Yandex Blog

Yandex DeepHD Technology Revolutionizing Photo & Video Quality for Yandex Services

As the quality and resolution of the screens in our homes and pockets reaches ever higher levels, the comparatively lower quality of photos and videos made decades ago becomes more obvious.  Here at Yandex, the Computer Vision team within our Machine Intelligence and Research Division has developed a neural-network based super-resolution technology, DeepHD, to help bring these media classics into the digital age. DeepHD enhances the quality of both photos and videos and with its application to Yandex images and content hosted on our video and TV streaming services, it is the first technology of its kind in production.

Image and video frame processing in DeepHD requires two steps, each using its own neural network. The first stage is the preliminary preparation of the image.  The neural network removes compression artifacts from the image, which are otherwise referred to as “noise” and commonly occur in images that have been digitally processed.

After the image is cleared of noise, it is transmitted to the second neural network -- the generator, which increases the resolution of the image. The Yandex Computer Vision team uses GANs (Generative Adversarial Networks, Goodfellow et al. (2014)) architectures to train this network.  GANs are neural network architectures in which one network generates high-resolution artificial images from low-resolution ones, while trying to make them indistinguishable from real high-resolution images to another network.  The process of increasing the resolution of images and videos is very similar; DeepHD enhances videos by processing each individual frame.

The technology is able to sharpen various aspects of images in videos such as improving the visibility of an object in the shade or even bringing the texture of an actor’s clothing to life by making the smallest details of the fabric more visible.  DeepHD does not alter the image by adding things that aren’t already there; the frames of images are semantically identical before and after running DeepHD. 

DeepHD is particularly useful when only small or low-quality images are available. One example of this is when a photo has been cropped and a user can’t find the original photo.  DeepHD has been applied across our entire database of Yandex images to provide users searching for photos with better, large versions of images.

Our Computer Vision team has also applied DeepHD to a number of low-resolution films and cartoons on our streaming service KinoPoisk.  These include famous Soviet-era movies that have, up to now, usually only been available in low quality.  Thanks to DeepHD, people can now experience these classic titles in stunning, modern quality online. 

An easy way for users to find these videos is searching on Yandex.ru, where they can enter queries on films and cartoons with “DeepHD”.  Viewers will know they have found a DeepHD stream when it is accompanied by the “dHD” logo.

Similarly, DeepHD has also improved the user experience for our streaming television service, Yandex.Efir.  The service provides viewers with live streams of television channels, and with the new integration of DeepHD, users can view higher quality streams that the technology upscales in real time, also marked by the “dHD” logo.

We look forward to expanding our collection of films and videos enhanced with DeepHD super-resolution technology so that our users can experience even more content in stunning high quality.

Yandex Celebrates Our Community of Coders

During the month of September, the global tech community recognizes three important celebrations relating to Yandex:  Programmers’ Day, Online Learning Day, and National Coding Week.  Here in the Yandex community, we are always appreciative of the work of our thousands of programmers and learners of all ages in our educational programs. For these occasions,  we wanted to extend a special thank you to our community of professionals, instructors, and students who are all helping grow the data science field and working to shape the future of tech in our lives. 

In honor of the celebrations, we will share stories from some of our programmers on what inspired them to study coding and how they began their programming journeys, plus share more about our educational initiatives.

Michael Roizner

Team Lead, Machine Learning Marines

When I was about 10 years old, my father first showed me how to code some very simple things. I had already had a computer for many years but used it only for games. My first program had just a green circle moving to the right, but it was like a wow-effect for me. I realised for the first time that things in a computer work not in a magically predefined way, but I that I can make them work how I want them to work.

Kostya Kramlih

Lead Developer, Yandex.Browser

I think that the beginning of my interest appeared in my childhood, when I was about 7 years old. My dad was interested in computers and in our home we had the ZX Spectrum, and after that a PC based on i486. I suspect that this partly prompted me to get a degree in mathematics-programming.  Then I got my first job as a developer in a small Novosibirsk company.  Since then I have realized that programming is exactly the area in which I would like to work.

