Abstract: Provided are improved end-to-end self-supervised pre-training frameworks that leverage a combination of contrastive and masked modeling loss terms. In particular, the present disclosure provides framework that combines contrastive learning and masked modeling, where the former trains the model to discretize input data (e.g., continuous signals such as continuous speech signals) into a finite set of discriminative tokens, and the latter trains the model to learn contextualized representations via solving a masked prediction task consuming the discretized tokens. In contrast to certain existing masked modeling-based pre-training frameworks which rely on an iterative re-clustering and re-training process or other existing frameworks which concatenate two separately trained modules, the proposed framework can enable a model to be optimized in an end-to-end fashion by solving the two self-supervised tasks (the contrastive task and masked modeling) simultaneously.

Abstract: Classification techniques are disclosed that take into account the “cost” of each type of classification error for minimizing total cost of errors. In one example embodiment, a pre-trained cost-sensitive auto-encoder can be used in combination with a training (fine-tuning) stage for cost-sensitive deep learning. Thus, cost information is effectively combined with deep learning by modifying the objective function in the pre-training phase. By minimizing the modified objective function, the auto-encoder not only tries to capture underlying pattern, it further “learns” the cost information and “stores” it in the structure. By later fine-tuning at the training stage, the classification system yields improved performance (lower cost) than a typical classification system that does not take cost information into account during pre-training.

Abstract: An electronic device having a judging system to facilitate an operation of the electronic device. The judging system includes a searching module for searching MAC addresses of WiFi devices in a presently working environment. A storing module stores the MAC addresses in a data base. A comparing module judges whether the present working environment is a previously encountered environment by comparing the MAC addresses for the present working environment with MAC addresses for the previously encountered environment which have been stored in the data base. A processing module has the electronic device to perform a predetermined function if the present working environment is judged as the previously encountered environment.

Abstract: Classification techniques are disclosed that take into account the “cost” of each type of classification error for minimizing total cost of errors. In one example embodiment, a pre-trained cost-sensitive auto-encoder can be used in combination with a training (fine-tuning) stage for cost-sensitive deep learning. Thus, cost information is effectively combined with deep learning by modifying the objective function in the pre-training phase. By minimizing the modified objective function, the auto-encoder not only tries to capture underlying pattern, it further “learns” the cost information and “stores” it in the structure. By later fine-tuning at the training stage, the classification system yields improved performance (lower cost) than a typical classification system that does not take cost information into account during pre-training.

Abstract: In a method for unlocking an object of an electronic device, an unlocking rule is preset to unlock the object. After movement signals are received from a touch screen of the electronic device, a quantity of segments displayed on the touch screen generated by the movement signals and a quantity of areas of the touch screen partitioned by the obtained segments are obtained. When a touch signal on one area of the obtained areas is received, a sequence number of the area from which the touch signal is received is calculated. The object of the electronic device is unlocked when the object is authorized to be unlocked according to the preset unlocking rule, the quantity of the obtained segments, the quantity of the obtained areas, and the sequence number of the area from which the touch signal is received.

Abstract: An electronic device receives data packets from wireless devices. The electronic device parses the data packets of each wireless device to obtain position information of each wireless device. The electronic device calculates a position of the electronic device according to the position information of each wireless device.

Abstract: An electronic device having a judging system to facilitate an operation of the electronic device. The judging system includes a searching module for searching MAC addresses of WiFi devices in a presently working environment. A storing module stores the MAC addresses in a data base. A comparing module judges whether the present working environment is a previously encountered environment by comparing the MAC addresses for the present working environment with MAC addresses for the previously encountered environment which have been stored in the data base. A processing module has the electronic device to perform a predetermined function if the present working environment is judged as the previously encountered environment.

Abstract: In a method for unlocking an object of an electronic device, an unlocking rule is preset to unlock the object. After movement signals are received from a touch screen of the electronic device, a quantity of segments displayed on the touch screen generated by the movement signals and a quantity of areas of the touch screen partitioned by the obtained segments are obtained. When a touch signal on one area of the obtained areas is received, a sequence number of the area from which the touch signal is received is calculated. The object of the electronic device is unlocked when the object is authorized to be unlocked according to the preset unlocking rule, the quantity of the obtained segments, the quantity of the obtained areas, and the sequence number of the area from which the touch signal is received.

Abstract: In a method for monitoring a second mobile device using a first mobile device and wireless communication, a prompting file, a time interval, and a name of a second communication module of the second mobile device are set in the first mobile device. A first communication module of the first mobile device searches for discoverable devices within a predetermined coverage area of the first mobile device. Each found discoverable device is entered in a device list of the first mobile device. If the second communication module is found, a timer starts timing the time interval. After the time interval has elapsed, the first communication module searches for the discoverable devices again, for updating the device list. If the second communication module is not in the updated device list, the prompting file is outputted to alert a user of the first mobile device.