1. General
Category
SDG 4: Quality Education
SDG 8: Decent Work and Economic Growth
SDG 10: Reduced Inequality
Category
Computer Software
2. Project Details
Company or Institution
Talov
Project
Talov
General description of the AI solution
Talov creates accessible technology for people with disabilities. Currently, smartphones have a massive presence in almost every place, becoming a big opportunity for the development of accessible mobile apps that help or assist people with disabilities in daily-life common but challenging tasks.
With the previously described background, Talov’s work is one step in the right direction towards independence improvements for people with disabilities. Currently, we have two mobile apps in the market (both iOS App Store and Android Play Store), able to work offline in real-time and compatible with up to 48 languages:
SpeakLiz, for deaf people: cutting edge AI technology together with phone sensors to improve the communication levels between deaf people and hearing ones, through features like:
Sign language detection just by phone camera, allowing users to understand the gestures of different sign languages without third-party sensors.
Sound identification through the phone’s microphone allows users to constantly know what is happening around them in terms of sounds.
Conversation transcriptions in real time allow users to know what people are saying in the surroundings.
Text to speech engine (emoticons included), for generating a voice aloud just typing like in a common messaging app, so hearing people can hear what deaf people want to express.
Vision for blind people: phone’s sensors like camera and speaker together with advanced AI to improve interaction between blind people and the world by audible descriptions of their surroundings, through features like:
Detection of hundreds of objects and their relative position.
Money bills identification, currently 24 world currencies and growing.
Text reading in many surfaces, no more dependence only on the Braille system.
Navigation through accessible maps tool.
Color detection and contextual places detection (example: many computers and chairs could generate “Office” as audible output).
Approximate distance estimation through phone’s camera photogrammetry analysis.
Website
Organisation
Talov
3. Aspects
Excellence and Scientific Quality: Please detail the improvements made by the nominee or the nominees’ team or yourself if your applying for the award, and why they have been a success.
I have been working in AI since 2008 when this project was developed (it’s in Spanish, please turn on the translated subtitles):
Together with my team, we have been working in different AI algorithms:
Sound and speech recognition: feature extraction over signal time and other parameters like frequency domain, and then Convolutional Neural Networks (CNN) with a certain structure balancing between accuracy and computational costs.
Sign languages identification: Neural Networks with better performance for dynamic data (like sign language gestures), these NN are optimized for spatial graphs analysis.
Computer vision: Convolutional and Multilayer Neural Networks for the real-time extraction of patterns from the live camera feed.
We innovate by building very complex AI algorithms for very useful cases when it comes to assisting people with disabilities in daily life tasks. Our research work is now demonstrated in our current two apps, already launched to the App Stores, so we are in a commercial implementation of this AI research process, of course research and new developments never ends, we are constantly improving and updating our software. Our TRL is 8 and we are close to a 9th level, because we have users around the world that are sending us comments about how to improve, but we already left in the past the only-laboratory testing towards real-life usage. I believe we will be fully in 9th level when our technology can recognize for example all possible countries' sign languages, instead of the current four (American, French, German and Japanese). That’s only a matter of time, while we build our own datasets from scratch, but the operational model and its performance is now working. Regarding IP, our code is now under Apple’s & Google’s Stores IP protection, but we started our own IP processes depending on the countries we operate.
Scaling of impact to SDGs: Please detail how many citizens/communities and/or researchers/businesses this has had or can have a positive impact on, including particular groups where applicable and to what extent.
Currently, our freemium solutions are being used by more than 23.000 people in 86 countries in the free mode, 1800 have become paying subscribers with all Pro unlocked features. We have achieved these milestones without support from governments or large global organizations, just with our endeavours and a couple of startups contests where we have obtained some seed capital. We research and develop for all the disability spectrum, but considering just the two ones for which we have commercial solutions (deafness & blindness), in the world there are 760 million people (470M with hearing impairment and 290M with visual impairment).