Valentin Sinitsyn

Team Lead of Overlay Networking, Yandex.Cloud

I was around 9 when my dad brought me to computer classes. Frankly speaking, I didn't like them at first. Computers were not yet everywhere in Russia at that time, so I quickly realized that being able to program simple games in BASIC earned me some brownie points in classmate's eyes. That was my secret superpower, much like a superhero’s, but real.

Sasha Ganshin

Developer, Yandex.Weather

It's funny that now I can say computer games drove me to learn more about coding. When I was about 5, we had an old computer, ‘Master.’ On this computer, you had to load games from a cassette. There was a start screen before loading the game - a BASIC console. One day the cassette was broken, so the only way I could use the computer was to try commands I found in the manual.

As the experiences of our programmers show, we are often inspired by a subject at a young age when we have welcoming opportunities and positive experiences.  At Yandex, we believe in helping create more of those opportunities for young learners. 

Beyond the Yandex company programmers, our community includes Yandex.Lyceum, a program available to teenagers interested in coding, which is taught at 129 locations in 58 cities across Russia and Kazakhstan.  Currently, up to 4,000 students are enrolled in Yandex.Lyceum to learn the fundamentals of programming in Python and those of industrial programming.  Since 2011, we have also hosted a coding competition, Yandex.Algorithm, which is open to aspiring developers as young as six.  In 2018, Yandex.Algorithm welcomed over 11,500 people from 93 countries to test their skills.

Our community and journeys in programming don’t stop there.  A big part of our data science community is centered around the Yandex School of Data Analysis (YSDA), which was established in 2007 to offer a free Master’s Level program in Data Science and Data Analysis. Today, YSDA spans five branches and boasts over 700 graduates who are working in the world’s leading tech companies. 

Most recently, YSDA launched a one-year career advancement program in Israel, where Yandex also created the Yandex Machine Learning Initiative at Tel Aviv University to support undergraduate studies.  The YSDA community also expands a global reach through Coursera courses, 15 of which are taught in English.  Since 2016, nearly 64,000 people have learned C++, Python, and JavaScript via our online courses.

Many Yandex employees and graduates of YSDA realized their potential and interest in programming through opportunities in higher education:

Sergey Zhigalov

Senior Frontend Developer

My love of writing code began at the university with programming robots. This is a good example of how something virtual, expressed as lines of code, can con control something physical - the "hands" of robots.

Anton Slesarev

Developer for Self-Driving Technology

During my third year at Moscow Institute of Physics and Technology (MIPT), we had to choose a science company to work with.  We had a tour of a company which worked on computer vision and object detection, areas I really liked, so I decided to apply there. It was then that it occured to me that I should work in programming, and that’s how I chose this profession and have never regretted it.

Irina Goltsman

Analyst-Developer, Geotargeting

I didn’t plan on being a programmer - I started university wanting to be the head of an IT startup.  But I always thought that you should be well-versed in what you are going to lead - so I went to MIPT and learned to program.  After my second year I went to work as a C++ developer, after my fourth year I ended up getting a job at Yandex as an analyst-developer.

Whether you are a young student just learning Python or a professional looking to make a career change into machine learning, there are countless opportunities ahead. Congratulations to everyone in the coding community on Programmers Day, Online Learning Day, and National Coding Week!

Safety and the Future of Autonomous Vehicles Explored by Yandex Self-Driving and Russian Federal Traffic Officials

The Yandex self-driving team first began working on our driverless technology in 2016, combining our expertise in machine learning, navigation, mapping tools, and cloud technologies.  In May of 2017, we introduced our prototype and since then have rapidly advanced from testing on closed tracks to driving on the busy streets of Moscow and our most recent fully autonomous ride-hailing launch.

As we introduced our autonomous vehicles to the challenging roads of Russia, we have been able to test our technology against diverse traffic conditions and situations in which pedestrians or other vehicles don’t follow the rules of the road.  Safely and efficiently managing our vehicles  is the highest priority in our self-driving development and requires extensive testing and knowledge of the roads and other drivers.

