Method, computer system, and computer program for processing of digital content (automatic processing of meetings for confidential content)

Abstract

To provide a system for automatically producing post processing of meeting content through artificial intelligence (AI).SOLUTION: A method, a computer system and a computer program product for processing digital content are provided. The present invention may include building a sensitive sentence classification model. The present invention may include receiving digital content, where the digital content is intended for one or more groups of content consumers. The present invention may include processing the digital content using the sensitive sentence classification model. The present invention may include generating consumer-specific digital content for each of the one or more groups of content consumers.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates generally to the field of computing, and more specifically to digital content processing. [Background technology]

[0002] Due in part to remote work trends, many organizations may be conducting many of their most important meetings, public service announcements, and speeches, among other discussions, through online software that may enable digital content to be replayed by multiple consumers within the organization. Often, these meetings may be recorded, and portions of the meeting may only be viewable under certain conditions. Current approaches typically require post-processing of meeting content for each specific group of consumers, specific encryption devices to encrypt audio data, unique authentication with specific equipment, and / or encoding and / or decoding equipment to encode and / or decode the protected data. Current approaches may be inefficient and suffer from human error and / or institutional barriers that may reduce the number of consumers who can access digital content replay.

[0003] Therefore, a system for automatically producing post-processing of conference content through Artificial Intelligence (AI) is needed. More specifically, a system that can recognize sensitive data segments within digital content based on the roles of content producers and content consumer groups may be beneficial. Summary of the Invention [Problem to be solved by the invention]

[0004] A system is needed to automatically produce post-processing of meeting content through Artificial Intelligence (AI). [Means for solving the problem]

[0005] Embodiments of the present invention disclose methods, computer systems, and computer program products for digital content processing. The invention may comprise constructing a sensitive text classification model. The invention may comprise receiving digital content, where the digital content is targeted to one or more groups of content consumers. The invention may comprise processing the digital content with the sensitive text classification model. The invention may comprise generating consumer-specific digital content for each of the one or more groups of content consumers.

[0006] In another embodiment, the method may include training the sensitive sentence classification model using a customized training dataset and an existing dataset, where the customized training dataset is comprised of keywords and entities derived from organizational information and the existing dataset is comprised of one or more existing sensitive word datasets sourced from one or more publicly available resources, and the customized training dataset and the existing dataset are preprocessed using a Word2Vec algorithm.

[0007] In a further embodiment, the method may include storing the received digital content in its original form in an original content database, identifying one or more sensitive data segments using the sensitive text classification model, and storing the one or more sensitive data segments in a processed content database according to at least one or more of a data type, a timestamp, or a corresponding consumer role, where the corresponding consumer role includes at least one of one or more groups of content consumers.

[0008] In yet another embodiment, the method may include assigning a plurality of consumers within an organization to one or more groups of content consumers based on a hierarchical structure of the organization constructed based on an analysis of the organizational information, and displaying consumer-specific digital content to the plurality of consumers, where each of the plurality of consumers may access the consumer-specific digital content corresponding to their content consumer group.

[0009] In addition to methods, additional embodiments are directed to computer systems and computer program products for automatically post-processing digital content for sensitive data segments for different groups of content consumers.

[0010] This Summary is provided to introduce a selection of concepts in a simplified form that are further described below in the Detailed Description. This Summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used as an aid in determining the scope of the claimed subject matter. [Brief description of the drawings]

[0011] These and other objects, features and advantages of the present invention will become apparent from the following detailed description of illustrative embodiments thereof, which is to be read in connection with the accompanying drawings, in which the illustrations are for clarity in facilitating understanding of the invention by those skilled in the art in conjunction with the detailed description, and various features of the drawings are not drawn to scale. The drawings are as follows:

[0012] [Figure 1] FIG. 1 illustrates a block diagram of an exemplary computing environment in accordance with at least one embodiment.

[0013] [Diagram 2] 1 is an operational flowchart illustrating a process for digital content processing, according to at least one embodiment.

[0014] [Diagram 3] 3 shows a block diagram and flow chart illustrating a training process utilizing some of the components discussed in FIG. 2, according to at least one embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0015] Exemplary embodiments described below provide systems, methods, and program products for digital content processing. Thus, the embodiments have the ability to improve the technical fields of digital content processing and machine learning by automatically post-processing digital content for sensitive data segments for different groups of content consumers. More specifically, the invention may comprise building a sensitive text classification model. The invention may comprise receiving digital content, where the digital content is targeted to one or more groups of content consumers. The invention may comprise processing the digital content with the sensitive text classification model. The invention may comprise generating consumer-specific digital content for each of the one or more groups of content consumers.

[0016] As explained above, due in part to remote work trends, many organizations may be conducting many of their most important meetings, public service announcements, speeches, among other discussions, through online software that may enable digital content to be played by multiple consumers within the organization. Often, these meetings may be recorded and portions of the meeting may only be viewable under certain conditions. Current approaches typically require post-processing of meeting content for each specific group of consumers, specific encryption devices to encrypt audio data, unique authentication with specific equipment, and / or encoding and / or decoding equipment to encode and / or decode the protected data. Current approaches may be inefficient and suffer from human error and / or institutional barriers that may reduce the number of consumers who can access digital content playback.

[0017] Therefore, there is a need for a system for automatically producing post-processing of meeting content through Artificial Intelligence (AI), more specifically, a system that can recognize sensitive data segments within digital content based on the roles of content producers and content consumer groups.

