Panasonic

About: Panasonic is a company organization based out in Kadoma, Ôsaka, Japan. It is known for research contribution in the topics: Signal & Layer (electronics). The organization has 49129 authors who have published 71118 publications receiving 942756 citations. The organization is also known as: Panasonikku Kabushiki-gaisha & Panasonic.

Towards Deep Neural Network Architectures Robust to Adversarial Examples

Abstract: Recent work has shown deep neural networks (DNNs) to be highly susceptible to well-designed, small perturbations at the input layer, or so-called adversarial examples. Taking images as an example, such distortions are often imperceptible, but can result in 100% mis-classification for a state of the art DNN. We study the structure of adversarial examples and explore network topology, pre-processing and training strategies to improve the robustness of DNNs. We perform various experiments to assess the removability of adversarial examples by corrupting with additional noise and pre-processing with denoising autoencoders (DAEs). We find that DAEs can remove substantial amounts of the adversarial noise. How- ever, when stacking the DAE with the original DNN, the resulting network can again be attacked by new adversarial examples with even smaller distortion. As a solution, we propose Deep Contractive Network, a model with a new end-to-end training procedure that includes a smoothness penalty inspired by the contractive autoencoder (CAE). This increases the network robustness to adversarial examples, without a significant performance penalty.

Apparatus for retrieving similar documents and apparatus for extracting relevant keywords

Abstract: After three kinds of data, i.e., a keyword frequency-of-appearance (103), a document length (105), and a keyword weight (107) are produced, a document profile vector (111) and a keyword profile vector (109) are calculated. Then, by independently performing the weighted principal component analysis (112,114) considering the document length and the keyword weight, a document feature vector and a keyword feature vectors are obtained. Then, documents and keywords having higher similarity to the feature vectors calculated with reference to the retrieval and extracting conditions are obtained and displayed.

Notification system for television receivers

Abstract: A notification system for television receivers including a visible alarm and/or an audible alarm that is activated when a certain event occurs. An information signal is extracted from a received television signal and is processed to determine the status of an event. If it is determined that the event occurred, then a control signal is sent to the appropriate alarm to activate that alarm, thereby notifying the user of the event's occurrence. The event may include the reception of a text or graphic message, the televising of a certain television program, and the televising of a television program with a specific program rating. The user may also deactivate the alarm. A personal identification number may be used to restrict access to certain features of the notification system.

Semiconductor device and method for producing the same

Abstract: A semiconductor region having an upper surface and a side surface is formed on a substrate. A first impurity region is formed in an upper portion of the semiconductor region. A second impurity region is formed in a side portion of the semiconductor region. The resistivity of the second impurity region is substantially equal to or smaller than that of the first impurity region.

Communication device, communication system and authentication method

Abstract: A communication device, a communication method, and an authentication method where dishonest impersonation by a third person is prevented and safety and reliability of authentication processing are improved. A slave (20) sends an authentication request including device information to a mater (10). The master (10) receives the authentication request and causes the device information to be displayed on the screen of a display section (13). A user visually confirms the device information displayed on the screen of the display section (13), determines whether or not to authorize the authentication, and instructs the master (10) about the result through an input section (14). The master (10), having been instructed about the approval/disapproval of the authentication, sends a response according to the instruction to the slave (20).