Abstract: A system, method and computer program product for training a deep neural network. The deep neural network can be trained using a learning process that is defined to optimize both an error function of the deep neural network as well as a network mapping function of the deep neural network. The network mapping function can represent a predicted label distribution geometry property of the deep neural network. This learning process can improve the accuracy of the trained deep neural network model as well as its robustness again adversarial attacks. Optimizing the network mapping function can also provide increased insight into the operation of the trained deep neural network model, which may promote increased interpretability of the trained model and thus encourage uptake of the trained model.

Abstract: An identification circuit for a CATV plug-in fixed attenuator is disclosed. The disclosed identification circuit may include an attenuator, a power supply module to supply power, a detection module to detect an output divided voltage value, and a microcontroller unit (MCU) identification module to determine an attenuation value of the CATV plug-in fixed attenuator based on the detected divided voltage value. The power supply module is connected to an input terminal of the plug-in fixed attenuator. The detection module is connected to an output terminal of the plug-in fixed attenuator and to the MCU identification module. With the disclosed embodiments, attenuation parameters can be determined remotely and failures on the line can be located.

Abstract: Embodiments of the present disclosure disclose an equalization circuit for a CATV plug-in fixed attenuator which comprises an equalizer, an equalization compensation circuit and a MCU module. The equalizer comprises a CATV plug-in fixed attenuator identification module and an equalizing module. The CATV plug-in fixed attenuator identification module is connected to the equalizing module and the MCU module. The equalization compensation circuit is connected to an output terminal of the equalizer and the MCU module. The embodiments enable identification of a changing attenuation value and compensation of the frequency response.

Abstract: Embodiments of the present disclosure disclose an equalization circuit for a CATV plug-in fixed attenuator which comprises an equalizer, an equalization compensation circuit and a MCU module. The equalizer comprises a CATV plug-in fixed attenuator identification module and an equalizing module. The CATV plug-in fixed attenuator identification module is connected to the equalizing module and the MCU module. The equalization compensation circuit is connected to an output terminal of the equalizer and the MCU module. The embodiments enable identification of a changing attenuation value and compensation of the frequency response.

Abstract: An identification circuit for a CATV plug-in fixed attenuator is disclosed. The disclosed identification circuit may include an attenuator, a power supply module to supply power, a detection module to detect an output divided voltage value, and a microcontroller unit (MCU) identification module to determine an attenuation value of the CATV plug-in fixed attenuator based on the detected divided voltage value. The power supply module is connected to an input terminal of the plug-in fixed attenuator. The detection module is connected to an output terminal of the plug-in fixed attenuator and to the MCU identification module. With the disclosed embodiments, attenuation parameters can be determined remotely and failures on the line can be located.