Patents by Inventor Serg Bell

Serg Bell has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).

  • Publication number: 20260350029
    Abstract: Example embodiments are directed to taxonomy-based knowledge graph construction. A document is received via a user interface. A plurality of keywords is extracted from the document using a large language model based on user context data. A subset of keywords is identified from the plurality of keywords as being relevant to user-provided data entries. A taxonomy for each keyword in the subset of keywords is imported from the user-provided data entries. The subset of keywords and corresponding taxonomy data are stored as a connected graph in a database.
    Type: Application
    Filed: July 8, 2026
    Publication date: November 5, 2026
    Inventors: Vasyl Shandyba, Sergey Ulasen, Andrey Adashchik, Serg Bell, Stanislav Protasov, Nikolay Dobrovolskiy, Laurent Dedenis
  • Publication number: 20260351383
    Abstract: A device includes a plurality of ingredient interfaces each couplable to a replaceable container. The device includes a hydraulic system for pumping liquids with metering pumps. The device includes an electronic controller configured to: receive an instruction to prepare a mixture based on a general recipe, detect one or more replaceable containers currently coupled to at least one of the plurality of ingredient interfaces, determine an ingredient identifier list associated with the one or more replaceable containers and associated properties of ingredients in the ingredient identifier list; and in response to the instruction, actuate selected metering pumps of the hydraulic system to deliver ingredient volumes from the one or more replaceable containers in accordance with the general recipe and initiate a rinsing flow through the hydraulic system toward an outlet.
    Type: Application
    Filed: January 28, 2026
    Publication date: November 5, 2026
    Inventors: Vladimir SITNIKOV, Serg BELL
  • Patent number: 12783296
    Abstract: Disclosed herein are systems and method for performing network accelerated scheduling. In one aspect, a method may receive, at a programmable switch comprising a scheduler, a job comprising a plurality of tasks from a client. The method may store, on the programmable switch, the plurality of tasks in a queue. The method may receive an indication from a first executor on a working node that the first executor is available for task execution. The method may schedule, using a scheduling policy, at least one task in the queue to be performed by the first executor. The method may transmit the at least one task to the first executor. The method may receive a task completion indication from the first executor.
    Type: Grant
    Filed: October 5, 2023
    Date of Patent: November 3, 2026
    Assignee: Acronis International GmbH
    Inventors: Sreeharsha Udayashankar, Samer Al-Kiswany, Serg Bell, Stanislav Protasov
  • Patent number: 12783648
    Abstract: Systems and methods for mitigating potential security incidents. A system includes a data ingestion module, a graphical user interface (GUI), a generative AI model, an enrichment module, and a mitigation module. The generative AI model is pretrained on a large language model (LLM) using a dataset of known security incidents and the computer system's infrastructure, and analyses potential security incidents and generates incident overviews, leveraging its understanding of attack frameworks and previous incident data. The enrichment module incorporates user interactions, enhancing the incident overviews with accurate information. The mitigation module proposes mitigation actions based on the generative AI model's insights gained from prior incidents. The system enables natural language interaction through the GUI and provides graphical representations of the incidents.
    Type: Grant
    Filed: September 15, 2023
    Date of Patent: November 3, 2026
    Assignee: Acronis International GmbH
    Inventors: Candid Wuest, Serg Bell, Stanislav Protasov
  • Patent number: 12784119
    Abstract: Disclosed herein are systems and methods for a machine-learning based method for assessing academic performance of a learner and providing learning recommendations. The method includes obtaining assessment materials for a learner. The method also includes analyzing answers using a trained test assessment LLM. The method includes analyzing the test feedback using a trained course assessment LLM configured to evaluate a proficiency with the course materials and to generate a course feedback comprising expected answers and indicating a level of proficiency with the course materials. The method includes analyzing the course feedback to evaluate a proficiency with topics of study based on the course feedback for given topics and the past academic performance data of the learner, and to generate a topic feedback. The method further includes generating a plurality of personalized learning recommendations for topics of study in which the learner has a level of proficiency below a proficiency threshold.
