5分で分かる!? 有名論文ナナメ読み：Jacob Devlin et al. : BERT : Pre-training of Deep Bidirectional Transformers for Language Understanding

An intelligent automated assistant system engages with the user in an integrated, conversational manner using natural language dialog, and invokes external services when appropriate to obtain information or perform various actions. The system can be implemented using any of a number of different platforms, such as the web, email, smartphone, and the like, or any combination thereof. In one embodiment, the system is based on sets of interrelated domains and tasks, and employs additional functionally powered by external services with which the system can interact.

Intelligent Automated Assistant

Recent developments and technological advancements in wireless communication, MicroElectroMechanical Systems (MEMS) technology and integrated circuits has enabled low-power, intelligent, miniaturized, invasive/non-invasive micro and nano-technology sensor nodes strategically placed in or around the human body to be used in various applications, such as personal health monitoring. This exciting new area of research is called Wireless Body Area Networks (WBANs) and leverages the emerging IEEE 802.15.6 and IEEE 802.15.4j standards, specifically standardized for medical WBANs. The aim of WBANs is to simplify and improve speed, accuracy, and reliability of communication of sensors/actuators within, on, and in the immediate proximity of a human body. The vast scope of challenges associated with WBANs has led to numerous publications. In this paper, we survey the current state-of-art of WBANs based on the latest standards and publications. Open issues and challenges within each area are also explored as a source of inspiration towards future developments in WBANs.

Wireless Body Area Networks: A Survey

A detecting, monitoring and reporting system includes a sensor device (10) which generates data indicative of physiological parameters of an individual and data derived therefrom when placed in proximity with the body. A remote central monitoring unit (30) generates analytical status data from the data indicative of physiological parameters, the derived data, and analytical status data that has previously been generated. A data storage device stores the data received and generated by the central monitoring unit (30). The system includes means (50, 55) for establishing electronic communication between the sensor device (10) and the central monitoring unit (30), and means for transmitting the data to a recipient. The sensor device (10) may also generate data indicative of contextual parameters of the individual, which may then be used when generating the analytical status data. The system may be used for monitoring and providing feedback relating to the degree to which an individual has followed a suggested routine.

System for monitoring health, wellness and fitness

A monitoring system comprises a module (55) having at least one sensor (120) and preferably skin and ambient temperature sensors (120) within a housing (95). The device may be durable or disposable. The housing (95) may be provided with certain surface features and shapes to facilitate mounting on and interface with the skin of the wearer for more accurate temperature measurement. A receiver (not shown) may be provided to obtain and display data from the module (55). The module (55) may also display the output data. The output data comprises both detected and derived data relating to physiological and contextual parameters of the wearer and may be transmitted directly to a local recipient or remotely over a communications network. The system is capable of deriving and predicting the occurrence of a number of physiological and conditional states and events and reporting the same as output data.

Non-invasive temperature monitoring device

A wearable life support apparatus worn by a user is disclosed. A physiological information acquirement unit acquires the user's physiological information. An action information acquirement unit acquires the user's action information. A status recognition unit recognizes the user's status in accordance with the action information. A physiological information decision unit decides whether the physiological information is normal in accordance with the user's status. A presentation unit presents the decision result of the physiological information decision unit to the user.

Wearable life support apparatus and method

According to one embodiment, a voice recognition apparatus includes a determination unit, an estimating unit, and a voice recognition unit. The determination unit determines whether a component with a frequency of not less than 1000 Hz and with a level not lower than a predetermined level is included in a sound input from a plurality of microphones. The estimating unit estimates a sound source direction of the sound when the determination unit determines that the component is included in the sound. The voice recognition unit recognizes whether the sound obtained in the sound source direction coincides with a voice model registered beforehand.

Voice recognition apparatus

A bio-information measuring includes a detector including a light source emitting light to a hand of an examinee and an optical receiver receiving the light emitted from the light source; and a supporter bent to clamp a paddle between fingers of the examinee. The clamping portion of the supporter is provided with the detector. The bio-information measuring apparatus also includes a measurer worn on the examinee, measuring bio-information based on pulse wave data obtained from the optical receiver; and a puller pulling the support member toward the measurer.

Bio-information measuring apparatus

A biotic sleep state determining apparatus that determines a sleep state of a subject based on a series of pulse interval data that indicate a time interval of one cycle of a pulse wave of the subject, and on body movement data that indicates a body movement of the subject, the apparatus includes a body movement determining unit that determines that the body movement occurs if a fluctuation amount of the body movement data is greater than a first predetermined threshold, a pulse interval data processor that processes the series of pulse interval data after removing pulse interval data measured in parallel with the body movement data from the series of pulse interval data, if the body movement determining unit determines that the body movement occurs, and a sleep state determining unit that determines the sleep state based on autonomic nerve indexes acquired from the series of pulse interval data processed by the pulse interval data processor.

Apparatus for and method of biotic sleep state determining

Introduces a prototype of wearable healthcare support system 'LifeMinder', which consists of a wristwatch-shaped wearable sensor module and a personal digital assistant (PDA). The wearable sensor module, equipped with sensors of accelerometer, pulse meter thermometer galvanic skin reflex (GSR) electrodes and Bluetooth module to communicate with the PDA, monitors the user context: health conditions, movements and behaviors. The system uses this information to guide the user in daily self-care in real time. Diet care and exercise care are especially significant to prevent the "lifestyle-related disease". The authors developed algorithms to recognize the user movement from wrist motion and to detect the beginning of a meal from pulse rates and GSR values. The accuracy of both algorithms is about 90%.

Kazushige Ouchi

Papers

Wearable life support apparatus and method

Voice recognition apparatus

Bio-information measuring apparatus

Apparatus for and method of biotic sleep state determining

LifeMinder: a wearable healthcare support system using user's context