We tackle challenges related to SDG 4, 8 and 10. The main contribution is reducing inequality of people with disabilities for better accessing for information or improving their interaction and independence levels regarding all their surrounding environments (social, educational, work, etc.), this utterly has positive consequences in accessing better job opportunities and their independence levels.
Since our technology is today being used in 86 countries, is compatible with up to 48 languages and can work in real-time in offline mode, makes it a potential candidate for massive spreading in a larger number of countries beneficiaries, we dream to reach and positively change the life of hundreds of millions of people, everyone with a mobile phone (now is not luxury, is a daily life tool and this will increase constantly), can be touched by our warm-purposed and social driven technology.
Our technology, with the correct global scale support by the main organizations in this field, can reach a lot of people with lower pricing or even free. All depends on many powerful organizations taking our work and spread it massively across the globe, converting it into a very vibrant AI development process to solve the most urgent SDG world challenges.
Scaling of AI solution: Please detail what proof of concept or implementations can you show now in terms of its efficacy and how the solution can be scaled to provide a global impact ad how realistic that scaling is.
By now, our two apps already launched with active users, you can see them here:
SpeakLiz iOS: https://apple.co/2SXTf7i
SpeakLiz Android: https://bit.ly/2Spnvb6
Vision iOS: https://apple.co/3wRNVkx
Vision Android: https://bit.ly/3xOq1GK
The main problem we have now is to go on massive numbers of users, now we have subscription plans for sustainability, but with the help of governments or large organizations we can search new ways to continue on sustainability, but making it more affordable, or even in some cases it can be free.
Our scalability approach to increase our AI platform for new user communities is extremely committed to them since the nature of Talov is research and development for all types of disabilities, we are always seeking new technologies and tools for improving current software and creating new tools. If Talov makes it, in terms of survival in this field, this is a very heavy precedent for a new generation of companies that are dedicated to put AI to the service of new social driven projects. The networking effects of this operation is very powerful, because for developing high tech is necessary the joining of many sectors like academical, corporative, financial, etc.
Currently, our freemium solutions are being used by more than 23.000 people in 86 countries in the free mode, 1800 have become paying subscribers with all Pro unlocked features. Regarding data protection, we are extremely committed to the fulfillment of GDPR or equivalent standards, due to a simple fact, our AI models always run offline inside the device, never collect or send users data to remote servers or anything similar, user’s privacy is one of our main concerns and that’s why we develop with that concept as a priority together with accessibility.
Ethical aspect: Please detail the way the solution addresses any of the main ethical aspects, including trustworthiness, bias, gender issues, etc.
Is important to notice that our mission will NEVER be replacing human assistance (like blindman guidance or sign language interpreters), but we are highly committed to bring access to assistive tools when users cannot continue paying such expensive services, our solutions address accessibility from a privacy-oriented perspective, in many cases is probably that people with disabilities don’t want third-party people assisting them in private situations. With that in mind, we can continue to the next stage, trying to solve the access to our technology in territories where salient power asymmetries persist.
Being a company whose core is the accessible technology to positively impact millions of people, definitely gives us the required trustworthiness, lawful principles fulfilling, regulations complying and the correct adherence to ethical principles and values, our dream to create the best accessible technology definitely becomes a robust ethical aspect of our startup.
Inclusiveness is a foundation factor at Talov, we add value to some technology tools for a social driven purpose, our innovative approach of putting high-end technology on people's phones is highly valuable because this allows to increase fair access to everyone. Our team is small but highly diverse and inclusive, some of our members are:
Diana Ceron: she’s blind since she was born, contributes with her amazing talent in Communications and also with her experience in product design & test for visual impairment.
Erick Ponce: he’s deaf since he was born, contributes with his amazing talent in Graphical Design and Photography and also with her experience in product design & test for hearing impairment.
We work on inclusiveness from inside the house, and of course our technology pushes the standard up in terms of searching for equality for everybody, despite any condition we all must have the same opportunities in our pursuing path of happiness and dreams fulfillment.