For those who can hear, it’s likely that we haven’t really thought about how delays, cancellations, and safety notices are usually delivered through loudspeakers, and updates on information screens are rare if they exist at all. While this might irritate people who can hear, it’s a much bigger problem for people who are Deaf or have hearing difficulties. In some situations, it can even be dangerous or scary. Even basic travel information isn’t usually provided in the sign language that many Deaf people use. Instead, they’re expected to understand written information in their second or third language, which can be tough. Enters the idea of AI and sign language!
The UK’s Network Rail is one of several transportation companies across the world trying to fix a global issue. Since April 2023, they’ve introduced screens at eight major railway stations in the UK that provide travel advice in British Sign Language (BSL). These screens give Deaf people access to regular announcements and updates about problems or emergencies in their preferred language. This is a good step forward to make public places more accessible and welcoming, although it’s been a long time coming.
Solving an Emerging Issue: AI and Sign Language Translation
The challenge of using artificial intelligence for sign language translation and interpretation has sparked a lot of interest. Companies like Signapse, which is based in Surrey Research Park, are working on this. They’re developing software that can translate sign language in real-time, and they’ve even created artificial signers for things like broadcasting, transportation, and making websites more accessible.
What sets Signapse apart is how they’re doing it. They have a team of people who are Deaf and who speak sign language, as well as a team of people who can hear and use spoken language, and they’re closely connected with organizations that support Deaf people. This way, they’re ensuring their technology matches the needs of the Deaf community. Instead of going with the popular choice of using animated characters for sign language, which Deaf people don’t really like, they’ve created artificial signers that look like real people. They’re using special computer techniques to make videos that look like they’re real Deaf people signing. And those real Deaf people are paid when their videos are used.
Similar to how computers talk to us in the train station, the BSL videos are made ahead of time and put together to create regular messages. Artificial intelligence helps make the videos smooth and natural. If something unusual needs to be said and it’s not in the messages they already have, a team of experts is ready to help turn the new information into sign language so they can have a video ready in an hour.
Going Beyond Travel: Signapse’s Big Plan
Signapse doesn’t just want to help with travel announcements. They’re also aiming to use their technology in different areas beyond transportation and sign language applications. It’s a big goal that won’t be easy to achieve.
The Special Challenge of Sign Languages
The way sign languages work makes them different from spoken languages, making them a challenge for AI. Unlike spoken languages, there’s no one way to write down sign languages that everyone agrees on. This makes it hard to use the computer programs we already have for understanding text. Sign languages use hand movements, facial expressions, and body motions to communicate; they can convey many ideas simultaneously. This makes things quite complex. While some people have devised ways to write down sign languages, these methods have yet to become widely accepted.
Because of this complexity, there’s not a lot of information available for computers to learn from. People who understand these special writing systems or ways of talking in sign language are needed to help computers learn. But even with these experts, there’s little information to learn because sign languages aren’t used as much as spoken ones.
This also makes it hard for computers to learn sign languages on their own, like they’ve been learning spoken languages from huge amounts of information on the internet.
The Good News for AI and Sign Language
There’s some good news, though. Computers are improving at creating sign language, at least in some cases. If the way Signapse makes signers can be used in more situations, it could help computers learn even more. However, there’s no doubt that we need more real videos of people using sign language and better ways to hide their identity. We also need to work closely with the Deaf community to make sure this technology works well for them.
In the journey to make sign language understandable for computers, we need to come up with new ideas, work together from different fields, and listen to what the Deaf community has to say. As technology advances, we should remember to make sure everyone can communicate easily and that no one is left out because of how they talk or hear.
Seldean is a multi-skilled content wizard that dedicates herself to writing content that goes beyond merely sparking interest in the audience.
0 Comments