Abstract: In some examples, recruitment process graph based unsupervised anomaly detection may include obtaining log data associated with a recruitment process for a plurality of candidates, and generating knowledge graphs and graph embeddings. The graph embeddings may be trained to include a plurality of properties such that graph embeddings of genuine candidate hires and fraudulent candidate hires are appropriately spaced in a vector space. The trained graph embeddings may be clustered to generate a plurality of embedding clusters that include a genuine candidate cluster, and a fraudulent candidate cluster. For a new candidate graph embedding for a new candidate, a determination may be made as to whether the new candidate graph embedding belongs to the genuine candidate cluster, to the fraudulent candidate cluster, or to an anomalous cluster, and instructions may be generated to respectively retain or suspend the new candidate.

Abstract: In some examples, natural language unification based robotic agent control may include ascertaining, by a robotic agent, an image of an object or an environment, and ascertaining a plurality of natural language insights for the image. A semantic relatedness may be determined between each insight of the plurality of insights, and a semantic relatedness graph may be generated for the plurality of insights. For each insight of the plurality of insights, at least one central concept may be identified. Based on the semantic relatedness graph and the identified at least one central concept, the plurality of insights may be clustered to generate at least one insights cluster. For insights included in the least one insights cluster, a unified insight may be generated. Further, an operation associated with the robotic agent, the object, or the environment may be controlled by the robotic agent and based on the unified insight.

Abstract: In some examples, natural language dialogue system perturbation testing may include identifying semantic segments for conversation data for a natural dialogue system. For each semantic segment, a perturbed variant that includes a perturbation may be generated, and forwarded to the natural dialogue system. An updated response to the perturbed variant may be obtained from the natural dialogue system. A semantic similarity may be determined between an original response to a semantic segment and the updated response, and based on the semantic similarity between the original response and the updated response, a perturbability of the natural dialogue system may be determined. A determination may be made as to whether the perturbability of the natural dialogue system is greater than a specified perturbability threshold, and if so, a training corpus that includes a failed response to a perturbed variant may be utilized to train the natural dialogue system.

Abstract: A device may obtain information identifying a base application. The device may extract a set of sub-application artifacts associated with the base application based on structural information associated with the base application. The device may define a set of metadata attributes associated with the set of sub-application artifacts associated with the base application. The device may generate a set of hash tuples for the set of metadata attributes associated with the set of sub-application artifacts associated with the base application. The device may generate a base composite identity of the base application based on the set of hash tuples. The device may store the base composite identity in a blockchain and in connection with storage of the base application in the blockchain to enable subsequent identification and verification of the base application.

Abstract: In some examples, extended reality based positive affect implementation for product development may include determining, for a user wearing or utilizing a device, characteristics and an environment of the user. Based on the characteristics and the environment of the user, a plurality of augmentation elements may be determined. A recommendation of a plurality of augmentation elements from the determined plurality of augmentation elements may be generated. Based on selection of an augmentation element from the recommended plurality of augmentation elements, the augmentation element may be rendered in an extended environment and/or an immediate environment of the user. Interaction of the user may be controlled with the rendered augmentation element. Feedback of the user may be analyzed with respect to the rendered augmentation element.

Abstract: In some implementations, a device may receive a machine learning model to be tested. The device may process the machine learning model, with generalization testing methods, to determine generalization which identifies responsiveness of the machine learning model to varying inputs. The device may process the machine learning model, with robustness testing methods, to determine robustness which identifies responsiveness of the machine learning model to improper inputs. The device may process the machine learning model, with an interpretability testing method, to determine decisions of the machine learning model. The device may calculate a score for the machine learning model based on the generalization data, the robustness data, and the interpretability data. The device may perform one or more actions based on the score for the machine learning model.

Abstract: Systems, apparatuses, and methods are directed towards identifying that an adversarial patch image includes a plurality of pixels. The systems, apparatuses, and methods include dividing the adversarial patch image into a plurality of blocks, that each include a different group of the pixels in which the pixels are contiguous to each other, and assigning a first plurality of colors to the plurality of blocks to assign only one of the first plurality of colors to each pixel of one of the plurality of blocks.

Abstract: In some examples, machine learning based semantic structural hole identification may include mapping each text element of a plurality of text elements of a corpus into an embedding space that includes embeddings that are represented as vectors. A semantic network may be generated based on semantic relatedness between each pair of vectors. A boundary enclosure of the embedding space may be determined, and points to fill the boundary enclosure may be generated. Based on an analysis of voidness for each point within the boundary enclosure, a set of void points and void regions may be identified. Semantic holes may be identified for each void region, and utilized to determine semantic porosity of the corpus. A performance impact may be determined between utilization of the corpus to generate an application by using the text elements without filling the semantic holes and the text elements with the semantic holes filled.

Abstract: A device may receive a first sample. The device may identify a plurality of second samples that are similar to the first sample. The plurality of second samples may correspond to service tickets that have previously been processed. The device may obtain or determine log information identifying a plurality of events associated with the first sample. The device may determine respective probabilities of occurrence corresponding to the plurality of second samples. A probability of occurrence, for a particular second sample, may identify a likelihood of receiving a corresponding service ticket within a particular time period of occurrence of the plurality of events. The device may assign ranks to the plurality of second samples based on the respective probabilities of occurrence. The device may provide information identifying the ranks and the plurality of second samples.

