Abstract:
Disclosed is an operating method of an apparatus operated by at least one processor, the operating method including: fusing at least some feature maps generated by layers of a target model performing online continual learning to generate a fused feature map; identifying features with high attention in the fused feature map as shortcut features in a ratio based on a drop intensity; and removing the shortcut features from a target feature map output from a predetermined layer of the target model and input to a next layer.
Abstract:
Provided is a method of providing a location-based social search service. The method includes collecting, from a social network system, one or more pieces of visit history information including at least one of visit information of social search service users in association with at least a place and review data associated with the place, and storing the collected information; selecting a topic for each topic category from the visit history information; when a location-based query is received from at least one of the users, extracting a topic for each topic category from the location-based query; analyzing, for each topic category, a relationship between the topic extracted from the location-based query and the topic selected from the visit history information, and selecting at least one of the users as an expert.
Abstract:
A method of detecting an overlapping community in a network including nodes and links between the nodes, includes calculating a similarity between the links, and generating a line graph of the network. The method further includes detecting one or more cores in the line graph, and growing a cluster for each of the one or more cores. The method further includes converting the cluster into a cluster of nodes of a node graph.
Abstract:
A method for operating an apparatus for predicting confirmed cases of an infectious disease is provided. The method comprises predicting infectious disease information per country, including an infection risk per country, expected number of entrants per country, and number of imported cases per country, based on collected epidemic statistics data per country and inflow data between a corresponding country and a destination country, grouping two or more countries based on geographic or economic relevance, and correcting the infectious disease information per country of countries within a grouped group according to a contagion risk impact set depending on a correlation between the countries within the group, and predicting total number of imported cases flowing into the destination country by re-correcting the infectious disease information per country through applying a correlation for the confirmed cases of the infectious disease between groups to the infectious disease information per country.
Abstract:
An outlier detection device sets a weight for a kernel center of a grid cell based on a distribution of the data disposed on the grid cell region, calculates a cumulative change of a weight for each corresponding kernel center, sets a stationary region in the grid cell region based on the cumulative change, maintains a density of a kernel center of the stationary region as a previous density, calculates a density of a kernel center excluding the stationary region to update the calculated density, estimates a density of multidimensional data at the current time, and detects an arbitrary number of outliers based on a relative difference between the density of the multidimensional data and a density of a kernel center nearest to the multidimensional data.