Last week, as part of our commitment to safety, our self-driving team met with the Moscow division of The General Administration for Traffic Safety of the Ministry of Internal Affairs of Russia (GIBDD) to discuss the development and future of self-driving technology in Russia.  The meeting marks a very important step in our collaboration with the GIBDD to better understand their expertise on traffic management and road safety, plus their views on the future integration of autonomous cars on Russian roads.

During the meeting, the Yandex team provided a demo of the self-driving car to the highest ranking officer of GIBDD, Lieutenant General Mikhail Chernikov, explaining our technologies, testing, and vision for the future.  Together our teams discussed the major challenges of integrating self-driving cars on public roads and our joint commitment to guaranteeing the safety of all people on the road, as well as the adoption of legislation on the emergence of new self-driving cars.

The two organizations addressed ways in which self-driving cars can help reduce road accidents and improve safety.  Citing official road accident statistics, Chernikov explained how human error accounts for more than 85% of accidents and his belief that self-driving will eventually help reduce the number of accidents.  Self-driving vehicles will not violate traffic rules, do not engage in distracting behavior, and utilize sensors to continuously monitor 360 degrees of traffic conditions around the vehicle.  The meeting also covered how legislation will work when there are issues on the road with a self-driving car. 

Moving ahead, Yandex looks forward to a continued fruitful relationship with The General Administration for Traffic Safety of the Ministry of Internal Affairs of Russia and together advancing the agenda on developing and integrating self-driving cars to improve urban living and the daily commutes of Russian citizens.

Images courtesy of GIBDD

Yandex Provides Cloud Platform to Companies Operating in Russia

At Yandex, we have always been committed to spreading technology and supporting other companies’ development by providing various APIs and open sourcing Yandex technologies.  More recently, in recognition of the need for an advanced cloud service in Russia, we started building a cloud platform as an important step in advancing the global tech ecosystem and offering useful solutions to external companies.

And last July we announced that we planned to ramp up our Yandex.Cloud initiative beyond the experimental phase.  Today in a press release, Yandex announced that it has advanced Yandex.Cloud to provide companies with low-cost and flexible IT solutions that include scalable virtual infrastructure with multiple management options, automated services for the labour-intensive management tasks of popular databases systems, and AI-based Yandex services.  Yandex.Cloud also offers flexible pricing to suit different companies’ diverse needs and is deployed across multiple local data centers, enabling companies to store and use databases containing personal data in Russia as required by law.

As part of the development of Yandex.Cloud, Yandex invited over 50 leading Russian and international companies to test the service and provide feedback on the platform. Since April 2018, companies have included industry leaders such as Tinkoff Bank, S7 Airlines and X5 Retail Group. Other companies that shared their feedback included Bitrix24, which offers CRM as a service, English learning platform SkyEng, and call center voice platform Voxiplant.

“We launched Bitrix24 in 2012 having chosen to use a cloud infrastructure right at the outset. This was what let us quickly scale up or down. Starting from 2015, the Russian segment of Bitrix24 is hosted in Russia to comply with the legal requirements. None of the leading global providers of cloud infrastructure run their data centers in Russia. This is why we are closely watching local market players and support the development of Russia's cloud platforms. Having participated in Yandex.Cloud's beta testing, we can appreciate the seriously systematic approach even at this point. This approach will help the company build a genuinely geographically distributed cloud platform with a full range of services,” says Alexander Demidov, head of cloud services at Bitrix24.

“Our product lets hundreds of thousands of people make progress in learning English. It is crucial for us that a cloud platform provides a stable and reliable service, as this is something that directly affects our customer experience. We have tested Yandex.Cloud and can admit that even at this point it satisfies the needs of a part of our services. We are positive that the new cloud platform will soon meet our expectations regarding our other services and challenges as well,” says Sergey Safonov, head of software development at SkyEng.

