Abstract: A model generating apparatus and method are provided. The apparatus receives a plurality of sample images. The apparatus generates a plurality of adversarial samples corresponding to the sample images. The apparatus inputs the sample images and the adversarial samples respectively to a first encoder and a second encoder in a self-supervised neural network to generate a plurality of first feature extractions and a plurality of second feature extractions. The apparatus calculates a similarity of each of the first feature extractions and the second feature extractions to train the self-supervised neural network. The apparatus generates a task model based on the first encoder and a plurality of labeled data.

Abstract: A model determination apparatus is configured to execute the following operations. Multiple candidate models are validated based on first adversarial validation data to generate a first accuracy corresponding to each of the candidate models, wherein the first adversarial validation data is generated by a first adversarial attack adjustment performed on validation data based on an initial model. A second adversarial attack adjustment is performed on the validation data based on each of the candidate models to generate a plurality of second adversarial validation data corresponding to each of the candidate models respectively. The candidate models are validated based on the corresponding second adversarial validation data to generate a second accuracy corresponding to each of the candidate models. At least one output model is selected from the candidate models based on the first accuracy and the second accuracy corresponding to each of the candidate models.

Abstract: A model generating apparatus and method are provided. The apparatus receives a plurality of sample images. The apparatus generates a plurality of adversarial samples corresponding to the sample images. The apparatus inputs the sample images and the adversarial samples respectively to a first encoder and a second encoder in a self-supervised neural network to generate a plurality of first feature extractions and a plurality of second feature extractions. The apparatus calculates a similarity of each of the first feature extractions and the second feature extractions to train the self-supervised neural network. The apparatus generates a task model based on the first encoder and a plurality of labeled data.

Abstract: A planar inverted-F antenna for use in a wireless network device comprises a connecting member and two radiators. The connecting member has at least one input end and at least one ground end. Each radiator has a first end portion perpendicularly connected to one of the two ends of the connecting member, and the two radiators are parallel and correspond in shape to each other. Each radiator further has an L-shaped notch and thus forms a barb. A second end portion of each radiator is bent to form an engaging end which is generally parallel to the connecting member and configured to fasten with a substrate of the wireless network device.

Abstract: A dual-band antenna for use in a wireless network device comprises first, second, and third radiators. The first and second radiators are connected by a stand portion. The second radiator is a generally C-shaped plate having a connecting section and a free-end portion. A ground end and an input end are provided at predetermined positions of the connecting section and are respectively and electrically connected to a grounding portion and a control circuit of a substrate. The free-end portion overlaps with the orthogonal projection of the first radiator and is parallel to the first radiator. The third radiator is electrically connected to the second radiator via a conductive post and is parallel to the second radiator. The second and third radiators are provided on the substrate while the first radiator is provided outside the substrate. The first, second, and third radiators are parallel to and spaced apart from one another.

Abstract: A planar inverted-F antenna for use in a wireless network device comprises a connecting member and two radiators. The connecting member has at least one input end and at least one ground end. Each radiator has a first end portion perpendicularly connected to one of the two ends of the connecting member, and the two radiators are parallel and correspond in shape to each other. Each radiator further has an L-shaped notch and thus forms a barb. A second end portion of each radiator is bent to form an engaging end which is generally parallel to the connecting member and configured to fasten with a substrate of the wireless network device.

Abstract: A dual-band antenna for use in a wireless network device comprises first, second, and third radiators. The first and second radiators are connected by a stand portion. The second radiator is a generally C-shaped plate having a connecting section and a free-end portion. A ground end and an input end are provided at predetermined positions of the connecting section and are respectively and electrically connected to a grounding portion and a control circuit of a substrate. The free-end portion overlaps with the orthogonal projection of the first radiator and is parallel to the first radiator. The third radiator is electrically connected to the second radiator via a conductive post and is parallel to the second radiator. The second and third radiators are provided on the substrate while the first radiator is provided outside the substrate. The first, second, and third radiators are parallel to and spaced apart from one another.

Abstract: A data security device and the method thereof are provided for transmission security via a USB port. A USB protocol analyzer of the data security device is provided for determining the signal type, so as to encrypt a storage data signal but not a command signal. Therefore, the part of the USB device not for storage (such as a mouse) can work normally. The data security device determines the file system format of the storage data signal by a file system analyzer and encrypts a data block of the storage data signal by an encrypt unit, thereby a filename of the encrypted storage data signal can be obtained and the USB device can be used on other hosts.