Nanosemantics, a leading developer of artificial intelligence (AI) technologies and a resident of the Information Technologies Cluster of the Skolkovo Foundation, has launched the NLab Marker service. It transforms data into information that can be understood by neural networks.
NLab Marker is an industrial platform ready for implementation and operation on tasks with large amounts of data that require special manual processing: markup and training by specialists of training examples for machine learning algorithms. NLab Marker allows you to highlight objects on video, transcribe audio recordings, mark up medical images. The service minimizes the work time of data preparation specialists (markers) and the number of errors when forming a set of training data (dataset).
Stanislav Ashmanov, CEO of Nanosemantics: “Machine learning is impossible without training data - examples by which algorithms learn. We, as the developers of AI algorithms, know how important high-quality labeled data is. Our team has developed the NLab Marker platform for ourselves. But now we are ready to offer this product to the market, as we see the demand for industrial data markup platforms that allow flexible implementation of any markup tasks and independently administer this process for companies with a strong Data Science department. ”
Errors in the dataset critically affect the quality of neural network training. For example, a trained neural network for video analytics will miss a defect on a production line or incorrectly transfer personal data from a completed questionnaire to the MFC. NLab Marker has implemented a system of automatic checking of markers operation using trap examples (honeypots). The time and volume of completed tasks are also monitored. In addition, various modules for working with text and audio are built into NLab Marker. For example, a module for speakers allows them to record audio for speech synthesis, and a categorization module allows you to assign a category to a specific text.
Unlike analogs, NLab Marker guarantees high accuracy of data preparation, quality control at all stages of the marking process and protection of personal data, since the platform can be deployed in a secure customer loop. Also, NLab Marker allows you to clean personal data from the finished dataset so that it does not get publicly available.
Pavel Krivozubov, Head of Robotics and Artificial Intelligence, Skolkovo Foundation: “NLab Marker has developed a convenient system for organizing and managing a marking team: the curator distributes tasks and instructions to project managers or executors, and also sets individual deadlines for completing tasks for a specific project. This saves time and money for the company. The service allows residents of the most remote regions to work, the unemployed, people with disabilities. It does not matter where in the world the stripper is located, the main thing is that he has a computer and access to the Internet. This is especially important now, because the trend is remote work with flexible hours without being tied to the workplace.”
The Skolkovo Foundation is a non-profit organization established in 2010 and acting as the managing company of the Skolkovo Innovation Center. The purpose of the Foundation is to support technological entrepreneurship in Russia and commercialize the results of research activities. The number of Skolkovo residents has already approached 2.5 thousand. The Fund provides them with a range of services necessary to bring developments to the Russian and international markets. Resident status gives the company tax and customs benefits. In 2019, the residents' revenue amounted to more than 100 billion rubles, and the amount of investments attracted by them exceeded 13 billion rubles. Since 2019, the law allows any organization engaged in research activities in Russia to gain access to all the services and benefits of the Innovation Center. The Skolkovo Innovation Center is home to the Skolkovo Institute of Science and Technology (Skoltech), which is one of the top 100 young universities in the world according to Nature Index. The infrastructure of the Innovation Center is made up, in particular, of Europe's largest Technopark and research centers of industrial partners. In 2019, three new partner R&D centers were opened in the Innovation Center - Sibur, TMK and Tatneft. On the territory of Skolkovo, a total of more than 750 thousand square meters have already been commissioned. m. residential areas, research and educational centers. Off-budget investments in infrastructure creation exceeded RUB 100 billion.