[0018] Thus, it may be advantageous, inter alia, to build a sensitive text classification model and receive digital content, where the digital content is targeted to one or more groups of content consumers, process the digital content using the sensitive text classification model, and generate consumer-specific digital content for each of the one or more groups of content consumers.

[0019] In accordance with at least one embodiment, the present invention may improve the processing of digital content by providing a system that may recognize sensitive data segments within digital content (e.g., video and / or audio of a web conference) based specifically on the roles of content producers and content consumers (e.g., speakers and listeners).

[0020] In accordance with at least one embodiment, the present invention may improve post-processing of conference video for confidential content by retaining and / or masking sensitive data segments identified within reproduced digital content in a role-based manner.

[0021] In accordance with at least one embodiment, the present invention may improve monitoring of sensitive content in conferences conducted through online software by allowing content providers and / or speakers to optionally and / or manually trigger the marking of sensitive data segments using a Sensitive Sentence Classification Model (SSCM), in addition to automatic recognition.

[0022] In accordance with at least one embodiment, the present invention may improve the efficiency with which digital content may be played by multiple consumers, where multiple consumers perform different roles within an organization, by applying different filter rules for different roles using a Sensitive Text Classification Model (SSCM), as opposed to editing video separately for each of the multiple roles.

[0023] In accordance with at least one embodiment, the present invention may improve generating consumer-specific digital content for different consumer groups by training an SSCM model with a customized training dataset and a pre-existing dataset, where the customized training dataset is comprised of keywords and entities derived from organizational information and the pre-existing dataset is comprised of one or more pre-existing sensitive word datasets sourced from one or more publicly available resources, and where the customized training dataset and the pre-existing dataset are pre-processed using the Word2Vec algorithm.

[0024] According to at least one embodiment, the present invention may improve processing of digital content using an SSCM model by storing digital content received in its original form in an original content database; identifying one or more sensitive data segments using a sensitive text classification model; and storing the one or more sensitive data segments in a processed content database according to at least one or more of a data type, a timestamp, or a corresponding consumer role, where the corresponding consumer role includes at least one of one or more groups of content consumers.

[0025] In accordance with at least one embodiment, the present invention may improve identifying sensitive data segments by allowing content producers to manually identify sensitive content within a user interface, and by allowing content consumers to manually mark sensitive content that can be stored in a processed content database and used to retrain the SSCM.

[0026] 1, a computing environment 100 includes an example of an environment for the execution of at least a portion of computer code involved in performing a method of the present invention, such as automatically post-processing digital content for sensitive data segments for different groups of content consumers using a content processing module 150. In addition to block 150, the computing environment 100 includes, for example, a computer 101, a wide area network (WAN) 102, an end user device (EUD) 103, a remote server 104, a public cloud 105, and a private cloud 106. In this embodiment, the computer 101 includes a processor set 110 (including processing circuitry 120 and cache 121), a communication fabric 111, a volatile memory 112, a persistent storage 113 (including an operating system 122 and block 150, as identified above), a peripheral device set 114 (including a user interface (UI) device set 123, a storage 124, and an Internet of Things (IoT) sensor set 125), and a network module 115. The remote server 104 includes a remote database 130. The public cloud 105 includes a gateway 140, a cloud orchestration module 141, a set of host physical machines 142, a set of virtual machines 143, and a set of containers 144.

[0027] Computer 101 may take the form of a desktop computer, a laptop computer, a tablet computer, a smart phone, a smartwatch or other wearable computer, a mainframe computer, a quantum computer, or any other form of computer or mobile device now known or to be developed in the future that is capable of executing programs, accessing a network, or querying a database, such as remote database 130. As is well understood in the field of computer technology, and depending on the technology, the execution of a computer-implemented method may be distributed among multiple computers and / or among multiple locations. However, in this presentation of computing environment 100, in order to keep the presentation as concise as possible, the detailed discussion focuses on a single computer, specifically computer 101. Although computer 101 is not shown in the cloud in FIG. 1, it may be located in the cloud. However, computer 101 is not required to be in the cloud except to any extent that may be categorically indicated.

[0028] Processor set 110 includes one or more computer processors of any type now known or to be developed in the future. Processing circuitry 120 may be distributed across multiple packages, e.g., multiple tailored integrated circuit chips. Processing circuitry 120 may implement multiple processor threads and / or multiple processor cores. Cache 121 is memory located within the processor chip package and is typically used for data or code that should be available for quick access by threads or cores executing on processor set 110. Cache memory is typically organized into multiple levels depending on relative proximity to the processing circuitry. Alternatively, some or all of the caches for a processor set may be located "off-chip." In some computing environments, processor set 110 may be designed to operate with qubits and perform quantum computing.

[0029] Computer-readable program instructions are typically loaded into the computer 101 and cause a sequence of operational steps to be executed by the processor set 110 of the computer 101, thereby effecting a computer-implemented method, such that the instructions so executed instantiate the method specified in the computer-implemented method flowcharts and / or narrative descriptions contained herein (collectively, the "methods of the invention"). These computer-readable program instructions are stored in various types of computer-readable storage media, such as cache 121 and other storage media discussed below. The program instructions and associated data are accessed by the processor set 110 to control and direct the execution of the methods of the invention. In the computing environment 100, at least some of the instructions for executing the methods of the invention may be stored in block 150 in persistent storage 113.

