Abstract: Methods, apparatus, and systems for using sensors to detect potential security risks include tracking, by a sensor, a component of a system, comparing sensor data from the sensor with a baseline measurement associated with the component, and flagging, based on detecting that the sensor data differs from the baseline measurement by a threshold amount, the sensor data as indicating a potential security risk.

Abstract: Continual neural network training in an edge computing environment includes deploying a plurality of copies of a centralized neural network respectively to a corresponding plurality of edge servers, wherein each of the copies of the centralized neural network is independently operated and trained at a respective one of the edge servers based on inputs received at that edge server to create independently trained neural networks. At periodic intervals, copies of the independently trained neural networks and a corresponding fitness measure for each of the copies of the independently trained neural networks are sent from the plurality of edge servers to a cloud-based data center. The centralized neural network is updated at the cloud-based data center, including performing neural network breeding based on the copies of the independently trained neural networks sent from the plurality of edge servers.

Abstract: Deploying a neural network to a new edge server in an edge computing environment includes deploying a plurality of copies of a centralized neural network respectively to a corresponding plurality of edge servers, wherein each of the copies of the centralized neural network is independently operated and trained at a respective one of the edge servers based on inputs received at that edge server to create independently trained neural networks. Each of the edge servers stores edge devices information including a physical location of each edge device that accessed the independently trained neural network operating at the edge server. A new edge neural network is generated for deployment to the new edge server, including performing neural network breeding based on the independently trained neural networks and the stored edge devices information, and based on anticipated edge devices information for edge devices expected to access the new edge server.

Abstract: Selective breeding for divergent neural networks in an edge computing environment includes deploying a plurality of copies of a centralized neural network respectively to a corresponding plurality of edge servers, wherein each of the copies of the centralized neural network is independently operated and trained at one of the edge servers based on inputs received at that edge server and becomes an independently trained neural network. Each of the edge servers at periodic intervals sends a copy of the independently trained neural network at that edge server to other ones of the edge servers. At each of one or more of the edge servers, the independently trained neural network at that edge server is updated, including performing neural network breeding based on the independently trained neural network at that edge server and one or more copies of the independently trained neural networks sent to the edge server.

Abstract: Methods, apparatus, and systems for altering refractive index of an expanded beam connector lens include receiving signals from an expanded beam connector, detecting a decibel loss of the signals in the expanded beam connector that exceeds a loss threshold, and directing, via a transmitter included in the expanded beam connector and in response to detecting the decibel loss, light at an expanding lens included in the expanded beam connector.

Abstract: Methods, apparatus, and systems for indicating cable temperature with chromogenic polymer coatings include determining, based on a temperature threshold, a material for a chromogenic polymer coating, wherein the chromogenic polymer coating is configured to change color based on a temperature of the chromogenic polymer coating exceeding the temperature threshold, and applying the chromogenic polymer coating to a surface of a cable.

Abstract: Methods, apparatus, and systems for inspecting and cleaning optical interconnects include inspecting, by a light scattering detection system and an imaging processing system of an inspection and cleaning tool, a surface of an optical cable interconnect; determining, based on measurements from the light scattering detection system and the imaging processing system, whether the surface of the optical cable interconnect is clean; and cleaning, based on determining that the surface of the optical cable interconnect is not clean, the surface of the optical cable interconnect.

Abstract: A method of testing fiber optic communication lanes includes sending a plurality of test parameters by a first end unit of a testing tool to a second end unit of the testing tool through at least one communication lane of a plurality of communication lanes that extend from a first cable endpoint to a second cable endpoint through at least one fiber optic cable. The first cable endpoint is connected to the first end unit and the second cable endpoint is connected to the second end unit, and the plurality of test parameters comprises wavelengths of a plurality of test signals. The method further includes sending the plurality of test signals by the first end unit to the second end unit through the plurality of communication lanes, and receiving test results through the at least one communication lane of the plurality of communication lanes.

Abstract: Methods, apparatus, and systems for cooling optical interconnects using light include monitoring an output power of a transceiver, where the transceiver includes an alkali activated geopolymer (AAGP) coating, and cooling, based on a determination that the output power has decreased below a threshold value, the transceiver by directing infrared light towards the transceiver.

Abstract: A computer-implemented method, system, and computer program product for maximizing bandwidth utilization of PCIe links. The bandwidth utilization of a PCIe link involving a PCIe card is measured. A bandwidth utilization of the PCIe link at a future time is predicted based on the measured bandwidth utilization of the PCIe link using a machine learning model trained to predict bandwidth utilization of PCIe links. If the predicted bandwidth utilization of the PCIe link exceeds a threshold value, then the PCIe card is configured to implement a first mode of operation that utilizes more bandwidth if not implementing the first mode of operation at the future time. If the predicted bandwidth utilization of the PCIe link does not exceed a threshold value, then the PCIe card is configured to implement a second mode of operation that utilizes less bandwidth if not implementing the second mode of operation at the future time.

