Abstract: Techniques for determining a classification probability of an object in an environment are discussed herein. Techniques may include analyzing sensor data associated with an environment from a perspective, such as a top-down perspective, using multi-channel data. From this perspective, techniques may determine channels of multi-channel input data and additional feature data. Channels corresponding to spatial features may be included in the multi-channel input data and data corresponding to non-spatial features may be included in the additional feature data. The multi-channel input data may be input to a first portion of a machine-learned (ML) model, and the additional feature data may be concatenated with intermediate output data from the first portion of the ML model, and input into a second portion of the ML model for subsequent processing and to determine the classification probabilities. Additionally, techniques may be performed on a multi-resolution voxel space representing the environment.

Abstract: Techniques for determining a classification probability of an object in an environment are discussed herein. Techniques may include analyzing sensor data associated with an environment from a perspective, such as a top-down perspective, using multi-channel data. From this perspective, techniques may determine channels of multi-channel input data and additional feature data. Channels corresponding to spatial features may be included in the multi-channel input data and data corresponding to non-spatial features may be included in the additional feature data. The multi-channel input data may be input to a first portion of a machine-learned (ML) model, and the additional feature data may be concatenated with intermediate output data from the first portion of the ML model, and input into a second portion of the ML model for subsequent processing and to determine the classification probabilities. Additionally, techniques may be performed on a multi-resolution voxel space representing the environment.

Abstract: Techniques to predict object behavior in an environment are discussed herein. For example, such techniques may include determining a trajectory of the object, determining an intent of the trajectory, and sending the trajectory and the intent to a vehicle computing system to control an autonomous vehicle. The vehicle computing system may implement a machine learned model to process data such as sensor data and map data. The machine learned model can associate different intentions of an object in an environment with different trajectories. A vehicle, such as an autonomous vehicle, can be controlled to traverse an environment based on object's intentions and trajectories.

Abstract: Techniques for determining a response of a simulated vehicle to a simulated object in a simulation are discussed herein. Log data captured by a physical vehicle in an environment can be received. Object data representing an object in the log data can be used to instantiate a simulated object in a simulation to determine a response of a simulated vehicle to the simulated object. Additionally, one or more trajectory segments in a trajectory library representing the log data can be determined and instantiated as a trajectory of the simulated object in order to increase the accuracy and realism of the simulation.

Abstract: Techniques to predict object behavior in an environment are discussed herein. For example, such techniques may include inputting data into a model and receiving an output from the model representing a discretized representation. The discretized representation may be associated with a probability of an object reaching a location in the environment at a future time. A vehicle computing system may determine a trajectory and a weight associated with the trajectory using the discretized representation and the probability. A vehicle, such as an autonomous vehicle, can be controlled to traverse an environment based on the trajectory and the weight output by the vehicle computing system.

Abstract: Techniques for determining a classification probability of an object in an environment are discussed herein. Techniques may include analyzing sensor data associated with an environment from a perspective, such as a top-down perspective, using multi-channel data. From this perspective, techniques may determine channels of multi-channel input data and additional feature data. Channels corresponding to spatial features may be included in the multi-channel input data and data corresponding to non-spatial features may be included in the additional feature data. The multi-channel input data may be input to a first portion of a machine-learned (ML) model, and the additional feature data may be concatenated with intermediate output data from the first portion of the ML model, and input into a second portion of the ML model for subsequent processing and to determine the classification probabilities. Additionally, techniques may be performed on a multi-resolution voxel space representing the environment.

Abstract: Techniques for determining a classification probability of an object in an environment are discussed herein. Techniques may include analyzing sensor data associated with an environment from a perspective, such as a top-down perspective, using multi-channel data. From this perspective, techniques may determine channels of multi-channel input data and additional feature data. Channels corresponding to spatial features may be included in the multi-channel input data and data corresponding to non-spatial features may be included in the additional feature data. The multi-channel input data may be input to a first portion of a machine-learned (ML) model, and the additional feature data may be concatenated with intermediate output data from the first portion of the ML model, and input into a second portion of the ML model for subsequent processing and to determine the classification probabilities. Additionally, techniques may be performed on a multi-resolution voxel space representing the environment.

