Abstract: Robot control systems, methods, control modules and computer program products that leverage one or more large language model(s) (LLMs) in order to achieve at least some degree of autonomy are described. Robot control parameters and/or instructions may advantageously be specified in natural language (NL) and communicated with the LLM via an NL prompt or query. An NL response from the LLM may then be converted into robot control parameters and/or instructions. In this way, an LLM may be leveraged by the robot control system to enhance the autonomy of various operations and/or functions, including without limitation task planning, motion planning, human interaction, and/or reasoning about the environment.

Abstract: A method for controlling a serving robot is provided. The method includes the steps of: acquiring at least one of weight information on a support coupled to the serving robot and image information on the support; recognizing at least one serving object placed on or removed from the support on the basis of at least one of the weight information and the image information; and determining at least one destination of the serving robot on the basis of a result of the recognition and order information of at least one customer in a serving place.

Abstract: Techniques are disclosed that enable model predictive control of a robot based on a latent dynamics model and a reward function. In many implementations, the latent space can be divided into a deterministic portion and stochastic portion, allowing the model to be utilized in generating more likely robot trajectories. Additional or alternative implementations include many reward functions, where each reward function corresponds to a different robot task.

Abstract: The technology disclosed also initializes a new hand that enters the field of view of a gesture recognition system using a parallax detection module. The parallax detection module determines candidate regions of interest (ROI) for a given input hand image and computes depth, rotation and position information for the candidate ROI. Then, for each of the candidate ROI, an ImagePatch, which includes the hand, is extracted from the original input hand image to minimize processing of low-information pixels. Further, a hand classifier neural network is used to determine which ImagePatch most resembles a hand. For the qualified, most-hand like ImagePatch, a 3D virtual hand is initialized with depth, rotation and position matching that of the qualified ImagePatch.

Abstract: Computer-implemented digital image analysis methods, apparatuses, and systems for robotic installation of surgical implants are disclosed. A disclosed apparatus plans a route within an anatomy of a patient from an incision site to a surgical implant site for robotic installation of a surgical implant. The apparatus uses digital imaging data to identify less-invasive installation paths and determine the dimensions of the surgical implant components being used. The apparatus segments the surgical implant into surgical implant subcomponents and modifies the surgical implant subcomponents, such that they can be inserted using the identified less-invasive installation paths.

Abstract: A robot is equipped with a processor. The processor detects external appearance or audio of a living being, and by controlling the robot, causes the robot to execute an operation in accordance with liking data indicating preferences of the robot regarding external appearance or audio and the detected external appearance or audio of the living being.

Abstract: A method for large scene elastic semantic representation and self-supervised light field reconstruction is provided. The method includes acquiring a first depth map set corresponding to a target scene, in which the first depth map set includes a first depth map corresponding to at least one 5 angle of view; inputting the first depth map set into a target elastic semantic reconstruction model to obtain a second depth map set, in which the second depth map set includes a second depth map corresponding to the at least one angle of view; and fusing the second depth map corresponding to the at least one angle of view to obtain a target scene point cloud corresponding to the target scene.

Abstract: A ventilation system with automatic flow balancing derived from a neural network to consistently achieve a desired flow rate for inlet flow and/or outlet flow in various operating environments to optimize system performance. The system includes a ventilation device that includes an exhaust blower assembly with a blower motor and a control circuit having a mathematical equation that determines an estimated exhaust blower flow based upon select inputs. The ventilation device also includes a supply blower assembly with a blower motor and control circuit having a mathematical equation that determines an estimated supply blower flow based upon select inputs. When the estimated exhaust blower flow is different than an exhaust flow set point, the exhaust control circuit selectively alters power supplied to the exhaust motor. When the estimated supply blower flow is different than a supply flow set point, the supply control circuit selectively alters power supplied to the supply motor.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a storage device, for using a robotic device to manipulate a manual control of a device. In one aspect, the system includes a robotic device, a first device that is located at a property and that has a manual control, and a monitoring unit. The monitoring unit may include a network interface, a processor, and a storage device that includes instructions to cause the processor to perform operations. The operations may include determining an operating state of the first device, determining the state of the monitoring system, determining whether one or more of the manual controls associated with the first device should be manipulated to alter the operating state of the first device, and transmitting one or more instruction to the robotic device that instruct the robotic device to manipulate one or more manual controls that are associated the first device.