Abstract: Systems, apparatuses, and methods are directed towards identifying that an adversarial patch image includes a plurality of pixels. The systems, apparatuses, and methods include dividing the adversarial patch image into a plurality of blocks, that each include a different group of the pixels in which the pixels are contiguous to each other, and assigning a first plurality of colors to the plurality of blocks to assign only one of the first plurality of colors to each pixel of one of the plurality of blocks.

Abstract: A device may receive information regarding a software environment and/or one or more modules associated with the software environment; receive information regarding one or more compliance requirements; determine one or more smart contracts based on the information regarding the one or more compliance requirements and the information regarding the software environment and/or the one or more modules associated with the software environment; obtain data regarding the software environment and/or the one or more modules; determine, using the one or more smart contracts, whether the data satisfies the one or more compliance requirements; and add at least part of a block, to a blockchain, to indicate whether the data satisfies the one or more compliance requirements, for realization of a reward or penalty associated with the one or more compliance requirements.

Abstract: In some examples, machine learning based quantification of performance impact of data irregularities may include generating an irregularity feature vector for each text analytics application of a plurality of text analytics applications. Normalized data associated with a corresponding text analytics application may be generated for each text analytics application and based on minimization of irregularities present in un-normalized data associated with the corresponding text analytics application. An un-normalized data machine learning model may be generated for each text analytics application and based on the un-normalized data associated with the corresponding text analytics application. A normalized data machine learning model may be generated for each text analytics application and based on the normalized data associated with the corresponding text analytics application.

Abstract: According to an example, partitioning based migration of systems to container and micro-service based-platforms may include determining, based on an analysis of source code for an application that is to be partitioned, an entity model corresponding to the application, identifying resources associated with the application, and determining a mapping of the identified resources to entities of the entity model. Further, partitioning based migration of systems to container and micro-service based-platforms may include identifying dependencies for each of the mapped resources, generating dependency and control flow metrics for the application, generating affinity values between the mapped resources, generating a resource affinity graph, determining an affinity score between each of the mapped resources, and generating resource clusters that correspond to partitions of the application.

Abstract: System and methods for cryptologic digital twin segregation with blockchain for development operations are provided. By way of introductory example, a system may receive a first development object from a first devops tool and a second development object from a second devops tool. The system may identify, in a segregation model graph, a first node and a second node, the first node corresponding to the first development object and the second node corresponding to the second development object. The system may determine the first node is mapped to an off-chain storage classification and the second node is mapped to an on-chain storage classification. The system may store the first development object in an off-chain storage, generate a hash value of the first development object, and store the hash value of the first development object on a blockchain. The system may store the second development object on the blockchain.

Abstract: A device may identify a first concept hierarchy of a first multimedia presentation and a second concept hierarchy of a second multimedia presentation. The device may determine a set of concepts associated with the first concept hierarchy and the second concept hierarchy and may determine a plurality of similarity scores associated with the set of concepts. The device may generate a new concept hierarchy based on the plurality of similarity scores, the first concept hierarchy, and the second concept hierarchy by: performing a first process to merge a concept with an additional concept; performing a second process to position a concept sequentially before or after an additional concept; performing a third process to position a concept and a different concept to sequentially follow an additional concept; and performing a fourth process to position a concept to sequentially follow an additional concept and a different concept.

Abstract: In some examples, extended reality based positive affect implementation for product development may include determining, for a user wearing or utilizing a device, characteristics and an environment of the user. Based on the characteristics and the environment of the user, a plurality of augmentation elements may be determined. A recommendation of a plurality of augmentation elements from the determined plurality of augmentation elements may be generated. Based on selection of an augmentation element from the recommended plurality of augmentation elements, the augmentation element may be rendered in an extended environment and/or an immediate environment of the user. Interaction of the user may be controlled with the rendered augmentation element. Feedback of the user may be analyzed with respect to the rendered augmentation element.

Abstract: A device may perform a static analysis of a codebase associated with a software application to determine a structure of the codebase and one or more metrics associated with the codebase. The device may generate a structured format of the codebase based on the structure of the codebase. The device may generate a domain specific language format of the codebase based on the structured format of the codebase and the one or more metrics associated with the codebase. The device may generate an extended reality rendered view of the domain specific language format of the codebase. The device may provide, to an extended reality device, the extended reality rendered view of the domain specific language format of the codebase.

Abstract: In some examples, extended reality based immersive project workspace creation may include obtaining sensing data from at least one sensor of a device worn by a user, and analyzing, based on the sensing data, an environment of the user. A plurality of insights associated with the user may be determined, and prioritized based on a plurality of prioritization criteria. An insight of the plurality of prioritized insights associated with the user may be rendered in the environment of the user. Further, the rendering of the insight may be controlled based on an interaction of the user with the rendered insight.

Abstract: In some examples, natural language eminence based robotic agent control may include ascertaining, by a robotic agent, an image of an object or an environment, and ascertaining a plurality of natural language insights for the image. For each insight of the plurality of insights, an eminence score may be generated, and each insight of the plurality of insights may be ranked according to the eminence scores. An operation associated with the robotic agent, the object, or the environment may be controlled by the robotic agent and based on a highest ranked insight.

Abstract: A device receives a machine learning model, model data associated with the machine learning model, and identifier generation data. The identifier generation data includes data utilized to generate identifier pairs that may be used to authenticate the machine learning model. The device selects an identifier model, for generating the identifier pairs, based on the machine learning model, the model data, and the identifier generation data. The device processes the machine learning model, the model data, and the identifier generation data, with the selected identifier model, to generate the identifier pairs and identifier pair data. The device stores the identifier pairs and the identifier pair data in one or more data structures, and utilizes the identifier pairs to identify and provide authentication for the machine learning model.