A: I was born in Greece, but my family moved to the USA when I was five years old, so I grew up in a number of states in the US. As a result, I was exposed to numerous languages and dialects and I later studied international business and government, as I’ve always been interested in global perspectives and cultures.
I am a native speaker of Greek and English, I also speak Spanish, some French, beginner’s Russian and Arabic, and I am now also learning American sign language, so I can more easily have a conversation with our partners from Gallaudet University. I love languages of all forms and I would like to speak even more. This is what intrigued me about AppTek as a company, the automation of language and the option to offer everyone the ability to communicate in any of them.
A: I am an amateur photographer and do some landscaping too, but my passion project is my work with children’s charities. I’ve been involved in this for years. I have my own non-profit with the mission to provide technology and skills to underserved children in inner-city Baltimore. I had a number of mentors and supportive people myself at a young age, which has been a great motivation for me for this project, on top of my desire to help children in the memory of two friends of mine. I am also on the board of directors of Casa DC, a wonderful organization who I’ve collaborated with, that also helps at risk children, who are in foster care and in the juvenile justice system.
A: I got interested and involved with accessibility when I joined AppTek and met Dr Christian Vogler, professor at Gallaudet University and director of the Technology Access Program research group. As often happens, many solutions are developed out of some personal experience. In my case, my background in languages made AppTek more interesting to me as a company to work in, while collaborating with the people at Gallaudet University made it more relevant to me to try and solve this accessibility problem as a product person.
Accessibility to me means equal experiences for everyone. In terms of the deaf and HOH, it means that they have the same access to communication, the same information and opportunities as hearing individuals. Many podcasts and online courses for instance are not accessible for the deaf and HOH because they weren’t developed with universal accessibility in mind so as to include their needs too. This goes beyond captioning, it is a major oversight of many digital platforms.
Imagine not being able to have a conversation with your doctor and tell them what’s wrong or to take a class so you can get hired for a new job. This is why accessibility is important and why we’ve been working with the deaf and HOH community to develop solutions. What got me passionate about it is the friendship I developed with people who are deaf and seeing how much work we had to go through just to have a simple exchange. I got a better understanding of what that medical scenario must have been like for them. Add to that the isolation of the pandemic quarantines and it became that much more critical to work on accessibility applications.
Overall, there’s much work to be done for the spectrum of disabilities. In terms of the deaf and HOH, we are behind in our inclusive thinking. The USA as a country is more aware today, but if you think that closed captioning on television has been around since the 1960’s and yet not all broadcast content is 100% captioned, it’s obvious that we still have work to do. The digital information space is also at a deficit and it’s difficult to catch-up. The evolution of greater computing and the understanding and adoption of machine learning and neural networks in the natural language processing and understanding (NLP/U) field is giving us real opportunities to provide communication and information intake applications that really work.
A: When I joined AppTek for the second time in 2018, I was asked to manage a project with Gallaudet University which was researching ways in which to improve the informativeness of closed captions for the deaf and HOH community, such as latency, punctuation, capitalization etc. As I got to understand this new area of communication, sign language vs. verbal languages, it became very interesting to me because it was a completely new set of challenges that I hadn’t had been exposed to before.
I then got to know our partners better and I realized how challenging it was to communicate with them and how much information and meaning got lost, even when going through an interpreter. This became frustrating and impactful, so solving these challenges became personal to me at that point. Especially because as a person who speaks a few languages, I enjoy so much communicating and connecting with people. It excited me that we could use technology to solve a communication problem.
The pandemic created a massive crisis for the deaf and HOH community, because they were literally quarantined from communications, as they suffered from a technology deficit. This is how I started working on the GoVoBo project with the Technology Access Program research group of the Gallaudet University. The team is the gateway for the needs of the deaf and HOH community and they set the goals of the project.
