Abstract: A method for training an image classification model according to an embodiment includes training a feature extractor and a rotation angle classifier to predict a rotation angle of each of unlabeled first training images, training the image classification model to predict a label and rotation angle of each of labeled second training images, but predict a uniform label even though an actual rotation angle of each of the second training images is changed, generating a pseudo label based on a training image that satisfy a preset condition among unlabeled candidate images, and training the image classification model to predict a rotation angle of each of the third training images, and predict a label of each of the third training images based on the pseudo label, but predict a uniform label even though an actual rotation angle of each of the third training images is changed.

Abstract: A method for solving a problem and a system thereof are provided. The method according to some embodiments includes setting at least one current search node on a search tree corresponding to a solution space of a target problem; selecting candidate search nodes from among child nodes of the at least one current search node, a number of the candidate search nodes being equal to a number of items inferred by a machine-trained model; determining at least one next search node from among the candidate search nodes based on results of search simulation for the candidate search nodes; and determining a solution to the target problem based on a result of a search using the at least one next search node.

Abstract: Methods and apparatuses for accurately determining a model, which is to be the basis of transfer learning, among a plurality of source models, are provided. According to an embodiment, an apparatus for determining a base model to be used for transfer learning to a target domain is provided. The apparatus comprises a memory which comprises one or more instructions and a processor which executes the instructions to construct a neural network model for measuring suitability of a plurality of pre-trained source models, measure the suitability of each of the source models by inputting data of the target domain to the neural network model, and determine the base model to be used for the transfer learning among the source models based on the suitability.

Abstract: An apparatus includes a modified image generator generating modified images by modifying each unlabeled image, a pre-trainer to generate a feature vector for each modified image by using an artificial neural network-based encoder and train the encoder based on the feature vector for each modified image, a pseudo-label generator to generate a feature vector for each unlabeled training image, cluster the training images based on the feature vector for each training image, and generate a pseudo-label for at least one training image among the training images based on the clustering result, and a further trainer to generate a predicted label by using the trained encoder and a classification model including a classifier to generate a predicted label for an image input to the trained encoder based on a feature vector, and train the classification model based on the pseudo-label and predicted label for the at least one training image.

Abstract: An apparatus for encoding matrix data according to an embodiment includes an initial encoding layer that outputs a plurality of encoded row vectors and a plurality of encoded column vectors for one or more pieces of matrix data representing information on a relationship between items having a plurality of characteristics based on the matrix data, a plurality of initial row vectors corresponding to the number of rows of the matrix data, and a plurality of initial column vectors corresponding to the number of columns of the matrix data, and one or more encoding layers that are disposed after the initial encoding layer and perform additional encoding on the plurality of encoded row vectors and the plurality of encoded column vectors.

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: An apparatus includes a modified image generator generating modified images by modifying each unlabeled image, a pre-trainer to generate a feature vector for each modified image by using an artificial neural network-based encoder and train the encoder based on the feature vector for each modified image, a pseudo-label generator to generate a feature vector for each unlabeled training image, cluster the training images based on the feature vector for each training image, and generate a pseudo-label for at least one training image among the training images based on the clustering result, and a further trainer to generate a predicted label by using the trained encoder and a classification model including a classifier to generate a predicted label for an image input to the trained encoder based on a feature vector, and train the classification model based on the pseudo-label and predicted label for the at least one training image.

Abstract: A method for training an image classification model according to an embodiment includes training a feature extractor and a rotation angle classifier to predict a rotation angle of each of unlabeled first training images, training the image classification model to predict a label and rotation angle of each of labeled second training images, but predict a uniform label even though an actual rotation angle of each of the second training images is changed, generating a pseudo label based on a training image that satisfy a preset condition among unlabeled candidate images, and training the image classification model to predict a rotation angle of each of the third training images, and predict a label of each of the third training images based on the pseudo label, but predict a uniform label even though an actual rotation angle of each of the third training images is changed.

Abstract: An apparatus for training a reinforcement learning model according to an embodiment includes a starting point determinator configured to determine starting points from an input value of a combinatorial optimization problem, a multi-explorer configured to generate exploration trajectories by performing exploration from each of the starting points using a reinforcement learning model, a trajectory evaluator configured to calculate an evaluation value of each of the exploration trajectories using an evaluation function of the combinatorial optimization problem, a baseline calculator configured to calculate a baseline for the input value from the evaluation value of each exploration trajectory, an advantage calculator configured to calculate an advantage of each of the exploration trajectories using the evaluation value of each exploration trajectory and the baseline, and a parameter updater configured to update parameters of the reinforcement learning model by using the exploration trajectories and the advantage

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: Provided are extracellular vesicles (EVs) derived from a recombinant microorganism including one or more polynucleotides encoding one or more target proteins, extracellular vesicles isolated from the microorganism, and a use of the extracellular vesicles.

