RESUMO
Almond is an extendible open-source virtual assistant designed to help people access Internet services and IoT (Internet of Things) devices. Both are referred to as skills here. Service providers can easily enable their devices for Almond by defining proper APIs (Application Programming Interfaces) for ThingTalk in Thingpedia. ThingTalk is a virtual assistant programming language, and Thingpedia is an application encyclopedia. Almond uses a large neural network to translate user commands in natural language into ThingTalk programs. To obtain enough data for the training of the neural network, Genie was developed to synthesize pairs of user commands and corresponding ThingTalk programs based on a natural language template approach. In this work, we extended Genie to support Chinese. For 107 devices and 261 functions registered in Thingpedia, 649 Chinese primitive templates and 292 Chinese construct templates were analyzed and developed. Two models, seq2seq (sequence-to-sequence) and MQAN (multiple question answer network), were trained to translate user commands in Chinese into ThingTalk programs. Both models were evaluated, and the experiment results showed that MQAN outperformed seq2seq. The exact match, BLEU, and F1 token accuracy of MQAN were 0.7, 0.82, and 0.88, respectively. As a result, users could use Chinese in Almond to access Internet services and IoT devices registered in Thingpedia.
Assuntos
Aprendizado Profundo , Prunus dulcis , China , Humanos , Semântica , SoftwareRESUMO
The International Symposium on the Future ICT (Future-ICT 2019) in conjunction with the 4th International Symposium on Mobile Internet Security (MobiSec 2019) has been held on 17-19 October 2019 in Taichung, Taiwan [...].
RESUMO
Electrocardiogram (ECG) as one of the best methods to measure irregular heartbeats is a dispensable method for doctor to diagnose Acute Myocardial Infarction (AMI) patients. Most medical centers in Taiwan implement the reduction of Door to Balloon (D2B) time, which is defined as the time interval starting when an Acute-Myocardial-Infarction patient arrives at the Emergency Department, and ending when a catheter guide wire crosses the culprit lesion as the acute-myocardial-infarction treatment on the patient in the cardiac catheterization room. Generally, when a patient with acute-chest pain is sent into a hospital (always to Emergency Department), the hospital will collect his/her ECG which needs to be evaluated by a cardiologist to ensure that the patient really has Acute Myocardial Infarction. Then the medical workers deliver the patient to the cardiac catheterization room to operate balloon angioplasty. In previous years, the cardiologist must utilize a PC to connect to the Intranet of the hospital and employ a special PACS (Picture Archiving and Communication System) image browser before he/she can check the patient's ECG. But this will prolong the D2B time since the doctor may stay outdoors and he/she needs some time to find a PC and network. Of course, if the PC has not installed the PACS image browser, the doctor has to download and install it. Consequently, the D2B time should be worsened, thus possibly impacting the patient's life. Therefore, in this paper, we introduce a Cloud-based Electrocardiogram System, with which cardiologists can directly utilize their smart phones to browse the patient's ECG so as to shorten the D2B time. This system has been online in a medical center in middle Taiwan for more than one year. The shortened D2B time is longer than 10 min, i.e., receiving fine results.