Abstract: A system providing a barrier for an autonomous floor cleaner includes an artificial barrier generator that radiates one or more infrared signals. An autonomous floor cleaner can be configured to detect the infrared signals, and can react by altering course. Methods for containing an autonomous floor cleaner within a user-determined boundary are disclosed.

Abstract: Devices and techniques are generally described for application determination in speech processing. Input data corresponding to a spoken utterance may be received. Speech recognition processing may be performed on the input data to generate text data. A machine learning encoder may generate a vector representation of the input data. A first binary classifier may determine a first probability that the input data corresponds to a first speech-processing application. A second binary classifier may determine a second probability that the input data corresponds to a second speech-processing application. A selection between the first speech-processing application and the second speech-processing application may be made based at least in part on the first probability and the second probability.

Abstract: A robot system includes: a robot including an end effector connected to an arm thereof; a vision sensor mounted to the robot; and a controller configured to output an operation signal that enables the robot to operate when an input is generated through a touch screen. Each of an object and a target to which the object is placed is inputted through the touch screen. The touch screen displays a recommendation region of the target in a distinguished manner.

Abstract: Techniques for structuring knowledge bases specific to a user or group of users and techniques for using the knowledge bases to answer user inputs are described. A knowledge base may be populated with information provided by users associated with the knowledge base. Users associated with a knowledge base may be proactive in providing content to the knowledge base and/or a system may solicit an answer to a user input from users associated with a particular knowledge base. When the system receives an answer, the system may populate the knowledge base with the answer and may output the answer to the user that originated the user input. The system may output user inputs to be answered using messages or by establishing two-way communication sessions.

Abstract: Computer-implemented digital image analysis methods, apparatuses, and systems for robotic installation of surgical implants are disclosed. A disclosed apparatus plans a route within an anatomy of a patient from an incision site to a surgical implant site for robotic installation of a surgical implant. The apparatus uses digital imaging data to identify less-invasive installation paths and determine the dimensions of the surgical implant components being used. The apparatus segments the surgical implant into surgical implant subcomponents and modifies the surgical implant subcomponents, such that they can be inserted using the identified less-invasive installation paths.

Abstract: A method includes determining motion imitation information for causing a system to imitate a physical task using a first machine learning model that is trained using motion information that represents a performance of the physical task, determining a predicted correction based on the motion information and a current state from the system using a second machine learning model that is trained using the motion information, determining an action to be performed by the system based on the motion imitation information and the predicted correction; and controlling motion of the system in accordance with the action.

Abstract: One embodiment of the invention is directed to a method comprising receiving instruction data relating to a sample in a sample container. The method includes generating, by at least one processor using a workflow management layer, a process plan for the sample, and providing the process plan to a process control layer. The process plan comprises a plurality of possible routes. The method also comprises selecting, by the at least one processor using the process control layer, an optimized route within the plurality of possible routes in the process plan, and providing the optimized route to a middleware control layer. The at least one processor and middleware control layer are operable to cause a transport system to proceed along the selected route.

Abstract: A multi-operation unit integration device having scale expandability and includes a plurality of operation units each of which includes a movable unit; and an integration module. The integration module includes an operation timing unit that gives operation timings of the plurality of operation units based on an operation instruction input from an outside, and the operation unit includes: a plurality of operation learning units that generate a control signal given to the movable unit according to an operation timing signal from the operation timing unit of the integration module; drive means for driving the movable unit of the operation unit according to the control signal; and a sensor that detects a state quantity of the movable unit driven by the drive means. An autonomous learning type robot device is configured using the multi-operation unit integration device as a control portion.