Abstract: Methods and apparatuses for accurately determining a model, which is to be the basis of transfer learning, among a plurality of source models, are provided. According to an embodiment, an apparatus for determining a base model to be used for transfer learning to a target domain is provided. The apparatus comprises a memory which comprises one or more instructions and a processor which executes the instructions to construct a neural network model for measuring suitability of a plurality of pre-trained source models, measure the suitability of each of the source models by inputting data of the target domain to the neural network model, and determine the base model to be used for the transfer learning among the source models based on the suitability.

Abstract: There is provided a method of masking an object of non-interest by a masking apparatus. The method comprises acquiring first video information of a region of interest, determining a first object of non-interest area as an area in which motion vectors are present in the first video information of the region of interest, removing spatial noise from the first object of non-interest area to acquire a second object of non-interest area including at least a part of the first object of non-interest area and generating a mask corresponding to the second object of non-interest area.

Abstract: A method includes acquiring, by a device, encoded image data corresponding to an original image. The method includes decoding, by the device, the encoded image data. The method included acquiring, by the device, a foreground extraction target frame and an encoding parameter associated with an encoding process of the original image based on decoding the encoded image data. The method includes extracting, by the device, a first candidate foreground associated with the foreground extraction target frame based on the encoding parameter. The method includes extracting, by the device, a second candidate foreground associated with the foreground extraction target frame based on a preset image processing algorithm. The method includes determining, by the device, a final foreground associated with the foreground extraction target frame based on the first candidate foreground and the second candidate foreground.

Abstract: Disclosed is a method for reducing noise. The method comprises decoding a bitstream encoded with a first codec, acquiring Quantization Parameter (QP) values determined in a process of encoding the bitstream with the first codec, detecting a frame with noise by using the QP values and inputting the frame with noise to a predefined noise filter.

Abstract: A method for selecting an image for object recognition is provided. The method comprises receiving an image bitstream; acquiring predetermined first codec metadata information among codec metadata information from the received image bitstream; calculating a first quality measurement value using the acquired first codec metadata information; calculating a quality score of the image by using the calculated first quality measurement value; and selecting a predetermined number of images based on the calculated quality score of the image.

Abstract: Disclosed is a method for reducing noise. The method comprises decoding a bitstream encoded with a first codec, acquiring Quantization Parameter (QP) values determined in a process of encoding the bitstream with the first codec, detecting a frame with noise by using the QP values and inputting the frame with noise to a predefined noise filter.

Abstract: Provided is a cassette for deleting a target gene comprising (a) a promoter-specific homologous region having a sequence identity to a portion of a promoter region of the target gene, wherein the degree of sequence identity is sufficient to drive homologous recombination therebetween, (b) a marker gene operably linked to the promoter-specific homologous region, and (c) a gene-specific homologous region adjacent to 3?-end of the marker gene and having a sequence identity to at least a portion of the target gene, wherein the degree of sequence identity is sufficient to drive homologous recombination therebetween.

Abstract: A display device includes a plurality of reception units receiving a plurality of content, a storage unit, a plurality of scaler units reducing data sizes of the plurality of content, storing the respective content with the reduced data sizes in the storage unit, and reading the respective content stored in the storage unit according to an output timing, a plurality of frame rate conversion units converting frame rates of the respective read content, and a video output unit combining and displaying the respective content output from the plurality of frame rate conversion units. Accordingly, the resources can be minimized.

Abstract: A display apparatus which can simultaneously display a plurality of images such as a two-dimensional image or a three-dimensional image, and an image display method therein, the display apparatus including: a signal receiving part which receives a first image signal; an image processing part which generates a second image signal and a third image signal based on the first image signal; a display part which displays images based on the second and third image signals; and a controller which controls the image processing part to generate the second and third image signals, to display the image based on the second image signal in a first region of the display part, and to display the image based on the third image signal in a second region of the display part, the third image signal including a left-eye image and a right-eye image generated from the first image signal.