A: AppTek’s accessibility service is indeed the backend that GoVoBo uses. But GoVoBo is a turnkey, platform-agnostic application, with a user interface that utilizes AppTek’s core automatic speech recognition (ASR) and machine translation (MT) cloud service. The other differentiator is that GoVoBo is a separate legal entity, a spinoff formed by AppTek and Gallaudet University.
The plan is for the company to license the GoVoBo platform for profit, as a commercial entity. But there is a community-giving element too, whereby licenses will be provided for free to applicants from deaf and HOH organizations. Dr Christian Vogler and Dr Patrick Boudreault are principal founders of GoVoBo, who have a deep understanding of the deaf community, strong ties to our users, and first-hand experience with the technology deficit in digital platforms and media accessibility. This is precisely the focus of the company, i.e. to serve the needs of this community first and foremost, even if the GoVoBo application itself can be used by hearing users as well.
What makes GoVoBo different to other applications out there that may be native to commonly used platforms, such as Zoom, is that it is agnostic, it can be used universally across any digital platform. It consists of a desktop application that interfaces to AppTek’s cloud-based accessibility service. The app’s function is to set up the user preferences in terms of color, font, etc. and it can run in parallel to any other app on a laptop. It uses the laptop’s operating system to ingest the sound and get the native audio that is used as input for the speech recognition system. As the source audio is not degraded, e.g. via the speakers, it provides for better quality recognition.
A. We are currently launching Version 1 of the GoVoBo app, for laptops. The next version will be for mobile devices, so users can literally take the app with them wherever they go. In the future, we want to integrate other technologies too, such as computer vision, for automatic sign language recognition, so that deaf users can communicate easier with people that are not familiar with sign language. The end goal is to provide freedom of communication and interaction to deaf users.
An ever-growing challenge for any ASR system is, of course, solving the problem of language variety, dialects and accents. Deaf-accented speech, in particular, is a focus for our ASR development, so as to improve the functionality of the application for deaf users and their ability to understand and engage in the digital space.
Natural language understanding is also key. What is currently missing from accessibility solutions and closed captioning systems is emotion. The ability to recognise and tag the meta elements of speech, like pace, tone, volume, speed, etc. to identify emotion. We capture this information and want to be able to convey all the elements in a conversation, so deaf users have the same complete experience in a conversation a hearing user has. We can do this by exposing such information in a visual manner, e.g. by colour coding emotion. This way we can enhance the informativeness of every conversation a deaf person has with the hearing world.
A: GoVoBo is already in use, it was used recently to caption a live disability technologies summit. It has also completed a user testing process with Verizon Communications and a group of their deaf users. The latter tested the app’s captioning accuracy, translation quality, latency and ease of use. The result was favourable, with the majority of the users having answered yes to the option “I would like to use this today”.
As I mentioned before, the core ASR service that is the backend of the GoVoBo app, fuels all of AppTek’s accessibility solutions. It also comes in the form of an appliance used on premise by US and Canadian broadcasters for the provision of automatic live captioning for their programming.
A: My vision for the future from the deaf users’ perspective is that deaf users will have the same experience in the digital world as hearing users do. That they will have the freedom to communicate and engage in the same way that hearing users do, directly, in their own terms, in real time, in the fashion that they need, at a moment’s notice via an app on their phone, or through smart glasses, incorporating speech and sign language. We as GoVoBo want to be the AI company that delivers those solutions to them. That provides enough tools to them so that they don’t have to put as much planning and forethought into having a simple act of communication, and so that they do it in real time just like all the rest of us do.
AppTek is a global leader in artificial intelligence (AI) and machine learning (ML) technologies for automatic speech recognition (ASR), neural machine translation (NMT), natural language processing/understanding (NLP/U) and text-to-speech (TTS) technologies. The AppTek platform delivers industry-leading solutions for organizations across a breadth of global markets such as media and entertainment, call centers, government, enterprise business, and more. Built by scientists and research engineers who are recognized among the best in the world, AppTek’s solutions cover a wide array of languages/ dialects, channels, domains and demographics.