Abstract: The invention relates to an autonomous domestic appliance (1) in the form of a floor cleaning robot or a lawn mower robot with a platform (3) which comprises means of locomotion (4), environmental sensors (5a, 5b, 5c) for detecting properties of the environment of the autonomous domestic appliance (1) and a control unit (6). The environmental sensors (5a, 5b, 5c) are connected to the control unit (6), and the control unit (6) is designed for autonomous movement of the domestic appliance (1) by actuating the means of locomotion (4) as a function of the detected environmental properties and for permanent travel in an area for carrying out its actual domestic activity. The autonomous domestic appliance (1) has a communication unit (7) for establishing a communication link with a remotely arranged service unit (2) and, in addition to its actual domestic activity, is designed to detect the need of help of a person as a function of the environmental properties.

Abstract: An article searching method includes: receiving a search task for searching for an article to be searched; acquiring, based on the search task, a three dimensional model corresponding to the article to be searched; determining a search task group for searching for the article to be searched; and searching for the article to be searched based on the acquired three dimensional model and in combination with the search task group, wherein the search task group shares a search result in the process of searching for the article to be searched.

Abstract: A surveillance performance (information expressing uncertainty of information relating to a potential risk area in a region that requires monitoring) and a tracking performance (information expressing the product of an importance level of a potential risk area and a coverage ratio (an overlap rate between potential risk areas and moving bodies)) are mutually analyzed, and plural moving bodies are allotted to either surveillance or tracking. Potential risk areas that require monitoring are determined such that the surveillance performance and the tracking performance respectively satisfy predetermined levels. The plural moving bodies are allocated to potential risk areas determined to be potential risk areas that require monitoring (simultaneous optimization). The respective moving bodies subject to simultaneous optimization start moving toward their allocated potential risk areas to monitor the potential risk areas.

Abstract: Person recognition device for person re-identification in a monitoring region, having a camera apparatus and an evaluation module, wherein the camera apparatus comprises a first camera unit and a second camera unit, wherein the first camera unit is configured to record a first monitoring image of a portion of the monitoring region, wherein the second camera unit is configured to record a second monitoring image of the portion of the monitoring region, wherein the camera apparatus is configured to feed the first monitoring image and the second monitoring image to the evaluation module, wherein the evaluation module is configured to re-identify a person in the monitoring region based on the first monitoring image and the second monitoring image, wherein the second camera unit has a cut-off filter for wavelength cutting of incident light in a stop band.

Abstract: Described herein is a system and method for providing a conversation session with an artificial intelligence entity that is associated with a business entity. In some aspects, input is provided to an artificial intelligence entity advertisement system. The input is analyzed to determine the subject matter of the input. An artificial intelligence entity associated with the subject matter is then selected and provided to the user. The artificial intelligence entity recommends products or services that are provided by the business entity to the user.

Abstract: Techniques for programmatically interfacing with mobile chat sessions are discussed herein. Some embodiments may include one or more servers configured to: access the mobile chat session hosted by the mobile chat server; receive input chat data from the mobile chat server; determine promotions scores of candidate promotions based at least in part on the input chat data; and provide the output chat data to the mobile chat server including an electronic marketing communication of one or more promotions selected based on the promotion scores. The one or more servers may be further configured to programmatically interact with chat accounts within the mobile chat session, such as to request additional search terms for improved promotion relevance targeting, among other things.

Abstract: Disclosed is a method of controlling a robot, the method including receiving user input of a request for fitting service, moving the robot to a position at which a distance from the user satisfies a predetermined reference distance or more, and a predetermined range or more of the body of the user is scannable, scanning the predetermined range or more of the body of the user through the camera at the moved position, and providing the fitting service by creating a virtual avatar character of the user based on a result of scanning.