[0030] Communications fabric 111 is the signal-conducting pathway that allows the various components of computer 101 to communicate with one another. Typically, this fabric is made up of switches and conductive pathways, such as the switches and conductive pathways that make up buses, bridges, physical input / output ports, etc. Other types of signal communication pathways may be used, such as fiber optic and / or wireless communication pathways.

[0031] Volatile memory 112 is any type of volatile memory now known or to be developed in the future. Examples include dynamic random access memory (RAM) or static RAM. Typically, volatile memory 112 is characterized by random access, although this is not required unless expressly indicated. In computer 101, volatile memory 112 is located in a single package and is internal to computer 101, although alternatively or additionally, volatile memory may be distributed across multiple packages and / or located external to computer 101.

[0032] Persistent storage 113 is any form of non-volatile storage for a computer, now known or later developed. The non-volatility of this storage means that the stored data is maintained regardless of whether power is provided to computer 101 and / or to persistent storage 113 directly. Persistent storage 113 can be read-only memory (ROM), but typically at least a portion of persistent storage allows data to be written, data to be deleted, and data to be rewritten. Some well-known forms of persistent storage include magnetic disks and solid-state storage devices. Operating system 122 can take several forms, such as various known proprietary operating systems or open source Portable Operating System Interface type operating systems that employ a kernel. The code contained in block 150 typically includes at least some of the computer code involved in carrying out the method of the present invention.

[0033] The peripheral device set 114 includes a set of peripheral devices of the computer 101. Data communication connections between the peripheral devices and other components of the computer 101 may be implemented in various ways, such as Bluetooth connections, Near-Field Communication (NFC) connections, connections made by cable (such as a Universal Serial Bus (USB) type cable), insertion type connections (e.g., a secure digital (SD) card), connections made through a local area communication network, and even connections made through a wide area network such as the Internet. In various embodiments, the UI device set 123 may include components such as a display screen, a speaker, a microphone, a wearable device (such as goggles and a smart watch), a keyboard, a mouse, a printer, a touchpad, a game controller, and a haptic device. The storage 124 is an external storage such as an external hard drive, or an insertable storage such as an SD card. The storage 124 may be persistent and / or volatile. In some embodiments, the storage 124 may take the form of a quantum computing storage device for storing data in the form of qubits. In embodiments where computer 101 is required to have a large amount of storage (e.g., where computer 101 stores and manages a large database locally), this storage may be provided by a peripheral storage device designed to store very large amounts of data, such as a storage area network (SAN) shared by multiple, geographically distributed computers. IoT sensor set 125 is made up of sensors that may be used in Internet of Things applications. For example, one sensor may be a thermometer and another sensor may be a motion detector.

[0034] The network module 115 is a collection of computer software, hardware, and firmware that enables the computer 101 to communicate with other computers over the WAN 102. The network module 115 may include hardware such as a modem or Wi-Fi® signal transceiver, software for packetizing and / or depacketizing data for communication network transmission, and / or web browser software for communicating data over the Internet. In some embodiments, the network control and network forwarding functions of the network module 115 are performed on the same physical hardware device. In other embodiments (e.g., embodiments utilizing Software-Defined Networking (SDN)), the control and forwarding functions of the network module 115 are performed on physically separate devices such that the control function manages several different network hardware devices. Computer readable program instructions for carrying out the methods of the present invention may be downloaded to the computer 101 from an external computer or external storage device, typically through a network adapter card or network interface included in the network module 115.

[0035] WAN 102 is any wide area network (e.g., the Internet) capable of communicating computer data over non-local distances by any technology for communicating computer data now known or to be developed in the future. In some embodiments, WAN 102 may be replaced and / or supplemented by a local area network (LAN) designed to communicate data between devices located in a local area, such as a Wi-Fi network. WANs and / or LANs typically include computer hardware such as copper transmission cables, optical transmission fiber, wireless transmission, routers, firewalls, switches, gateway computers, and edge servers.

[0036] End User Device (EUD) 103 is any computer system used and controlled by an end user (e.g., a customer of the business that operates computer 101) and may take any of the forms discussed above in connection with computer 101. EUD 103 typically receives useful and useful data from the operation of computer 101. For example, in a hypothetical case in which computer 101 is designed to provide recommendations to the end user, the recommendations would typically be communicated from network module 115 of computer 101 over WAN 102 to EUD 103. In this manner, EUD 103 can display or otherwise present the recommendations to the end user. In some embodiments, EUD 103 may be a client device such as a thin client, a heavy client, a mainframe computer, a desktop computer, and the like.

[0037] Remote server 104 is any computer system that provides at least some data and / or functionality to computer 101. Remote server 104 may be controlled and used by the same entity that operates computer 101. Remote server 104 represents a machine that collects and stores useful and useful data for use by other computers, such as computer 101. For example, in the hypothetical case where computer 101 is designed and programmed to provide recommendations based on past data, this past data may be provided to computer 101 from remote database 130 of remote server 104.