    Type: Grant
    Filed: December 17, 2024
    Date of Patent: November 3, 2026
    Assignee: Constructor Technology AG
    Inventors: Rasilia Rakhmatulina, Andrey Adaschik, Victor Batraev, Serg Bell, Stanislav Protasov, Nikolay Dobrovolskiy, Laurent Dedenis
  • Patent number: 12783650
    Abstract: Systems and methods for securing and managing API calls. A security dynamic link library (DLL) is a user mode component that intercepts API calls from an API client to an API server. A security kernel driver is a kernel mode component that hooks API calls from the security DLL to the API server. A protection service receives, analyzes, and processes API calls from the security DLL and the security kernel driver.
    Type: Grant
    Filed: March 22, 2024
    Date of Patent: November 3, 2026
    Assignee: Acronis International GmbH
    Inventors: Vladimir Strogov, Alexey Kostyushko, Aliaksei Dodz, Anastasia Pereberina, Serg Bell, Stanislav Protasov
  • Publication number: 20260340007
    Abstract: A system verifies a plurality of documents in the knowledge database by determining verification information comprising at least one of a source identifier and a creation time of each respective document of the plurality of documents. The system stores the verification information in the knowledge database. The system receives a user query for the machine learning model and identifies, in the knowledge database, two or more statements related to the user query. In response to detecting conflicting factual data in the two or more statements, the system assigns a respective comprehension level to each of the two or more statements based at least on respective verification information associated with each of the documents. The system provides at least the conflicting factual data and the respective comprehension level of each of the two or more statements to the machine learning model for generating a user query response.
    Type: Application
    Filed: April 29, 2025
    Publication date: October 29, 2026
    Inventors: Dmitrii Istomin, Alexander Tormasov, Laurent Dedenis, Stanislav Protasov, Serg Bell, Nikolay Dobrovolskiy
  • Publication number: 20260339920
    Abstract: A system generates the knowledge database using a plurality of documents. The system verifies the plurality of documents in the knowledge database by determining verification information comprising at least one of a source identifier and a creation time of each respective document of the plurality of documents, detecting conflicting factual data in two or more statements of the knowledge database. In response to detecting the conflicting factual data, the system assigns a respective comprehension level to each of the two or more statements based at least on respective verification information associated with each document corresponding to the two or more statements; and adjusts the knowledge database based on a comparison of the respective comprehension level of each of the two or more statements.
    Type: Application
    Filed: April 29, 2025
    Publication date: October 29, 2026
    Inventors: Dmitrii Istomin, Alexander Tormasov, Laurent Dedenis, Stanislav Protasov, Serg Bell, Nikolay Dobrovolskiy
  • Patent number: 12775654
    Abstract: A system receives a plurality of scientific documents for inclusion in an extended knowledge graph. For each respective scientific document of the plurality of scientific documents, the system classifies the respective scientific document with a theoretical framework of a plurality of theoretical frameworks each comprising terms and principles associated with a particular scientific topic, extracts metainformation of the respective scientific document, structures the metainformation in a document-specific ontology model that further comprises an indication of the theoretical framework, generates a plurality of text chunks from the respective scientific document of a given size, and generates, using a first ML model, one or more concepts from each of the plurality of text chunks. The system generates, using a second ML model, the extended knowledge graph using each of the plurality of text chunks, each concept, and the metainformation, and stores the extended knowledge graph in a graph document database.
    Type: Grant
    Filed: February 10, 2025
    Date of Patent: October 27, 2026
    Assignee: Constructor Technology AG
    Inventors: Andrey Ustyuzhanin, Alexander Tormasov, Serg Bell, Stanislav Protasov, Artem Maevskiy, Sergey Ulasen, Vasyl Shandyba
  • Publication number: 20260329972
    Abstract: Disclosed herein are systems and methods for providing large language model (LLM) embeddings. The method includes receiving an input query to a LLM service from a user while a main embedding model is unavailable. The method also includes generating an embedding of the input query using a fallback embedding model. The method further includes comparing the embedding of the input query to a plurality of secondary centroid embeddings of a plurality of subject matter clusters, each subject matter cluster comprising a secondary centroid embedding and a corresponding main centroid embedding. The method further includes determining that a similarity between the embedding of the input query and the secondary centroid embedding of a subject matter cluster satisfies a cluster similarity threshold. The method further includes transmitting a main centroid embedding of the subject matter cluster that satisfies the cluster similarity threshold to the LLM service.