Abstract: Methods and systems are provided for high-speed constrained motion planning. In one embodiment, a method includes computing, with a neural network trained on trajectories generated by a non-convex optimizer, a trajectory from one or more initial states of an autonomous system to one or more final states of the autonomous system, updating, with the non-convex optimizer, the trajectory according to kinematic limits and dynamic limits of the autonomous system to obtain a final trajectory, and automatically controlling the autonomous system from an initial state of the one or more initial states to a final state of the one or more final states according to the final trajectory. In this way, efficient and smooth trajectories can be rapidly computed for effective real-time control while accounting for obstacles and physical constraints of an autonomous system.

Abstract: A method includes providing a virtual representation of an environment of a robot. The virtual representation includes an object representation of an object in the environment. The method further includes determining an attribute of the object within the environment of the robot. The attribute includes at least one of occupancy data, force data, and deformation data pertaining to the object. The method further includes receiving a user command to control the robot to move with respect to the object, and modifying a received user command pertaining to the representation of the object based upon the attribute.

Abstract: A ventilation system with automatic flow balancing derived from a neural network to consistently achieve a desired flow rate for inlet flow and/or outlet flow in various operating environments to optimize system performance. The system includes a ventilation device that includes an exhaust blower assembly with a blower motor and a control circuit having a mathematical equation that determines an estimated exhaust blower flow based upon select inputs. The ventilation device also includes a supply blower assembly with a blower motor and control circuit having a mathematical equation that determines an estimated supply blower flow based upon select inputs. When the estimated exhaust blower flow is different than an exhaust flow set point, the exhaust control circuit selectively alters power supplied to the exhaust motor. When the estimated supply blower flow is different than a supply flow set point, the supply control circuit selectively alters power supplied to the supply motor.

Abstract: A system for gaze and gesture detection in unconstrained environments includes a 360-degree (omnidirectional) camera system, one or more depth sensors, and associated memory, processors and programming instructions to determine an object of a human user's attention in the unconstrained environment. The illustrative system may identify the object using eye gaze, gesture detection, and/or speech recognition. The system may generate a saliency map and identify areas of interest. A directionality vector may be projected on the saliency map to find intersecting areas of interest. The system may identify the object of attention once the object of attention is located.

Abstract: An autonomous mobile robot includes a robot body, a movement unit, a detector, and a controller. The movement unit moves the robot body to a destination point. The detector detects a state of a person present around the destination point. The controller controls the movement unit so as to make the autonomous mobile robot approach the person present around the destination point along a travel route that differs depending on the state of the person detected by the detector.

Abstract: According to one embodiment, a holding apparatus includes a holding part, a rotating part, a supporting part, and a moving mechanism. The holding part is configured to hold an article. The rotating part is configured to rotate the holding part. The supporting part is configured to support the article from below. The moving mechanism is configured to move the holding part with respect to the supporting part.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media for evaluating robot learning. In some implementations, one or more computers receive object classification examples from a plurality of robots. Each object classification example includes (i) an embedding that a robot generated using a machine learning model, and (ii) an object classification corresponding to the embedding. The object classification examples are evaluated based on a similarity of the received embeddings with respect to other embeddings. A subset of the object classification examples is selected based on the evaluation of the quality of the embeddings. The subset of the object classification examples is distributed to the robots in the plurality of robots.

Abstract: A robot system may include a robot controller; a target robot; and a host device. The target robot may include a first memory to store device specific data comprising information on the model of the target robot; individual identification data used to identify the target robot; and an individual difference parameter unique to the target robot. The robot controller may include a second memory to store information on the model of the target robot connected to the robot controller, the model configuration information, the individual identification data, and the individual difference parameter; and a controller to control the target robot on the basis of the model configuration information and the individual difference parameter stored in the second memory unit. The controller may read the model configuration information of the model and store the model configuration information in the second memory unit.

Abstract: An information processing system, a control method and a non-transitory storage medium through which emotional values of objects can be numerically expressed based on an interaction between the objects are provided. The information processing system includes a detection unit that detects information related to an interaction between a first object and a second object; and a generation unit that is capable of generating an emotion value of the first object and an emotion value of the second object on the basis of the information related to the interaction.