[0038] Public cloud 105 is any computer system available for use by multiple entities that provides on-demand availability of computer system resources and / or other computer functionality, especially data storage (cloud storage) and computing power, without direct active management by users. Cloud computing typically leverages the sharing of resources to achieve coherence and economies of scale. Direct active management of computing resources of public cloud 105 is performed by computer hardware and / or software of cloud orchestration module 141. Computing resources provided by public cloud 105 are typically implemented by virtual computing environments running on various computers that make up the computers in public cloud 105 and / or host physical machine set 142, which is the universe of physical computers available to public cloud 105. Virtual computing environments (VCEs) typically take the form of virtual machines from virtual machine set 143 and / or containers from container set 144. It is understood that these VCEs can be stored as images and can be transferred among and between various hosts of physical machines, either as images or after instantiation of the VCEs. Cloud orchestration module 141 manages the transfer and storage of images, deploys new instantiations of VCE, and manages active instantiations of VCE deployments. Gateway 140 is a collection of computer software, hardware, and firmware that enables public cloud 105 to communicate over WAN 102.

[0039] Some further discussion of virtualized computing environments (VCEs) is now provided. A VCE may be stored as an "image." A new active instance of a VCE may be instantiated from an image. Two well-known types of VCEs are virtual machines and containers. A container is a VCE that uses operating system level virtualization. This refers to a feature of an operating system where the kernel allows the existence of multiple isolated user space instances called containers. These isolated user space instances typically behave as real computers from the perspective of the programs running within them. A computer program running on a normal operating system may utilize all of the resources of that computer, such as connected devices, files and folders, network shares, CPU power, and quantifiable hardware features. However, a program running inside a container may only use the contents of the container and of the devices assigned to the container, a feature known as containerization.

[0040] A private cloud 106 is similar to a public cloud 105, except that the computing resources are only available for use by a single enterprise. Although the private cloud 106 is shown as being in communication with the WAN 102, in other embodiments, the private cloud may be completely disconnected from the Internet and only accessible through a local / private network. A hybrid cloud is a composite of multiple clouds of different types (e.g., types of private, community, or public clouds), often each implemented by a different vendor. While each of the multiple clouds remains a separate, discrete entity, the larger hybrid cloud architecture is bound together by standardized or proprietary technologies that enable orchestration, management, and / or data / application portability between the constituent clouds. In this embodiment, both the public cloud 105 and the private cloud 106 are part of a larger hybrid cloud.

[0041] According to this embodiment, the computing environment 100 may automatically post-process the digital content for sensitive data segments according to different groups of content consumers using the content processing module 150. The content processing method is described in more detail below with respect to Figures 2 and 3.

[0042] Referring now to FIG. 2, an operational flow diagram illustrating an exemplary content processing process 200 used by the content processing module 150 is shown, according to at least one embodiment.

[0043] At 202, the content processing module 150 builds a sensitive text classification model (SSCM) (e.g., a classification model). The content processing module 150 may build the SSCM (e.g., a classification model) to enable identification of sensitive data segments in digital content, particularly based on the roles of both speakers and / or consumer groups within the organization. The organization may be a business entity, a non-profit organization, an educational institution, or any other organization that may utilize online software in conducting meetings among its members. The members may consist of content producers (e.g., management, organizational executives, professors, faculty), which may be speakers and / or members of the organization who present within the meetings, and content consumers (e.g., employees, volunteers, students). Content consumers may be organizational participants for whom digital content may be produced. Content consumers may also include consumers outside the organization, such as clients and / or the general public. A training process that may be involved in building the SSCM (e.g., a classification model) may be described in more detail in FIG. 3.

[0044] The content processing module 150 may utilize one or more linguistic analysis techniques and / or one or more deep learning models with one or more pre-existing sensitive word datasets and / or customized training datasets in constructing the SSCM (e.g., classification model) for identifying audio and / or visual content within the organization's digital media content that may include sensitive words and / or sensitive visual content. The one or more sensitive words may be identified using one or more pre-existing sensitive word datasets in general and / or with specific association to a consumer role setting using the customized training dataset utilized in constructing the SSCM (e.g., classification model). As discussed in more detail below and in FIG. 3, the one or more pre-existing sensitive word datasets may be sourced from one or more publicly available datasets, and the customized training dataset may be based on at least pre-defined consumer data, consumer roles, and sensitive words, among other data.

[0045] The one or more linguistic analysis techniques may include, but are not limited to, machine learning models using Natural Language Processing (NLP), Latent Dirichlet Allocation (LDA), speech-to-text, Hidden markov model (HMM), N-gram, Speaker Diarization (SD), Semantic Textual Similarity (STS), Keyword Extraction, among other analysis techniques, such as those implemented in IBM Watson® (IBM Watson and all Watson-based trademarks are trademarks or registered trademarks of International Business Machines Corporation in the U.S. and / or other countries), IBM Watson® Speech to Text, IBM Watson® Tone Analyzer, IBM Watson® Natural Language Understanding, and IBM Watson® Natural Language Classifier, among other implementations. As described in more detail below, one or more linguistic analysis techniques may be utilized in generating a text transcript of the audio content that may be filtered using the SSCM (e.g., a classification model).

[0046] The one or more deep learning models that may be utilized by the content processing module 150 may include, but are not limited to, a Convolutional Neural Network (CNN), among other Image Recognition Neural Network deep learning models for images, objects, and / or other visual content. As described in more detail below, the one or more deep learning models may be utilized in identifying visual content that may be filtered using a SSCM (e.g., a classification model). A CNN may also be utilized for processing during the training process of a SSCM (e.g., a classification model), which is described in more detail below and in FIG. 3.

