We are creating some awesome events for you. Kindly bear with us.

Machine Learning to Translate Text Products into Non-English Languages

The National Weather Service (NWS) disseminates numerous forecast and warning text products to its partners and the general public to support its mission of protecting life and property. There is a need to translate NWS products into Spanish, the second most spoken language in the U.S, to better serve Spanish speaking communities across the country. It is estimated that in 2015, there are 45 million people, or roughly 13% of the population, who speak Spanish as a first or second language in the U.S.

NWS is requesting information on Machine Learning (ML) technology that can translate weather forecasts and warnings into Spanish and other languages. NWS often issues information and alerts to its partners and the general public via text messages. Currently, the NWS Weather Forecast Office (WFO) in San Juan manually translates information from the National Hurricane Centre’s Atlantic basin into Spanish However, WFO San Juan cannot sustain the manual translation of NHC Atlantic products due to the workload and limited personnel proficient in Spanish.

Within the NWS to date, there are numerous locally developed capabilities, many of which are labour intensive and not sustainable. As there is no nationally supported system to enable the generation of Spanish-language products or services, a near term and long-term enterprise solution needs to be developed.

Initially, NWS wants to use ML tools to create a “first guess” Spanish translation, so the San Juan forecasters can more quickly translate the products manually. Eventually, NWS aims to use a cloud-based solution to generate on-the-fly translations of free text, documents, webpage and social media content. The tool must use the same machine learning base model across messaging platforms to make sure that all the content is translated using the same baseline source material.

NWS Office of Central Processing (CP) will develop and incrementally implement a sustainable and scalable Spanish language translation approach that will integrate with existing enterprise systems NWS-wide. CP will automate the translation of the Atlantic basin products from the NHC by leveraging automated Artificial Intelligence (AI) followed by the translation of other key products at NWS WFOs prioritised with the largest percentage of Spanish speaking populations.

Additional offices will incorporate this capability as training of AI models and associated quality control of translated products are developed to address unique dialects of the local population. Also, CP will lay the groundwork to scale automated translations to additional languages in support of other Limited English Proficient (LEP) populations for the following languages: Spanish (Caribbean, Latin American, South American), Samoan, French, Tagalog.

U.S. Researchers have been utilising ML for various fields, including healthcare. As reported by OpenGov Asia, To help clinicians avoid remedies that may potentially contribute to a patient’s death, researchers at MIT have developed a machine learning model that could be used to identify treatments that pose a higher risk than other options. Their model can also warn doctors when a septic patient is approaching a medical dead end — the point when the patient will most likely die no matter what treatment is used — so that they can intervene before it is too late.

When utilised to a dataset of sepsis sufferers in a hospital intensive care unit, the investigator mannequin confirmed that about 12% of the therapies for deceased sufferers have been dangerous. The research additionally exhibits that about 3 p.c of sufferers who didn’t survive have been caught in a medical stalemate 48 hours earlier than demise.

Moving forward, the researchers also want to estimate causal relationships between treatment decisions and the evolution of patient health. They plan to continue enhancing the model so it can create uncertainty estimates around treatment values that would help doctors make more informed decisions. Another way to provide further validation of the model would be to apply it to data from other hospitals, which they hope to do in the future.

Send this to a friend