Abstract: An optical cable mismatch error detection and remediation process is provided which includes detecting, by at least one processor, an optical cable mismatch error in a connection path between optical transceivers of a computing network, where the connection path includes one or more optical cables. The detecting includes evaluating operation of an optical transceiver of the connection path. The evaluating includes determining that data on one or more optical lanes of the optical transceiver is lane-mismatched data. The process further includes restructuring one or more lane assignments of the optical transceiver, based on determining that the data on the one or more optical lanes of the optical transceiver is lane-mismatched data, where the restructuring addresses the detected optical cable mismatch error.

Abstract: A computer system, computer readable storage medium, and computer-implemented method for generating random seeds for encryption. The method includes emitting photons into a fluid. The fluid is maintained within a fluid containment structure. The method also includes capturing at least a portion of the photons. The captured photons are at least partially representative of a photonic reflection from the fluid containment structure and the fluid therein. The photonic reflection is random and unpredictable. The method further includes converting the photonic reflection into one or more random encryption key seeds.

Abstract: An optical assembly is provided and includes a first component configured to output light, a second component and an adhesive bonding agent to which a photochromic reactive agent is added and by which the first and second components are attached such that the first component is disposed to output the light toward the second component and such that the photochromic reactive agent is exposed to the light in an event of the first and second components being attached in a misaligned condition whereupon a color change of the photochromic reactive agent resulting from exposure of the photochromic reactive agent to the light is indicative of the misaligned condition.

Abstract: Systems and methods are provided for assigning an issue for resolution using natural language processing (NLP) and updating recognition scores for individual/teams accurately redirecting an issue to a different individual/team having a greater ability to resolve it. An issue is analyzed using NLP, and the text is compared to each individual/team's corpus of issues to derive a match percentage. A list is built which ranks individuals/teams by the match percentage. Weights are applied to each individual/team in the list, based on their corresponding recognition scores in their profiles in a profile database. The recognition scores indicate an ability to recognize correct reassignment with a degree of accuracy above a threshold. The list is reordered based on the applied weights, and the issue is assigned to the individual/team having a highest rank.

Abstract: An optical cable mismatch error detection and remediation process is provided which includes detecting, by at least one processor, an optical cable mismatch error in a connection path between optical transceivers of a computing network, where the connection path includes one or more optical cables. The detecting includes evaluating operation of an optical transceiver of the connection path. The evaluating includes determining that data on one or more optical lanes of the optical transceiver is lane-mismatched data. The process further includes restructuring one or more lane assignments of the optical transceiver, based on determining that the data on the one or more optical lanes of the optical transceiver is lane-mismatched data, where the restructuring addresses the detected optical cable mismatch error.

Abstract: An optical cable misplug error detection and remediation process is provided which includes detecting, by at least one processor, an optical cable misplug error in a connection path between optical transceivers of a computing network, where the connection path includes one or more optical cables. The detecting includes evaluating operation of an optical transceiver of the connection path. The evaluating includes determining whether a receive lane of the optical transceiver is receiving a signal during evaluating operation of the optical transceiver, and detecting the optical cable misplug error where the receive lane of the optical transceiver is not receiving the signal. The process further includes identifying, by the at least one processor based on detecting the optical cable misplug error, a location of a misplugged optical cable in the connection path between the optical transceiver of the computing network.

Abstract: According to one embodiment, a method, computer system, and computer program product for gathering relevant digital content. The embodiment may include receiving, from an online meeting scheduler, information within an online meeting invite. The embodiment may include analyzing the information using natural language processing (NLP) and machine learning (ML) techniques. Based on results of the analysis of the received information, the embodiment may include locating additional digital content items of the user which are relevant to the information of the online meeting invite. The embodiment may include outputting, to a user, a top-k listing of the additional digital content items for attachment to the online meeting invite before being sent.

Abstract: According to one embodiment, a method, computer system, and computer program product for gathering relevant digital content. The embodiment may include receiving, from an online meeting scheduler, information within an online meeting invite. The embodiment may include analyzing the information using natural language processing (NLP) and machine learning (ML) techniques. Based on results of the analysis of the received information, the embodiment may include locating additional digital content items of the user which are relevant to the information of the online meeting invite. The embodiment may include outputting, to a user, a top-k listing of the additional digital content items for attachment to the online meeting invite before being sent.

Abstract: A system to reduce surface bubbles from cell culture media using ultrasonic pulses is disclosed. The system includes a plurality of ultrasonic transducers, a photodetector, a camera, a control circuit, and a cell culture container that includes a cell culture media. The control circuit is communicatively coupled to the plurality of ultrasonic transducers, the temperature sensor, and the camera. The photodetector is used to determine a presence of a surface bubble on the cell culture media. The camera is used to determine a location of the surface bubble when the presence of the surface bubble is determined. The plurality of ultrasonic transducers is used to position an ultrasonic pulse at the location of the surface bubble and thereby reduce the presence of the surface bubble on the cell culture media when the location of the surface bubble is determined.