[0047] The content processing module 150 may utilize one or more linguistic analysis techniques and / or one or more deep learning models described above in extracting meaning and / or metadata from the unstructured text data. The unstructured text data may include organizational information received from the organization. The organizational information may include, but is not limited to, internal documents, organizational directories, management chains, job descriptions, employee titles, user profiles, text files, audio files, and / or video files, among other organizational information. The content processing module 150 may not utilize one or more linguistic analysis techniques in performing text analysis on the unstructured text data and / or any other data received from the organization until consent is received from the organization and / or an authorized user within the organization. The content processing module 150 utilizing one or more linguistic analysis techniques may extract at least categories, concepts, classifications, metadata, entities, keywords, sentiment, semantic roles, emotions, relationships, and / or syntax.

[0048] The content processing module 150 may additionally utilize multiple manually identified spoken language examples, including sensitive and non-sensitive content, in constructing the customized training dataset. The content processing module 150 may utilize a template (e.g., base text) and utilize one or more linguistic analysis techniques to identify entities and / or keywords within the template (e.g., base text). In addition, the content processing module 150 may generate variations of the template (e.g., base text) by using synonyms and / or antonyms to replace entities and / or keywords within the base text (e.g., template). The content processing module 150 may perform pre-processing of the variations of the template (e.g., base text) and the sensitive and / or non-sensitive content received from the existing dataset. The content processing module 150 may pre-process this input using an NLP model, including but not limited to a Word2Vec model. The NLP model may utilize a text corpus as input and produce an output that can be classified by a CNN as either sensitive or non-sensitive text. For example, the content processing module 150 may utilize a Word2Vec model, an algorithm that accepts a text corpus as input and outputs a vector representation of each word in the text corpus as output that can be classified by a CNN as either sensitive or non-sensitive.

[0049] The content processing module 150 may also utilize the consumer role settings in constructing the SSCM (e.g., classification model). The consumer role settings may be based on an organizational hierarchy. The organizational hierarchy may be created based on an organizational hierarchy analysis utilizing at least the organizational information detailed above, among other data provided by the organization to the content processing module 150. The content processing module 150 may not perform the organizational hierarchy analysis prior to receiving consent from the organization and / or authorized users of the organization. In performing the organizational hierarchy analysis, the content processing module 150 may utilize one or more tools and / or techniques, such as, but not limited to, the Galton-Watson branching process, the Lightweight Directory Access Protocol (LDAP), among other hierarchy analysis techniques. The organizational hierarchy may be a directory information tree (DIT) that indicates the relative positions of multiple users in the organization with respect to each other. The hierarchy may be utilized in specifying information that may be considered sensitive to multiple groups of consumers based on the role of each group in the organization. For example, compensation information may be considered confidential to consumers within an organization's sales group, but not to consumers working within the organization's payroll group, so the SSCM (e.g., classification model) may block discussion of compensation in digital content for the sales group, but not for the payroll group. Another example may relate to an organization's goal of maintaining trade secrecy, so information such as formulas, practices, processes, designs, equipment, patterns, and / or compilations of information may be considered confidential to different groups within multiple consumers.

[0050] The SSCM (e.g., classification model) may also be utilized in blocking and / or masking visual content identified in the digital content by one or more deep learning models described above. For example, in an internal meeting within an organization, the content processing module 150 may use one or more deep learning models to identify one or more number figures in a screen share deck. In this example, the SSCM (e.g., classification model) may mask and / or block an annual revenue number in the one or more numbers that are determined to be sensitive to at least a portion of the content consumers.

[0051] As described in more detail below, the consumer role configuration may include predefined consumer data, such as roles and sensitive words, that may be utilized by the SSCM (e.g., classification models) in generating the desired output to the processed content database. For example, the predefined consumer data for a XXX Manager may include: { "keywords":["confidential","Contract","XXX company","Deal"], "roles":["XXX manager"] }

[0052] In another embodiment, content may be confidential at one time and not confidential at a later time. As described in more detail below, content creators (e.g., executives, senior management, professors, faculty) may specify different masking periods for different groups of content consumers. For example, a professor may create digital content targeted to a group of teaching assistants and students before an exam. The digital content may be only partially available to students before the exam, but available in its entirety after the exam.

[0053] At 204, the content processing module 150 receives the digital content. The content processing module 150 may receive the digital content, which may be produced by a content creator (e.g., executives, organization executives, professors, faculty) for one or more groups of content consumers (e.g., employees, volunteers, students). The content processing module 150 may receive the digital content at least in an Internet browser, in a dedicated software application, and / or as an integration with a third party software application. The third party software application may include web conferencing and / or video recording software utilized by the organization in conducting meetings, announcements, and / or producing and / or distributing digital content to the content consumers (e.g., employees, volunteers, students).

[0054] The digital content may be produced for one or more groups of content consumers. A content producer (e.g., executives, organization executives, professors, faculty) may identify one or more groups of content consumers (e.g., employees, volunteers, students) for which the digital content is intended in at least one user interface displayed in an internet browser, a dedicated software application, and / or a third-party software application. The content processing module 150 may also identify one or more groups of content consumers for which the content is intended based on the organizational hierarchy generated in step 202. As described in more detail below, the content processing module 150 may utilize a content extractor in extracting at least audio and / or visual content from the received digital content. The SSCM (e.g., classification model) may then be utilized in identifying sensitive data segments within the audio and / or visual content that correspond to one or more groups of content consumers. As described in more detail below with respect to step 206, the content producer (e.g., executives, organization executives, professors, faculty) may also manually trigger the masking and / or blocking of content in the user interface. A content creator may manually select at least one of the one or more groups within a user interface during the creation of digital content for a particular segment that may be automatically stored in the processed content database by the content processing module 150.

