Abstract: Systems and methods for creating a mixed index in a computerized data storage system. The systems and methods configured to receive a file for input into a data storage system, generate one or more segments of data from the plurality of data of the file, store, in a first index in the data storage system, the one or more segments of data, compare the one or more segments of data to an electronic library to retrieve results in the library at or above a threshold of similarity, store the results in a second index in the data storage system, and generate a mixed index in the data storage system based on the first index and the second index, the mixed index comprising the one or more segments of data from the first index and semantically related terms from the second index.

Abstract: The invention relates to a method and a system for generating accurate speech-to-text transcriptions in communication sessions. The invention uses a hybrid approach leveraging the strengths of both, machine learning models and traditional NLP techniques. Thus, the present invention allows for maintaining transcription quality and context awareness while keeping compute costs reasonable and enables easy updates to domain-specific vocabulary without solely relying on complex model capabilities.

Abstract: An automated computer-implemented process can be executed by an authentication server for real-time authentication of a user. The process includes receiving user data about the user from a user device, comparing the user data with a user profile database to authenticate the user data, receiving a first image from the user device, and providing the first image and instructions for evaluating the first image to a multi-modal model. The process further includes generating a list of items detected in the first image using the multi-modal model and generating a context-specific challenge for the user based on the list of items detected in the first image. The process further includes sending the context-specific challenge to the user device with a request for the user to complete the context-specific challenge, receiving a second image from the user in response to the context-specific challenge, and verifying the second image with the multi-modal model.

Abstract: A computer-implemented method for switching communication mediums during a communication is disclosed that includes monitoring, by a signal strength engine of a carrier server, a signal strength of the communication between a first user device and a second user device; determining, by the carrier server, if the signal strength is subject to a signal loss or the reception is determined below the threshold; and switching by the carrier server, the communication from a first communication medium to a second communication medium.

Abstract: The present invention relates to a method for Conference Privacy Screening on Edge with Sensitive Aware Solution and a system for Conference Privacy Screening on Edge with Sensitive Aware Solution wherein the method comprises joining via a client or client node, a conference session held on a collaboration tool's backend of a server node by one or more users and capturing the audio stream of the conference session, by the client node and/or the server node. Then an Automatic Speech Recognition, ASR, model transcribes the audio stream of the conference session and sending the transcript to a transformer model. The transformer model evaluates if a sensitive topic occurs in the transcript and marks it with a flag in case a sensitive topic occurs. A Privacy Escalation Control Process analyzes then the flagged sensitive topic and confirms or not the flagged sensitive topic.

Abstract: A computerized method for converting an interactive voice response (IVR) tree into a natural language conversion architecture. The computerized method includes converting IVR tree audio data into IVR tree text data, the IVR tree audio data includes spoken language generated using the IVR tree during a phone call. The computerized method further includes providing the IVR tree text data to a machine learning model with instructions on how to evaluate the IVR tree text data. The machine learning model can generate a set of natural language prompts based upon the IVR tree text data. The set of natural language prompts can include instructions for directing a received call based upon caller voice input.

Abstract: A telecommunication apparatus and method can be configured to facilitate disconnection for conference calls. Embodiments can include the monitoring of an ongoing call (e.g. conference call or other type of call involving multiple parties), detecting a possible termination of the call based on a period of silence during the call for triggering an analysis of at least a portion of the call and evaluating the analysis results to determine whether the call is to be terminated or not. In the event termination occurs, resources can be unassigned to the call to free up resources. In the event it is determined that the call is not yet terminated, monitoring of the call can resume. In some embodiments, the conducted analysis can involve analyzing context specific aspects of the call (e.g. at least a portion of a transcript of the call, metadata related to the call, sound patterns, etc.).

Abstract: A method for protecting communications from sound leakages, and a corresponding system can be configured to offer a possibility to detect sound leakages and to alert participants in an online conference that information shared in the conference may be overheard by individuals not participating in the conversation but being in the environment of one participant. Thus, the privacy of end-users from sound leakages can be protected by taking advantage of multiple devices in the same physical space.

Abstract: A method for the secure and efficient distribution of documents in online conferences can integrate Speech-to-Text and AI technologies to allow a host to specify sections of a document for sharing, automatically adjusting access based on the clearance level of each participant. This can help enhance the efficiency and security of document sharing in virtual collaborative environments that can balance accessibility with confidentiality.

Abstract: A method and a system for remote meeting appointment priority overwrite can be configured to allow for scheduling higher priority remote meeting appointments although the invited meeting participant's calendar shows as busy. For instance, embodiments can provide such functionality by pre-empting and re-scheduling other already scheduled but lower priority meeting appointment to a later date. Embodiments can help prioritize scheduling so that important things are being scheduled, discussed, and decided first in relation to other lower priority things, which can provide enhanced efficiency as well economic and other benefits.

Abstract: A method and a system can be configured for meeting opening prediction to conference participants, while waiting in a conference call for the conference call to start (e.g. how long it will take until the moderator can open the meeting). In case of waiting users waiting in an isolated waiting room, the time until the users get connected to each other can be indicated for example as a countdown. The indicated countdown can allow waiting users to do other productive work while waiting and allows the moderator to better know how long he should wait until opening the meeting as likely the missing users will have joined by then. Also, if the conference has not yet been started (no one has joined yet), a system can be configured to automatically re-schedule a conference or provide conference participants with an indication of the meeting starting later.

Abstract: A method and system for collaboration session swap between at least two devices of a user and a system for performing said method can be configured to offer an efficient and secure method of unlocking a device and transferring a session, enhancing the overall user experience with a collaboration tool. In some embodiments, a session transfer request for transferring a collaboration session from a second device to a first device can be sent to facilitate transferring of the collaboration session after the first device is unlocked so a session can be seamlessly transferred to the first device.

Abstract: Methods and systems for secure validation of a machine learning model can facilitate encryption of a machine learning model and evaluating the performance of the model. Validation data can be encrypted and processed by the machine learning model so that results of the model's processing can be received and evaluated to evaluate how the machine learning model functioned.

Abstract: A computer-implemented method of detecting abnormal events in a telecommunication system can include a training phase and a production phase. The training phase can include recording data events in a log file during the execution of the system and creating a standard for the data events representing a properly working environment of the telecommunication system by applying a training procedure. The production phase can include receiving at least one new data event and determining whether the at least one new data event is a normal event that matches the standard or not. If the at least one new data event matches the standard, it can be used for updating the standard. If the at least one new data event does not match the standard, it can be classified and reported as at least one an abnormal event.

Abstract: A method for secure validation of machine learning models and parallel validation data using homomorphic encryption can include providing, by a provider, a machine learning model and providing, by a user, validation data; encrypting, by the provider, the machine learning model; sending, by the provider, a public encryption parameter to the user; selecting, by the user and provider, a unifying encoding method; encrypting, by the user, the validation data; sending, by the user, the encrypted validation data; processing, the encrypted validation data with the encrypted machine learning model; providing encrypted results of said processing to the provider and the user; and decrypting the results and evaluating whether the performance of the machine learning model is satisfactory with the given valuation data of the user.

Abstract: A computer-implemented method of detecting abnormal events in a telecommunication system can include a training phase and a production phase. The training phase can include recording data events in a log file during the execution of the system and creating a standard for the data events representing a properly working environment of the telecommunication system by applying a training procedure. The production phase can include receiving at least one new data event and determining whether the at least one new data event is a normal event that matches the standard or not. If the at least one new data event matches the standard, it can be used for updating the standard. If the at least one new data event does not match the standard, it can be classified and reported as at least one an abnormal event.