    Type: Application
    Filed: July 2, 2026
    Publication date: October 22, 2026
    Inventors: Sergey ULASEN, Alexander TORMASOV, Serg BELL, Stanislav PROTASOV, Nikolay DOBROVOLSKIY, Laurent DEDENIS
  • Publication number: 20260329870
    Abstract: In response to determining that an input query is not answered by a user responder within a threshold period of time: a system tags the input query as an input for a response generation ML model; generates a score representing a likelihood of the response generation ML model outputting a correct response to the input query based on characteristics of the input query and a knowledge base; executes the response generation ML model to answer the input query based on the score exceeding a threshold query evaluation score; and outputs the response for display on a computing device based on the score exceeding a threshold response evaluation score. In response to determining that the input query is answered by the user responder within the threshold period of time, the system executes the response generation ML model to generates and outputs a supplemental response based on a user response.
    Type: Application
    Filed: April 18, 2025
    Publication date: October 22, 2026
    Inventors: Alexander Tormasov, Serg Bell, Stanislav Protasov, Nikolay Dobrovolskiy, Laurent Dedenis
  • Publication number: 20260329425
    Abstract: A system identifies a plurality of components in a computational chain model selected by a user for editing on a user interface. The system generates, for display on a computing device, the user interface comprising a plurality of visual objects corresponding to the plurality of components, wherein the plurality of visual objects are visually connected, and wherein at least one visual object includes configurable parameters. The system receives, via the user interface, a user request to configure a parameter associated with a visual object in the plurality of visual objects. The system modifies a chosen component of the computational chain model that corresponds to the visual object by modifying code of the chosen component that includes the parameter based on the user request. The system generates, for display on the computing device, an execution result of the computational chain model with the chosen component modified.
    Type: Application
    Filed: April 16, 2025
    Publication date: October 22, 2026
    Inventors: Alexander Tormasov, Laurent Dedenis, Stanislav Protasov, Serg Bell, Nikolay Dobrovolskiy
  • Publication number: 20260326785
    Abstract: A system collects a first plurality of real-time parameter values associated with an autonomous vehicle driving in a race track. The system generates, by a first controller of a plurality of controllers, at least one command for driving the autonomous vehicle based on the first plurality of real-time parameter values. In response to determining that an amount of time spent driving in a segment of the race track exceeds a threshold amount of time when the autonomous vehicle executes the at least one command, the system selects a second controller of the plurality of controllers to generate further commands for driving the autonomous vehicle. The system collects a second plurality of real-time parameter values, and generates, by the second controller of the plurality of controllers, at least one additional command for driving the autonomous vehicle based on the second plurality of real-time parameter values.
    Type: Application
    Filed: April 22, 2025
    Publication date: October 22, 2026
    Inventors: Maksim FILIPENKO, Aleksandr BUIVAL, Ruslan MUSTAFIN, Ilya SHIMCHIK, Stanislav Protasov, Serg Bell, Nikolay Dobrovolskiy
  • Publication number: 20260330987
    Abstract: Disclosed herein are systems and methods for machine-learning (ML)-based assessment of a presentation by a presenter and engagement of audience members. The method includes obtaining a presentation video stream, a presenter audio stream, and presentation material. The method also includes evaluating the presentation using a prepared presentation evaluation machine learning model (MLM) configured to generate a presentation score based on the presentation video stream, the presenter audio stream, and the presentation material. The method also includes obtaining a plurality of video streams capturing each respective audience member, a plurality of audio streams for each respective audience member, and a plurality of capture streams. The method also includes evaluating an engagement of at least one audience member using a prepared engagement evaluation MLM.
    Type: Application
    Filed: April 21, 2025
    Publication date: October 22, 2026
    Inventors: Svetlana Dergacheva, Nikita Zherebtsov, Alexander Tormasov, Laurent Dedenis, Stanislav Protasov, Serg Bell, Nikolay Dobrovolskiy
  • Publication number: 20260330852
    Abstract: Disclosed herein are systems and methods for associating digital images and text. The method includes: obtaining a digital image and text; analyzing the digital image using an image segmentation MLM trained to identify a plurality of distinct graphical elements in the digital image and to generate a plurality of image segments; splitting the text associated with the digital image using a text segmentation MLM into a plurality of text chunks; analyzing the plurality of image segments and the plurality of text chunks using the text segmentation MLM to identify semantic links between related text chunks and image segments that describe and depict, respectively, the same one or more district graphic elements; and storing into a multimodal database the digital image, the plurality of image segments, the plurality of text chunks, and the semantic links between the related text chunks and image segments.