[0055] At 206, the content processing module 150 processes the digital content. The content processing module 150 may process the digital content for each group of content consumers using the SSCM (e.g., classification model) and two or more content databases. The two or more content databases may include at least an original content database and a processed content database. The two or more content databases may be maintained by the content processing module in the database 130. The original content database may store the digital content in its original format as recorded by the content creator. The processed content database may store the sensitive data segments identified by the SSCM (e.g., classification model) and their corresponding consumer roles.

[0056] The digital content may be comprised of at least audio and / or video content. The content processing module 150 may utilize one or more linguistic analysis techniques, such as speech-to-text, in generating a text transcript of the audio content that may be filtered using the SSCM (e.g., classification models) in identifying sensitive content. The content processing module 150 may utilize one or more linguistic analysis techniques, such as speech-to-text, and / or one or more deep learning models to employ visual analysis techniques, such as image and / or object recognition, in identifying visual content that may be filtered using the SSCM (e.g., classification models). For example, sensitive visual content may be detected while a content creator shares their screen with a client during a product launch. The content creator may have forgotten to close out one or more internal confidential documents. In this example, when the content processing module 150 detects that the document contains the keyword "confidential," the content processing module 150 may mask and / or block the one or more internal confidential documents to ensure that no sensitive data is leaked. The content processing module 150 may store content identified within the audio and / or visual content as possibly containing sensitive information based on the SSCM (eg, classification models) in a processed content database.

[0057] The processed content database may store the output of the SSCM (e.g., classification model). The content processing module 150 may store the sensitive data segments according to data type, timestamp / location within the digital content, and / or corresponding consumer role. For example, the content processing module 150 may store sensitive audio segments in the processed content database using the following: { "type":"audio", "startTime":"00:01:23", "endTime":"00:01:59", "keyWords":["profit","cost"], "appliedRoles":["general public","non-employee"] } And, the sensitive video segments may be stored in the Processed Content Database using: { "type":"video", "startTime":"00:10:23", "endTime":"00:10:44", "keyWords":["confidential","Contract","XXX company","Deal"], "appliedRoles":["XXX manager"] }

[0058] The content processing module 150 may also store one or more sensitive data segments manually identified by the producer in the processed content database. In addition, the content processing module 150 may identify additional content segments that may be similar to the one or more data segments identified by the producer using one or more linguistic analysis techniques described in step 202. The content processing module 150 may present the producer in a user interface with a list of the manually identified and / or the content processing module 150 identified sensitive data segments for each group within the content consumers (e.g., employees, volunteers, students). The content producer (e.g., management, organizational executives, professors, faculty) may edit and / or approve each of the one or more sensitive data segments for each group.

[0059] As discussed in more detail below with respect to step 208, the content processing module may utilize the digital content stored in the original content database and the sensitive data segments stored in the processed content database in generating consumer-specific digital content.

[0060] In another embodiment, the content processing module 150 may utilize a delay when processing digital content so that it may be received in real time and processed using the SSCM (e.g., classification model) during the intervening delay, such that the content processing module 150 may be utilized for digital content intended for playback, as well as digital content intended to be consumed in real time, such as an organization's meetings.

[0061] At 208, the content processing module 150 generates consumer-specific digital content. The content processing module 150 may generate consumer-specific digital content for each consumer group according to the consumer role settings.

[0062] The content processing module 150 may generate consumer-specific digital content for each of one or more consumer groups by retrieving from the original content database the digital content in its original form as recorded by the content creator and one or more sensitive data segments including timestamps, corresponding consumer roles, and / or associated data stored in the processed content database. Thus, the content processing module 150 may determine which video objects and / or audio time slots may require masking and / or blocking for each of one or more consumer groups.

[0063] The content processing module 150 may utilize an output content processor in masking and / or blocking audio and / or visual sensitive data segments. The output content processor may mask and / or block audio and / or visual sensitive data segments by muting the sound for a time period associated with the audio sensitive data segment, deleting the entire segment from the digital content, and hiding, scrambling, or replacing keywords identified in the digital content transcript, among other methods of masking and / or blocking the digital content.

[0064] The content processing module 150 may further utilize an output content processor in transmitting and / or displaying the consumer-specific content to each of the multiple consumers according to the consumer groupings in the user interface presented by the internet browser, the dedicated software application, and / or the third-party software application, as described in detail above with respect to step 204. The content processing module 150 may also generate a unique download link for each content consumer for downloading via http, https, and / or ftp. The content processing module 150 may receive approval for each version of the generated consumer-specific content from the content creator (e.g., executives, organizational executives, professors, faculty) and / or another authorized party within the organization prior to transmitting and / or displaying the consumer-specific content. The content processing module 150 may also require verification of the content consumer group prior to allowing viewing by the content consumer. The content processing module 150 may receive the verification of the consumer group based on a consumer profile associated with the device, a consumer profile associated with the internet browser, the software application, and / or the third-party application, and / or a manual selection made by the consumer in the user interface.

[0065] The content processing module 150 may also enable content consumers to manually mark sensitive content within consumer-specific content. The content processing module 150 may request additional feedback from consumers regarding the identified sensitive content and store the sensitive data segments and associated data in a customized training data set. The content processing module 150 may utilize the identified additional sensitive data segments in retraining the SSCM (e.g., classification model), the training process being described in more detail below with respect to FIG.