“We have a long history with Yandex, especially using their SpeechKit technology in our Voximplant communication app platform. Our plans are to move our development and testing environment to Yandex.Cloud and use this platform for hosting some of our services aimed at our Russian customers,” says Andrey Kovalenko, CTO at Voximplant.

Thanks to the many companies who have helped test Yandex.Cloud, we have been able to build a service that meets the cloud-computing needs of the local market.  We will continue product development by adding new infrastructure services, additional services based on Yandex’s machine intelligence and analytics tools, and developer tools intended to simplify the entire code lifecycle management within Yandex.Cloud.

Access to Yandex.Cloud is currently available on request and will be open to all by the end of 2018.

Yandex Launches the First Autonomous Ride-Hailing Service in Europe

Today we are thrilled to announce that Yandex.Taxi’s self-driving car is now offering a ride-hailing service to passengers in the university city of Innopolis, Russia.  As the first service to offer autonomous ride-hailing in Europe, we are not only excited to help users safely and efficiently navigate to destinations but also to help the public adopt driverless technology and the future of transportation. 

Residents can order rides to multiple set destinations such as the university, the stadium, residential blocks, and the local business center.  Currently, approximately 100 passengers have agreed to participate in the testing of our driverless ride-hailing service and received a link to order rides through a Telegram chatbot.

The young city, centered around the tech-focused Innopolis University, offers Yandex a great urban environment to take the next significant step in publicly testing our self-driving cars that combines our expertise operating on busy public streets and our experience offering passenger rides on test tracks.  

In Innopolis, Yandex will first operate two self-driving vehicles to specific destinations within the city free of charge and keep a safety engineer in the passenger seat.  Our Yandex self-driving team plans to later expand the autonomous ride-hailing service to include more destinations, additional vehicles, and removing the safety driver in addition to improving the service based on user feedback.  

Yandex.Drive Hits 2 Million Rides, Doubles Its Fleet, & Announces Refueling Trucks

Earlier this year Yandex expanded its transportation services and efforts to improve urban living by launching its first car sharing service, Yandex.Drive, becoming one of the largest services in Moscow and surpassing over 2 million rides within the first six months of operation. Yandex.Drive, which offers drivers on-demand access to a fleet of vehicles through a mobile application, has since added a business class option and yesterday announced a refueling service for its fleet.

Starting in February 2018 Yandex.Drive offered users in Moscow and the surrounding region access to a fleet of 750 vehicles that included Kia Rios, Kia Rio X-Lines or Renault Kapturs for as low as five roubles ($0.09) per minute. Within months of this launch Yandex.Drive has more than doubled its fleet to over two thousand vehicles to meet consumer demand.

Recently Yandex.Drive added Audi and VW cars to the fleet and started offering qualified drivers access to 500 business class vehicles that include Mercedes E-Class and BMW 5 series vehicles.  The service also added new premium options that provide drivers access to crossover Porsche Macan vehicles and Porsche 911 Carrera 4S vehicles.

Drivers with a safe driving record over the age of 26 who have been licensed for a minimum of six years will have access to business class vehicles at a rate of 20 rubles per minute ($0.32)  for Porches and 10 rubles per minute ($0.16) for Mercedes and BMW vehicles.  Yandex.Drive rides are billed on a per minute basis and Yandex integrates traffic conditions and customer demand to provide dynamic pricing at the time of reservation. The price includes fuel, parking, liability and life insurance.

Moscow is an incredibly congested and challenging city in which to own a vehicle. This has led it to become the largest car-sharing market in Europe based on the overall available size of the fleet of vehicles. By offering users a viable alternative to car ownership through an accessible car sharing service that includes personalisation through the Yandex.Auto in-car infotainment system associated with each user’s account, plus a range of vehicle choices, Yandex.Drive has become one of the top services in the market.

Yandex also announced yesterday that it has plans to purchase a fleet of refueling trucks from Toplivovbak to integrate a more efficient refueling feature for Yandex.Drive cars.  The refueling trucks will not only help drive overall costs down but also serve to benefit drivers who need to refuel when taking long trips with their Yandex.Drive car.

