Let's Build
the AI Future
Together

Samsung R&D Institute Poland

Artificial Intelligence Team

As one of the largest AI teams in Europe, we work with the analysis of sound, image, speech, text, and other data used, for instance, in creating recommendation systems and indoor navigation solutions. Our visual computing algorithms are used in cameras, augmented reality (AR) and visual content analysis. We are one of the leading centers specializing in natural language processing (NLP), in particular: machine translation, dialogue systems (including chatbots), and text analytics in its broad sense.

We use eXplainable AI to better understand our models and to deliver our solutions in compliance with the ethics of AI (fairness, accountability, transparency). Our work is based on the state-of-the-art machine learning methods (including deep neural networks), extensive data resources, and adequate computing power (NPU, GPU, DSP).

We work on the equipment of the latest generation. In our research we use advanced microphone matrices, depth and RGB cameras, and laser projectors. We create large-scale cloud systems as well as lightweight on-device applications, optimized for the limited computing power of mobile devices.

Our ready-to-use solutions are released on Samsung devices used by hundreds of millions of people all over the world. Our ideas win AI competitions and feature in prestigious research conferences.

Join us, let’s build the future of the Artificial Intelligence together!

Natural language processing

Natural language processing

Human language understanding and processing are remarkably difficult for computers. This is because natural languages are not only complex, but also imprecise, ambiguous and nuanced.

The same word may mean something completely different, depending on its context. The same thought may be expressed in many different ways. Some expressions need to be understood metaphorically, otherwise they make no sense.

We are a large team of researchers, engineers and linguists working on natural language processing (NLP), in such areas as: question answering, machine translation, dialogue systems (including chatbots), text mining and many others. Using the newest approaches, (e.g. deep neural networks) we create solutions that are able to grasp the nuances of the natural language.

Sound

Sound

Sound brings joy to our life. Audio signal is very informative.

We believe that next-generation devices will react to the acoustic context just like people do. At Samsung R&D Center, we develop an advanced technology that takes the raw waveform and either transforms it into meaningful information, or enhances it to create new experiences. The technology recognizes sounds and matches them with visual clues.

Our work usually starts with the raw audio signal captured by microphones. Next, we use the newest audio processing algorithms to clean it and prepare it for the AI magic. We use state-of-the-art methods, backed by hundreds of GPUs, to create deep neural networks tailored for the audio domain. We like to think that we teach machines to hear.

Visual computing

Visual computing & AR

How about we taught the machines to see the world? That is what we actually do. We equip our products with these astonishing skills. To achieve it, we apply the state-of-the-art sensors and build platforms for spatial perception.

Our engineers create advanced solutions that make Samsung excel both in the popular area of digital photography, and in the new, exciting world of augmented reality (AR). We create technologies that enable our devices to see – they are aware of their spatial, physical environment, as well as have semantic understanding of the space around them.

With enthusiasm, we work on virtual scene rendering – our technology can seamlessly mix real and artificial objects. Our engineers take the best out of classic algorithms and AI, make them work on-device and in real time, and thus deliver powerful solutions to the market.

Recommendation systems

Recommendation systems

Recommendation systems are ubiquitous, however most often people use them unconsciously.

Imagine browsing through thousands of products in a web store, looking for a specific item. The website magically discovers your preferences for certain product features and selects products that are expected to attract your attention. At Samsung R&D Center, we go beyond such classic recommendation systems.

We employ deep learning techniques and a broad context to achieve the highest accuracy – in the end, who wants a pair of shoes when looking for a book? Our systems need to work efficiently for hundreds of millions of users, so you will meet Big Data here as well.

Explainable AI (XAI)

Explainable AI (XAI)

Artificial Intelligence is frequently associated with black magic. In fact, it can be easily explained and understood.

Explainable AI is an emerging domain of machine learning which aims at solving the low interpretability problem on decisions taken by AI systems. In our projects, we try to find the answers to some vital questions, such as: Why has the system made some specific decision or forecast? Did the system succeed or fail? When do AI systems have enough decision-making credibility to trust them?

We are constantly seeking the answer to the question: what should be done to make the AI systems work better? We are aware that development of high-class AI systems is a tough, time-consuming task. That is why we use the XAI tools to reduce the time dedicated to develop an AI system. XAI tools indicate the most suitable data to train the right model and tell us why. It also provides information on the features, weights and layers that were used by the model to make the prediction.

Join our AI team

See our job offers

Our office

Warszawa

Warsaw Spire
Plac Europejski 1
00-844 Warszawa