[0066] Referring now to FIG. 3, a block diagram and flow chart illustrating a process for training a SSCM (e.g., a classification model) utilizing some of the components discussed in FIG. 2 is provided, according to at least one embodiment.

[0067] As described in detail above with respect to at least step 202, building a sensitive sentence classification model (SSCM) 310 (e.g., a classification model) to enable identification of sensitive data segments within digital content based specifically on the roles of both speakers and / or consumer groups within an organization includes a training process that begins with a customized training dataset 302 and an existing dataset 304.

[0068] The customized training data set 302 may be constructed based on at least predefined consumer data, consumer roles, sensitive words, among other data that may be extracted utilizing one or more linguistic analyses detailed in step 202. One or more linguistic analysis techniques may extract meaning and / or metadata from unstructured text data received from organizations, such as organizational information described in step 202. One or more linguistic analysis techniques may extract meaning and / or metadata using text analysis to identify entities from the organizational information, such as, but not limited to, consumers, creators, companies, organizations, cities, geographic features, and / or other entities. Additionally, linguistic analysis techniques such as IBM Watson® Natural Language Processing® (IBM Watson and all Watson-based trademarks are trademarks or registered trademarks of International Business Machines Corporation in the United States and / or other countries) may be utilized in extracting keywords associated with each of the identified entities. The identified entities and keywords may be manually identified as either sensitive or non-sensitive and used to generate variations by modifying the entities and / or keywords with synonyms or antonyms. The customized training dataset 302 is comprised of the identified entities and keywords and generated variations. The existing dataset 304 may be built using one or more existing sensitive word datasets sourced from one or more publicly available datasets.

[0069] The customized training dataset 302 and the existing dataset 304 may then be provided as input to an NLP model, such as the Word2Vec model 306 shown for preprocessing. In preprocessing the customized training dataset 302 and the existing dataset 304 with the Word2Vec algorithm, a neural network model may be utilized to learn sensitive or non-sensitive word associations from the training dataset, such that the word embeddings output by the algorithm may enable grouping of vectors for similar words. In the case of audio content, the training process may require an additional speech-to-text step, which may be performed using IBM Watson® Speech to Text (IBM Watson and all Watson-based trademarks are trademarks or registered trademarks of International Business Machines Corporation in the United States and / or other countries).

[0070] A convolutional neural network (CNN) 308 may then be used to classify the sentences. The CNN 308 may utilize an nxk representation of the sentences, where n is the number of words and k is the number of vectors for each word. The CNN may additionally utilize static and non-static channels, convolutional layers with multiple filter widths and feature maps, maximum temporal pooling, and fully connected layers with dropout and SoftMax output in determining whether a sentence is classified or non-classified.

[0071] Thus, digital content intended for each group of content consumers may be automatically processed using the SSCM 310 (eg, classification models) such that sensitive data segments may be masked and / or blocked according to different content consumer groups.

[0072] It can be understood that Figures 2 and 3 provide only an illustration of one embodiment and do not imply any limitations on how different embodiments may be implemented. Many modifications to the illustrated embodiment can be made based on design and implementation requirements.

[0073] Various aspects of the present disclosure are described by narrative text, flow charts, block diagrams of computer systems, and / or block diagrams of machine logic included in computer program product (CPP) embodiments. For any flow chart, depending on the technology involved, operations may be performed in a different order than shown in a given flow chart. For example, again depending on the technology involved, two operations shown in successive flow chart blocks may be performed in reverse order, as a single integrated step, simultaneously, or in an at least partially overlapping manner in time.

[0074] A computer program product embodiment ("CPP embodiment" or "CPP") is a term used in this disclosure to describe any set of one or more storage media (also referred to as "media") collectively included in a set of one or more storage devices that collectively contain machine-readable code corresponding to instructions and / or data for performing a computer operation specified in a given CPP claim. A "storage device" is any tangible device that can hold and store instructions for use by a computer processor. The computer-readable storage medium may be, but is not limited to, an electronic storage medium, a magnetic storage medium, an optical storage medium, an electromagnetic storage medium, a semiconductor storage medium, a mechanical storage medium, or any suitable combination of the foregoing. Some known types of storage devices that include these media include diskettes, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), compact disk read-only memory (CD-ROM), digital versatile disk (DVD), memory stick, floppy disk, mechanically encoded devices (such as punch cards or pits / lands formed on a major surface of a disk), or any suitable combination of the above. Computer-readable storage media, as the term is used in this disclosure, is not to be construed as storage in the form of a transitory signal itself, such as electric waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through a waveguide, light pulses passing through fiber optic cables, electrical signals transmitted through wires, and / or other transmission media. As will be appreciated by those skilled in the art, data is typically moved at some occasional time during the normal operation of a storage device, such as during access, defragmentation, or garbage collection, but the above does not render a storage device temporary, since the data is not temporary while it is stored.

[0075] The description of various embodiments of the present invention has been presented for illustrative purposes, but is not intended to be exhaustive or limited to the disclosed embodiments. Many modifications and variations will become apparent to those skilled in the art without departing from the scope of the described embodiments. The terminology used in this specification has been selected to best explain the principles of the embodiments, practical applications, or technical improvements to the technology found in the market, or to enable other skilled in the art to understand the embodiments disclosed herein.

[0076] This disclosure should not be construed as violating or encouraging the violation of any local, state, federal, or international law regarding privacy protections.