Refueling Yandex.Drive vehicles will be done automatically using algorithms that gauge the fuel levels in vehicles and send fueling trucks to nearby cars.  By automating the process and monitoring of fuel levels, users won’t need to end their rides early or switch to another vehicle in order to refuel.

“We aim to provide our users with an efficient, convenient, and comfortable car sharing experience.  Over the past six months, the expansion of Yandex.Drive has led to an even better experience for our users that ranges from more vehicle choices to new app features. Optimizing the refueling process for our fleet of vehicles will further enhance the car sharing service we offer users and we look forward to further expansion and innovation.“ says Anton Ryazanov, Head of Yandex.Drive.

Users can start using Yandex.Drive by downloading the application for Android and iOS.

Expanding the Yandex Intelligent Assistant Ecosystem

Today at our annual Yet Another Conference (YAC) we announced major expansions to the Yandex intelligent assistant ecosystem that includes our new smart speaker, a new skills platform for our intelligent assistant Alice, and a new bundled subscription service.  Combined these developments offer the Russian audience more choice and the most advanced ability to navigate the online and offline world to date.

Over the years we have provided the Russian audience with many firsts and many highly localized smart products and services. And over time many of these products and services have come to operate together through our Search app to offer a more fluid user-experience. Last fall we enhanced that experience by introducing our Russian-speaking intelligent assistant Alice - first to mobile through our Search and Browser apps, then to desktops through Yandex Browser, and then to cars through our navigation app, Yandex.Navigator and the forthcoming integration in our Yandex.Auto connected car platform.

Building on our success helping users with tens of millions of daily interactions on Alice on multiple platforms, we are excited to welcome Alice to Russian homes in addition to building out more capabilities and benefits for users of our smart speaker and services.

A new platform & our first hardware launch

Our new smart speaker, Yandex.Station, is not only our first hardware product but it is the first smart speaker for the Russian market. With the introduction of Yandex.Station, Russian users now have the ability to rely on Alice and voice commands to manage a number of routine tasks in their homes.

Yandex.Station is the first smart speaker to incorporate a full video streaming experience.  A HDMI output allows Yandex.Station users to ask Alice to search for and playback videos, movies and television shows from a wide variety of sources.  Currently, Yandex.Station has access to Yandex's video streaming service KinoPoisk, as well as a wide range of other content including premium local video service providers ivi and Amediateka.

Yandex.Station will be available for purchase for 9,990 roubles ($160) this summer.

More collaboration and capabilities for our intelligent assistant

To further expand Alice’s functionality on Yandex.Station and all other platforms, we created a developers skills platform for Alice called Yandex.Dialogues.  This platform is designed to make it easy for any 3rd party developer to create new skills for Alice, ranging from a beginner programmer working on their first game to a major brand that can further enhance their customer experiences, such as a restaurant chain taking food orders.

Through Yandex.Dialogues developers can empower Alice to create custom conversations with users.  Developers choose a custom hotword that a potential customer can then use to activate a custom chat, game or whatever service has been developed. For example, a customer can say, “Alice, launch the Papa John’s skill” to activate Papa John’s skill that handles their pizza ordering.  Developers can also uniquely incorporate a live-chat feature to manage customer service that can include chatbots and live human operators to provide the best possible user experience.

The Dialogues chat functionality will also be available on Yandex’s search results page to further expand their reach.

Yandex.Dialogues is mutually beneficial for both our users and businesses. Users gain more choice and more efficient interactions with companies, while businesses have an innovative opportunity to reach and serve new and existing audiences.

Thanks to the creativity of hundreds of developers during the three-month beta testing period for Dialogues, Alice now allows users to interact with some of their favorite brands.   

Alice has the capability to help users order a pizza from Papa John’s, groceries from Utkonos and flowers from Flowwow.  Alice can also help users check MegaFon mobile phone account balances, exercise with Reebok, select air tickets from S7, search for a job with HeadHunter, study English with SkyEng, browse a McDonald’s menu, and consult with a Sberbank bot.  With Dialogues now publically available we look forward to all of the new skills that will help Alice serve users’ needs better.