Abstract: A technology to build emulated nervous systems is presented here, as well as the interface method for operating the emulated nervous system. The technology provides for inclusion of neuroanatomically accurate definitions organized hierarchically. This permits a highly realistic nervous system to be created and interact with its surrounding environment.

Abstract: A method and system for improving the software programming of a robot system, comprising monitoring of plurality of human user-robot interactive pairs' (HURIP) interactions. System comprises each of said plurality of HURIPs as using ‘front-end’ semi-autonomous robot component linked by wireless two-way communications to a ‘back-end’ cloud-based computerized component. Monitoring comprises review of robot sensor-gathered data and data from camera and audio data from homes of users during user-robot interactions. Analysis of said monitoring by authorized observers such as psychologist, parent, teacher, system administrator, software programmer(s), enables identification of areas for software improvement. Improved software is tested, wherein testing comprises at least similar monitoring of HURIPs, and wherein said testing comprises social robots comprising said updated software.

Abstract: The present disclosure provides an eye animated expression display method. The method includes: receiving an instruction for displaying an eye animated expression; parsing a JSON file storing the eye animated emoticon to obtain a parsing result; and displaying the eye animated emoticon on the eye display screen based on the parsing result. The present disclosure further provides a robot. In the above-mentioned manner, the present disclosure is capable of improving the interactive performance of the eyes of the robot while reducing the space for storing eye animated expressions.

Abstract: Techniques for structuring knowledge bases specific to a user or group of users and techniques for using the knowledge bases to answer user inputs are described. A knowledge base may be populated with information provided by users associated with the knowledge base. Users associated with a knowledge base may be proactive in providing content to the knowledge base and/or a system may solicit an answer to a user input from users associated with a particular knowledge base. When the system receives an answer, the system may populate the knowledge base with the answer and may output the answer to the user that originated the user input. The system may output user inputs to be answered using messages or by establishing two-way communication sessions.

Abstract: Techniques for determining an answer that relates to a user question when a system is unable to determine an exact answer to the user question are described. When the system receives a user question, the system may parse the user question to determine the words making up the question. The system may then process the words against an indexed dictionary of words to piece the words of the question together to from interpretations of the different possible relationships that may exist between the words. The system may attempt to determine an exact answer to the user question using the indexed words. If the system determines an exact answer, the system may output the answer to the querying user. If the system is unable to determine an exact answer, the system may navigate through a ontology of relationships within a knowledge base to determine an answer that is related to the user question but that may not directly answer the user question.

Abstract: An apparatus for recognizing a person includes a content separator configured to receive contents and separate the contents into video content and audio content; a video processor configured to recognize a face from an image in the video content received from the content separator and obtain information on a face recognition section by analyzing the video content; an audio processor configured to recognize a speaker from voice data in the audio content received from the content separator and obtain information on a speaker recognition section by analyzing the audio content; and a person recognized section information provider configured to provide information on a section of the contents in which a person appears based on the information on the face recognition section and the information on the speaker recognition section.

Abstract: A method of improving response times associated with responding to requests submitted at one or more front-end systems is disclosed. An utterance is listened for at an intelligent virtual assistant included in the one or more front-end systems. At least one of an intent, a context, and a classification is inferred from the utterance. One or more back-end system commands are generated based on the inferring. The one or more back-end system commands are selected based on machine-learned mappings of the at least one of the intent, the context, and the classification to machine-learned organization-specific pathways into the one or more back-end systems. The one or more back-end system commands are distributed across the one or more back-end systems. A response to the utterance is communicated for presentation via the intelligent virtual assistant, the response including an aggregation of the one or more results received.

Abstract: A system and method to manage communication between a plurality of heterogeneous devices and a cloud has been described. A speaker communication property of a speaker node executing at one of the plurality of heterogeneous devices and the cloud in a first network is compared with a listener communication property of a listener node executing at another of the plurality of devices in a second network. A message-type agnostic listener proxy included in the first network is then generated based on the comparison. Next a message-type agnostic buffer including serialized message corresponding to a message sent by the speaker node is received at the first communication bridge in the first network. Finally, the message-type agnostic speaker proxy forwards the serialized message included in the message-type agnostic buffer to the listener node in the second network.

