Because time is a crucial factor in providing help, Several Singaporean entities and universities are working on employing artificial intelligence in recording emergency calls and analyzing their content to speed up the response and increase the efficiency of emergency services.
Response teams in Singapore receive nearly 200,000 calls annually. In each of them, One second can determine the fate of a human being in need. Many details go into the drawing of this scene, Including the readiness of the response crews, their possession of the necessary equipment and the speed of their work. But it all depends first on the accuracy of the information they receive from callers as answers to their questions. Which determines the course of the entire intervention process.
Must In a multi-ethnic society, From taking into account the differences and diversity of the population when designing modern services, This is to provide experiences that are relevant to people's lives and able to achieve equal opportunities among users. Singaporeans speak several languages and dialects. Among them is the "English" language, which is English with a Singaporean accent in which Mandarin and Malay merge. The nature of this language is based on switching between the unique phrases of the languages it composes. In fact, speech recognition and automatic response programs fail to understand many of these languages.
The authorities established the Singapore AI Programme to qualify local talent and lead innovations in AI to serve the public interest. The Singapore Civil Defence Authority and other government bodies, including two universities, have cooperated. To employ modern technology in improving emergency services. I developed a new program that can recognize spoken languages spoken by the public.
The program, In addition to recording every call received by responders in real time, by transcripting these calls, It extracts the most important information such as the location, cause and severity of the injury. What distinguishes this program is the possibility of dealing with different languages and dialects, Unlike previous versions, which can only analyze speech in standard English, Its accuracy in transcripting emergency calls is close to 90%.
To design the program, The researchers collected more than a thousand hours of recorded call samples from Singapore and Penang (where different languages share the formation of the spoken dialect). Along with recordings of Singaporeans speaking their spoken dialects on radio, YouTube and SoundCloud. In the next stage, Specialized teams manually unloaded these recordings. Finally, Both versions were introduced into the system to use deep learning technology to build a link between spoken speech and written text. As a result, The system was able to learn nearly 40,000 English words and the same in Mandarin.
The National Research Foundation's Artificial Intelligence Program has allocated $1.25 million in funding to establish a laboratory specialized in spoken speech analysis. Led by researchers from the National University of Singapore and Nanyang Technological University. in the laboratory, Deep learning technology, which produces algorithms that mimic the neural pathways of the human brain, was used to help computers perform new tasks and analyze data.
More broadly, The Civil Defense Forces Authority can conduct analyses of the texts and data collected by the program, This is to determine the times and locations of peak need for emergency services. For example DEWA relied on accurate data obtained from previous response operations to select locations to distribute stationary ambulances with crews on standby to assist at vital points during the morning and evening peak periods.
Through this innovation, Singapore has joined the ranks of countries seeking to integrate AI into emergency response operations. In a similar experiment, Respondents in Copenhagen use a program that identifies cardiac arrest symptoms over the phone by monitoring verbal and nonverbal communication patterns such as tone of voice and breathing frequency. This is an important progress, Every minute of delay in treating the injured person reduces the chance of survival by 7 to 10%. In Sweden, Another system is being tested to speed up sea and air rescue operations.
Moreover, The Singapore government is using artificial intelligence in the areas of recruitment and fraud detection to gain support. It is currently seeking the participation of different sectors in capacity building in the field of AI. In order for services to be as diverse as the population, The Smart State and Digital Government Office uses a set of standards that detail the digital principles and the basics of designing services that take into account different groups such as the elderly, The team developed a personal button for emergency use. In normal scenarios, they try to alert family members or those in close proximity first. It is worth noting that several States have taken this approach. Like New Zealand, which works to enable people of determination to access government organisations' websites, Or Georgia, which has hired sign language interpreters and video calls to respond to deaf and dumb emergencies.
The Smart State Office is also working with the Land Transport Authority to predict how changing bus routes will affect passengers using available data. This analytical tool allows faster and better trip planning and prevents disruptions to public transport.
The text discharge system is still not available for commercial use, But it outperforms the options available in the market, The authorities intend to study its implications and challenges to assess the possibility of using it on a national scale as part of the government's plan to provide easy, usable and easily accessible digital services.
These technologies save the time that response teams need to intervene, This means delivering better performance and employing resources more efficiently in the long run.
References:
https://www.ship-technology.com/analysis/mayday-mayday-answering-emergency-calls-with-help-from-ai/
https://www.theverge.com/2018/4/25/17278994/ai-cardiac-arrest-corti-emergency-call-response