New benefits and options for our users  

When it comes to offering more benefits for our products and services, we launched a new bundled subscription service for Yandex products, Yandex.Plus,that provides advanced opportunities to utilize the Yandex ecosystem.

Owners of the new Yandex.Station will receive a complimentary one-year subscription to Yandex.Plus, which regularly costs 169 roubles ($2.75) per month, and includes:

  • Complete access to Yandex.Music, including unlimited downloads to mobile devices and high-quality music streaming.
  • A 10% discount for Yandex.Taxi rides in the Comfort, Comfort+ and Business categories.
  • Free delivery for customers in over 30 locations in Russia spending over 500 roubles on Yandex.Market's new marketplace recently launched in beta.
  • A 5% discount for any Yandex.Drive ride with an extended car reservation period.
  • Ad-free viewing of films and television series on the web-based version of KinoPoisk.
  • Additional 10GB of free storage space on Yandex.Disk.

As we reflect on these important expansions to the Yandex intelligent ecosystem, we are excited by the advancements and opportunities this creates for the Russian audience and the entire global technology community.  

An update on our intelligent assistant Alice

Last fall we launched our intelligent assistant Alice, providing the Russian market with an assistant equipped with high-quality conversational abilities and advanced understanding of context. We were especially excited to provide users with this unique interactive feature and we were humbled by the overwhelming positive feedback we received.  Today Alice engages in tens of millions of interactions every day to millions of daily active users.

While we have been thrilled to see the adoption (plus all our users’ funny memes and screenshots), we were also eager to keep Alice on track in advancing its development as a centralized tool that integrates multiple useful services, which started with Search, Weather, News, Maps, Transport, and Music.  Over the course of Alice’s first six months in production, we introduced several new features and capabilities to provide users with even more assistance.  We shared those updates to our Russian users as they happened but wanted to provide a recap to our English audience before we enter the next phase of development and expand Alice’s capabilities even further:

  • Integration with Yandex.Navigator: Alice is now helping drivers stay safer on the roads through integration with Yandex.Navigator, Yandex’s navigation system that serves over 18 million drivers every month.  With the help of Alice and advanced speech technologies, drivers can ask Alice to help provide routes to destinations, locate the nearest gas station, and find open parking spots among other features.  While Yandex.Navigator previously handled specific voice commands, the integration with Alice has enhanced the service to offer more natural interaction on relevant navigation topics.
  • Integration with Yandex.Taxi: Through integration with our market-leading Yandex.Taxi app, Alice can also help book taxi orders by checking the address or name of the desired organization and then launching the Yandex.Taxi order page with all relevant information. After a user makes the request, he or she just needs to hit the order button once Alice has loaded all relevant information.
  • Interactive Games: Alice offers users a chance to test their knowledge and entertain themselves with over 16 interactive games.  Alice’s games help challenge users with their knowledge of various topics through trivia style games such as This day in history, Guess the Animal, and Fact or Fiction.
  • Integration with Yandex Browser for desktop: Through integration with Yandex Browser, Alice offers users a new way to interact with their computer through voice.  In addition to other standard functions offered through mobile, Alice provides assistance with operations in both the browser and the computer. Alice can help open a favorite YouTube channel, play a song and change the volume in addition to other options such as launching Word or shutting down the computer.
  • Math and currency conversions: Alice helps users quickly solve their advanced banking and math questions through voice commands. Alice computes math questions in addition to helping with currency conversions.
  • Timer and Alarm: Alice helps users stay on schedule with the timer and alarm functions that are integrated into Android.  Whether it is setting a timer while cooking or a reminder, Alice users can simply ask Alice to help them stay organized.

As part of this process to expand Alice’s capabilities and ecosystem even further, we introduced a beta version of Yandex.Dialogues, which provides developers the opportunity to develop new skills for Alice.  We believe that some of the best products come from collaboration and diversity and we look forward to Alice’s growth through this initiative.  We are excited to share more news on the development of Alice in the coming future!

