time slot information including information about time slots allocated to each of the division regions, a content information storage storing content information including content list information about a list in which contents are listed for each of the division regions among the contents and content display order information including display orders each of which includes a display order of the contents listed foreach of the division regions, and a schedule generator to allocate the contents listed for each of the division regions to the time slots allocated to each of the division regions according to the content display order information so that the contents listed for each of the division regions are sequentially displayed on each of the division regions.

Abstract: An apparatus for congestion visualization according to an embodiment of the present invention includes an area-specific image storage configured to store area-specific images captured by a plurality of cameras which capture different regions in a specific space, respectively, a two-dimensional map storage configured to store a two-dimensional map for the specific space, an area-specific congestion calculator configured to calculate, based on the area-specific images, an area-specific congestion for each of capturing areas captured by the plurality of cameras in the specific space, a mapping area identifier configured to identify a mapping area corresponding to each of the capturing areas on the two-dimensional map, based on a mapping point set in each of the two-dimensional map and the area-specific images, and an area-specific congestion visualizer configured to visually displaying, on the two-dimensional map, the area-specific congestion for each of the capturing areas, based on the mapping area.

Abstract: A method and apparatus for correcting missing values in data are provided. A method of correcting missing values in basic data according to an embodiment includes a data extraction step, a prediction model configuration step, a first correction step, and a second correction step. The method corrects missing values in data by repeating the steps of generating a prediction model for correcting the missing value and correcting the missing value with the use of the prediction model.

Abstract: A content scheduling apparatus includes a content storage storing contents, a screen division information storage storing screen division information including a screen division form having division regions and time slot information including information about time slots allocated to each of the division regions, a content information storage storing content information including content list information about a list in which contents are listed for each of the division regions among the contents and content display order information including display orders each of which includes a display order of the contents listed foreach of the division regions, and a schedule generator to allocate the contents listed for each of the division regions to the time slots allocated to each of the division regions according to the content display order information so that the contents listed for each of the division regions are sequentially displayed on each of the division regions.

Abstract: A method of embedding a vector representing an arrangement state of objects in a 3D space according to an embodiment includes dividing a target space into M×N subspaces extending from a first surface among planes forming the target space having a cuboid shape, calculating, for each of the subspaces, a feature representing an object arrangement state in the subspace, inputting data representing the feature of each of the subspaces into an artificial neural network and obtaining, from the artificial neural network, an embedding vector representing an arrangement state of an object in the target space.

Abstract: A hyperparameter optimization method performed by a hyperparameter optimization apparatus to optimize hyperparameters of a model includes calculating an evaluation score for each of a plurality of hyperparameter samples constituting a first hyperparameter sample set by applying each of the hyperparameter samples to a target model, performing weighted clustering of the hyperparameter samples by using the calculated evaluation scores as weights and constructing a second hyperparameter sample set based on the clustering result.

Abstract: An image segmentation model training apparatus according to a disclosed embodiment includes a predictor, which generates a plurality of original masks using an unlabeled original image set and a pre-trained image segmentation model, a label generator, which generates a synthesized image set based on the original image set and the plurality of original masks, and generates a plurality of pseudo labels based on the synthesized image set, a pre-processor, which generates a training image set by performing pseudo labeling on the original image set and the synthesized image set using the plurality of pseudo labels, and a model trainer, which further trains the image segmentation model based on the training image set.

Abstract: A data management system according to an embodiment of this invention includes a blockchain node configured to store on-chain data on blockchain and a service server, upon receiving a service request associated with the original data from a client, i) accessing off-chain data stored in a storage node or recording off-chain data in the storage node according to the service request, and ii) sending a transaction proposal to the blockchain node for a record associated with the service request to be stored in the blockchain.

Abstract: An apparatus for set intersection operation according to an embodiment includes a ciphertext acquisition unit configured to acquire a ciphertext for a first vector corresponding to a first subset of a universal set including a plurality of elements from an encryption apparatus, a transform unit configured to generate a second vector corresponding to a second subset of the universal set, a computation unit configured to generate a ciphertext for a third vector corresponding to an intersection of the first subset and the second subset, based on the ciphertext for the first vector and the second vector, and a ciphertext providing unit configured to provides the ciphertext for the third vector to the encryption apparatus.

Abstract: An edge computing method for flexibly allocating a computing resource includes receiving an edge module and a container resource allocation value, generating a container based on the container resource allocation value and arranging the edge module in the generated container. The edge module is distributed to an edge computing device to perform a predetermined data processing. The container is a software component of the edge computing device providing an isolated computing environment for driving the edge module. The edge computing device is installed on-premise. The edge computing device can flexibly adjust available computing resources according to the current working situation and a resource requirement. In response to an environment, in which data traffic is flexible according to time or situation, such as a manufacturing plant, allocated resources of the edge computing device can be quickly adjusted, thereby a stable service can be provided.

Abstract: An apparatus for guaranteeing the integrity of a state database includes a memory for storing one or more instructions, a communication interface for communicating with one or more blockchain nodes and a processor. The processor is configured, by executing the stored one or more instructions, to receive a request to query for a state value, obtain a first state value by querying a state database in response to the request, forward the request to the one or more blockchain nodes to receive a second state value from the one or more blockchain nodes and compare the first state value with the second state value to determine integrity of a state value.

Abstract: A method for path routing according to an embodiment of the present disclosure may include selecting a first start point and a first end point with which path routing is performed in a circular frame generated by connecting all points included in one or more start point sets included in a layer, one or more end point sets paired with the start point set, and one or more edge point pair sets to one closed curve, generating a connectivity graph by connecting edge points included in one or more nodes corresponding to segments obtained by dividing the circular frame into one or more regions, and connecting the first start point and the first end point based on a cost for connecting the first start point and the first end point calculated using the connectivity graph.