Abstract: The control system includes a master device control part controlling a master device with a certain number of control axes by using master information and a slave device control part controlling a slave device with a certain number of control axes by using slave information. The control system includes an abstracted master information creating part for creating abstracted master information having a fixed number of elements based on a predetermined manner for allocation and according to the number of elements of the received master information. The slave device control part extracts elements in a number in accordance with the number of elements of the slave information from the fixed number of elements included in the abstracted master information based on a predetermined manner for extraction and according to the number of elements of the slave information to create the slave information.

Abstract: A technology to build emulated nervous systems is presented here, as well as the interface method for operating the emulated nervous system. The technology provides for inclusion of neuroanatomically accurate definitions organized hierarchically. This permits a highly realistic nervous system to be created and interact with its surrounding environment.

Abstract: In one embodiment, a method of processing a natural language input using a conversational knowledge graph in a virtual assistant is disclosed. The method includes receiving a natural language query from a user; translating the natural language query received from the user into corresponding intents; retrieving conversational knowledge context information based on the intents; using the retrieved conversational knowledge context information to customize back-end service calls to downstream applications; receiving a result of the customized back-end service calls; sending the result of the customized back-end service calls in a response to the natural language understanding system; translating the response from the fulfillment service system into a natural language response; and providing the natural language translated response to the user.

Abstract: An embodiment includes a plurality of transport modules forming a transport path on which a carriage travels and a plurality of control units each provided to corresponding one of the plurality of transport modules and configured to perform position control of the carriage in accordance with a pre-stored drive condition and a control start timing, and one of the plurality of control units which corresponds to a first transport module that is one of the plurality of transport modules corrects the drive condition by using a difference between an entry timing at which the carriage enters the first transport module and the control start timing.

Abstract: A processor-implemented method controls a self-driving vehicle (SDV). One or more processors receive, from one or more physical sensors, physical state readings that describe a physical environment of multiple persons that are in spatial proximity with one another. The processors determine, based on the physical state readings, a context of a physical setting of the multiple persons. The processors identify, based on the context of the physical setting of the multiple persons, an SDV that is known to modify a state of the multiple persons. The processor(s) then transmit, to a device controller, a device activation signal to activate the SDV, in order to transport cargo that one of the multiple persons was previously scheduled to transport.

Abstract: A household appliance in the form of a food processor has a controller that controls the preparation of a meal according to a sequence of prescribed steps that includes comminuting, mixing or heating the meal or ingredients. The controller is designed and set up to output a control command for controlling an external cleaning device depending on a progress of the sequence of steps. The control command to the external device includes a cleaning instruction for the external device to effect cleaning in an area where the household appliance is located.

Abstract: A mobile robot includes a receiving unit that receives designation of a destination region including a destination, a moving unit that moves toward the destination region, and a seeking unit that seeks a client after movement toward the destination region starts. The moving unit moves toward the sought client.

Abstract: A location of a transaction or payment request is compared with a location of a user device to determine whether the distance is great enough to send an alert for a possible fraudulent transaction. The user device location may be predicted based on a last known location and information about the area of the last known location and movement of the user device.

Abstract: A device configured to emulate a node in a correlithm object processing system that includes a node engine. The node engine is configured to receive an input correlithm object and to determine distances between the input correlithm object and source correlithm objects in a node table. A correlithm object is a point in an n-dimensional space represented by a binary string. The distance between the input correlithm object and a source correlithm object is determined based on differences between a binary string representing the input correlithm object and binary strings linked with the source correlithm objects. The node engine is configured to identify a source correlithm object from the node table with the shortest distance, to fetch a target correlithm object from the node table linked with the identified source correlithm object, and to output the identified target correlithm object.