Claims

1. Building a confidential text classification model; receiving digital content, said digital content targeted to one or more groups of content consumers; processing the digital content using the sensitive text classification model; and generating consumer-specific digital content for each of said one or more groups of content consumers; A method for processing digital content comprising:

2. The step of constructing the confidential text classification model includes: Training the sensitive text classification model using a customized training dataset and an existing dataset. The method of claim 1 further comprising:

3. 3. The method of claim 2, wherein the customized training data set is comprised of keywords and entities derived from organizational information, and the pre-existing data set is comprised of one or more pre-existing sensitive word data sets sourced from one or more publicly available resources.

4. The method of claim 2 , wherein the customized training dataset and the existing dataset are pre-processed using a Word2Vec algorithm.

5. The method of claim 1 , wherein the one or more groups of content consumers are identified by a content producer in a user interface.

6. Processing the digital content using the sensitive text classification model comprises: storing the digital content received in its original format in an original content database; identifying one or more sensitive data segments using the sensitive text classification model; and storing the one or more sensitive data segments in a processed content database according to at least one or more of a data type, a timestamp, or a corresponding consumer role, where the corresponding consumer role includes at least one of the one or more groups of the content consumers. The method of claim 1 , further comprising:

7. The step of identifying one or more sensitive data segments comprises: receiving one or more additional sensitive data segments manually identified by a content creator; and storing said one or more additional sensitive data segments in said processed content database. The method of claim 6 further comprising:

8. assigning a plurality of consumers within the organization to one or more groups of the content consumers based on an organizational hierarchical structure constructed based on the analysis of organizational information; and displaying said consumer-specific digital content to a plurality of consumers, where each of said plurality of consumers may access said consumer-specific digital content corresponding to their content consumer group. The method of claim 1 , further comprising:

9. 1. A computer system for processing digital content, comprising: a computer system comprising one or more processors, one or more computer readable memories, one or more computer readable tangible storage media, and program instructions stored on at least one of the one or more tangible storage media for execution by at least one of the one or more processors via at least one of the one or more memories; Steps to build a confidential text classification model; receiving digital content, said digital content being targeted to one or more groups of content consumers; processing the digital content using the sensitive text classification model; and generating consumer-specific digital content for each of said one or more groups of content consumers; A computer system capable of executing a method comprising:

10. The steps of constructing the confidential text classification model are as follows: program instructions stored on at least one of the one or more computer-readable storage media for execution by at least one of the one or more processors via at least one of the one or more memories for training the sensitive text classification model with a customized training dataset and an existing dataset; 10. The computer system of claim 9, further comprising:

11. 11. The computer system of claim 10, wherein the customized training data set is comprised of keywords and entities derived from organizational information, and the pre-existing data set is comprised of one or more pre-existing sensitive word data sets sourced from one or more publicly available resources.

12. The computer system of claim 10 , wherein the customized training dataset and the existing dataset are pre-processed using a Word2Vec algorithm.

13. 13. A computer system according to claim 9, wherein the one or more groups of content consumers are identified by a content producer in a user interface.

14. The steps of processing the digital content using the sensitive text classification model include: program instructions stored on at least one of the one or more computer-readable storage media for execution by at least one of the one or more processors via at least one of the one or more memories for storing the digital content received in its original form in an original content database; program instructions stored on at least one of the one or more computer-readable storage media for execution by at least one of the one or more processors via at least one of the one or more memories for identifying one or more sensitive data segments using the sensitive text classification model; and and program instructions stored on at least one of the one or more computer readable storage media for execution by at least one of the one or more processors via at least one of the one or more memories for storing the one or more sensitive data segments in a processed content database according to at least one or more of a data type, a timestamp, or a corresponding consumer role, where the corresponding consumer role includes at least one of one or more groups of the content consumers.

13. The computer system of claim 9, further comprising:

15. The step of identifying the one or more sensitive data segments comprises: program instructions stored on at least one of the one or more computer-readable storage media for execution by at least one of the one or more processors via at least one of the one or more memories for receiving one or more additional sensitive data segments manually identified by a content creator; and program instructions stored on at least one of the one or more computer-readable storage media for execution by at least one of the one or more processors via at least one of the one or more memories for storing the one or more additional sensitive data segments in the processed content database. The computer system of claim 14 further comprising:

16. 1. A computer program for processing digital content, comprising: Steps to build a confidential text classification model; receiving digital content, said digital content being targeted to one or more groups of content consumers; processing the digital content using the sensitive text classification model; and generating consumer-specific digital content for each of said one or more groups of content consumers; A computer program for executing the above.

17. The steps of constructing the confidential text classification model are as follows: Training the confidential text classification model using a customized training dataset and an existing dataset.

17. The computer program of claim 16, further comprising:

18. 20. The computer program product of claim 17, wherein the customized training data set is comprised of keywords and entities derived from organizational information, and the pre-existing data set is comprised of one or more pre-existing sensitive word data sets sourced from one or more publicly available resources.

19. 20. The computer program product of claim 17, wherein the customized training dataset and the existing dataset are pre-processed using a Word2Vec algorithm.

20. The steps of processing the digital content using the sensitive text classification model include: storing the digital content received in its original format in an original content database; identifying one or more sensitive data segments using the sensitive text classification model; and storing the one or more sensitive data segments in a processed content database according to at least one or more of a data type, a timestamp, or a corresponding consumer role, where the corresponding consumer role includes at least one of one or more groups of the content consumers.

20. A computer program according to any one of claims 16 to 19, further comprising:

Images