    Type: Application
    Filed: April 18, 2025
    Publication date: October 22, 2026
    Inventors: Alexander Tormasov, Andrey Adashchik, Sergey Ulasen, Serg Bell, Stanislav Protasov, Nikolay Dobrovolskiy, Laurent Dedenis
  • Publication number: 20260330099
    Abstract: A system receives a user query to change a parameter in a pre-existing scientific model. The system executes a first machine learning model configured to detect the parameter in a code of the pre-existing scientific model, wherein the first machine learning model outputs at least one fragment of the code that is associated with the parameter. The system executes a second machine learning model configured to: receive both the user query and the at least one fragment of the code as inputs; generate a new part of the code for the at least one fragment, wherein the new part of the code changes the parameter in accordance with the user query; and modify the code of the pre-existing scientific model to include the new part of the code. The system outputs a modified version of the pre-existing scientific model with the modified code.
    Type: Application
    Filed: April 16, 2025
    Publication date: October 22, 2026
    Inventors: Alexander Tormasov, Laurent Dedenis, Stanislav Protasov, Serg Bell, Nikolay Dobrovolskiy
  • Publication number: 20260329549
    Abstract: A system receives a user input to create a pre-existing data processing model on a user interface. The system extracts information about the pre-existing data processing model from at least one information source. The system identifies, based on the information, a plurality of components in the pre-existing data processing model. The system generates a wrapper for each of the plurality of components, wherein the wrapper comprises configurable wrapper parameters mapped to internal configurable parameters of the pre-existing data processing model. The system generates, for display on a computing device, the user interface comprising a plurality of visual objects corresponding to the plurality of components. The system generates, for a visual object of the plurality of visual objects, a user interface element that enables a user to modify a wrapper parameter of a component corresponding to the visual object.
    Type: Application
    Filed: April 16, 2025
    Publication date: October 22, 2026
    Inventors: Alexander Tormasov, Laurent Dedenis, Stanislav Protasov, Serg Bell, Nikolay Dobrovolskiy
  • Publication number: 20260321153
    Abstract: Disclosed herein are systems and method for providing personalized learning content, comprising: receiving different types of learning materials related to a plurality of concepts or topics; receiving personal learning parameters of a plurality of learner users; dividing the learning materials into content chunks using a machine learning model, wherein each content chuck includes portions of different types of learning materials that are related to the same concept or topic; personalizing the chunks for each user, using the machine learning model, based on the personal learning parameters of each user; determining an importance level of each chunk of each user, using the machine learning model, based on the learning parameters of each user; selecting, for each user, a first personalized subset of the chunks based at least on the importance level of the chunks for that user; and providing the subset of chunks to each user.
    Type: Application
    Filed: April 10, 2025
    Publication date: October 15, 2026
    Inventors: Svetlana Dergacheva, Laurent Dedenis, Stanislav Protasov, Serg Bell, Nikolay Dobrovolskiy
  • Publication number: 20260320189
    Abstract: Disclosed herein are systems and methods for analyzing safety of software code associated with scientific or technical documents before execution of the software code. The method includes: receiving a scientific or technical document and software code associated with the scientific or technical document; analyzing the software code using a first prepared machine learning model (MLM) to predict a first computational parameters (e.g.
    Type: Application
    Filed: April 11, 2025
    Publication date: October 15, 2026
    Inventors: Alexander Tormasov, Ivan Tsyganov, Kevin Reed, Laurent Dedenis, Stanislav Protasov, Serg Bell, Nikolay Dobrovolskiy
  • Publication number: 20260320173
    Abstract: Disclosed herein are systems and methods for analyzing safety of software code associated with scientific or technical documents during execution of the software code in a pre-configured virtual environment.
    Type: Application
    Filed: April 14, 2025
    Publication date: October 15, 2026
    Inventors: Alexander Tormasov, Ivan Tsyganov, Kevin Reed, Laurent Dedenis, Stanislav Protasov, Serg Bell, Nikolay Dobrovolskiy