Ascend to the World’s Highest Peak through Yandex’s Everest Panorama Project

In 2016, the Yandex panoramic maps team undertook a major project with Toyota of Russia to create the world’s first ever panoramic climb up Mt. Everest to enable users to experience one of the world’s most challenging climbs and highest peak through a life-like experience. After two attempts at capturing images and a few years of physical and technical work, the team launched the Yandex Everest page, where people can experience the same climb mountaineers made from every existing Everest base camp.  The panoramic climb offers viewers an interactive movie-like experience up to the peak with commentary on the history and facts on Everest.

Our team committed to this project with a mission to expand users’ horizons by taking them on a journey most will never experience.  In attempts to reach their mission, our team members became an important part of the modern history of Mt.Everest.  In their first efforts to capture images from the summit, the team experienced and survived the 2015 earthquake in Nepal, later returning in 2016 even more inspired to make the climb and create the first panoramic trek up Everest.

We are proud to have achieved our mission and to have received such incredible feedback not only from users but also the design and Internet community.  We are excited to announce that the project is nominated for a 2018 Webby Award, the leading international award honoring excellence on the Internet.  Since the project was published, it was also a silver winner at the 2017 Epica Awards, a 2017 Red Dot Award winner in Communication Design, a 2017 Golden Drum finalist, a 2017 ADCE finalist, a 2017 ADCR 2017 gold and silver winner, and a 2016 Red Apple 2016 finalist.  

“We are incredibly honored and humbled to be nominated for the 2018 Webby and won several other awards over the past year amongst other inspiring work on the Internet,” says Anatoly Topol, Special Projects Director, “For us, it is especially gratifying to receive such positive feedback from the community because our journey to create the Yandex Everest Project was a truly demanding challenge that represents our passion for creating innovative experiences for users - even those that can seem impossible at times.”

We welcome you to join the millions of users who have taken the virtual climb up Everest on our project page and to cast your vote for the Webby Awards until April 19th, 2018.  

New AI Driver Identification System for Yandex.Taxi

Every day our ridesharing service Yandex.Taxi provides millions of rides to and from work, friends’ and families’ homes, classes, the gym, and many other locations.  This past January alone our thousands of drivers provided over 62 million rides to users of the service. While helping facilitate millions of rides, we are also committed to making each and every ride a positive experience for both our drivers and passengers. One of the most important aspects of providing a high-quality user experience is ensuring passenger safety.

As part of our continued commitment to safety, Yandex.Taxi recently began testing a two-step driver identification system powered by our computer vision and voice technologies that requires drivers to send a selfie and a recording of their voice through the app during each driving shift.  We developed the two-step driver identification system for the driver side application for Yandex.Taxi as an important safety check that not only verifies our drivers identity to ensure passenger safety but also serves to protect drivers’ data.

Yandex.Taxi drivers are prompted to take and submit a selfie in addition to recording themselves reading a provided text.  This data combination createss a unique key for each driver and verifies their identity with more than 90% accuracy.  After the driver is positively identified they are able to receive ride orders. 

If a driver’s appearance changes – say by weight change or facial hair growth – the algorithms automatically update and factor in differences in appearance while using historical data to confirm it is the same driver. Yandex’s SpeechKit voice technologies are also able to identify a driver when there is a variation in voice tone or pitch.

In addition to the verification of drivers, Yandex.Taxi routinely runs vehicle inspections by requiring drivers to send photos of both the interior and exterior of their cars.  We also welcome passenger feedback on the condition of the vehicle and interactions with drivers.  We consider excellent vehicle condition and cleanliness to also be important aspects of Yandex.Taxi’s commitment to ensuring passenger and driver safety.  

Yandex.Taxi is piloting the AI driver identification system in Russia and shortly, it will expand to all other countries in which our service operates.  Currently, drivers are asked to submit a selfie and record their voice periodically throughout their driving shift when they aren’t active on the roads. In the future, the system will operate at the start of each driver’s shift.