Abstract: Examples of the present disclosure describe systems and methods for using AI to identify regions of interest (ROI) in medical images. In aspects, medical reports and images may be provided to a first environment. The first environment may use the medical report data/medical images to train a natural language processing (NLP)-based algorithm to identify the location in images of ROI described in the medical report data. The output of the NLP-based algorithm may be stored in an ROI repository in the first environment. After the NLP-based algorithm has been trained, a request to train a user-specific model may be received in a second environment. Data objects for the requested user-specific model may be provided to the first environment, which uses the ROI repository to train the model. The trained model may be provided to the second environment, where the trained user-specific model/algorithm may be tested and stored.

Abstract: Examples of the present disclosure describe systems and methods for using AI to identify regions of interest (ROI) in medical images. In aspects, medical reports and images may be provided to a second environment. The second environment may use the medical report data/medical images to train a natural language processing (NLP)-based algorithm to identify the location in images of ROI described in the medical report data. The output of the NLP-based algorithm may be stored in an ROI repository in the second environment. After the NLP-based algorithm has been trained, a request to train a user-specific model may be received in a first environment. Data objects for the requested user-specific model may be provided to the second environment, which uses the ROI repository to train the model. The trained model may be provided to the first environment, where the trained user-specific model/algorithm may be tested and stored.

Abstract: Techniques to predict object behavior in an environment are discussed herein. For example, such techniques may include determining a trajectory of the object, determining an intent of the trajectory, and sending the trajectory and the intent to a vehicle computing system to control an autonomous vehicle. The vehicle computing system may implement a machine learned model to process data such as sensor data and map data. The machine learned model can associate different intentions of an object in an environment with different trajectories. A vehicle, such as an autonomous vehicle, can be controlled to traverse an environment based on object's intentions and trajectories.

Abstract: Techniques to predict object behavior in an environment are discussed herein. For example, such techniques may include inputting data into a model and receiving an output from the model representing a discretized representation. The discretized representation may be associated with a probability of an object reaching a location in the environment at a future time. A vehicle computing system may determine a trajectory and a weight associated with the trajectory using the discretized representation and the probability. A vehicle, such as an autonomous vehicle, can be controlled to traverse an environment based on the trajectory and the weight output by the vehicle computing system.

Abstract: A method may include obtaining a task-specific state of a robotic device associated with performing a task by the robotic device. The method may also include mapping the task-specific state to a general state of the robotic device, where the general state may applicable to multiple tasks and may include a variable number of elements. The method may additionally include obtaining a set of generic actions represented by a set of parameters, where the set of generic actions may be used to perform the multiple tasks. The method may also include processing the general state of the robotic device to solve for the set of parameters and to predict a next action in performing the task, where the next action may be selected from the set of generic actions, and performing the next action using the set of parameters solved for.

Abstract: A bot network system may be provided. A system may include a plurality of bot agents, wherein at least one bot agent of the plurality of bot agents is configured to receive a request from a user in natural language. The system may further include a plurality of digital resources including one or more of a software program, a service, a web service and a dataset. Each digital resource of the plurality of digital resources may be configured to communicate with a dedicated bot agent of the plurality of bot agents. Also, each bot agent may be configured to interact with its associated digital resource via an application programming interface (API) of the associated digital resource and translate between the natural language and a language of the associated digital resource.

Abstract: A method may include obtaining a task-specific state of a robotic device associated with performing a task by the robotic device. The method may also include mapping the task-specific state to a general state of the robotic device, where the general state may applicable to multiple tasks and may include a variable number of elements. The method may additionally include obtaining a set of generic actions represented by a set of parameters, where the set of generic actions may be used to perform the multiple tasks. The method may also include processing the general state of the robotic device to solve for the set of parameters and to predict a next action in performing the task, where the next action may be selected from the set of generic actions, and performing the next action using the set of parameters solved for.

Abstract: A bot network system may be provided. A system may include a plurality of bot agents, wherein at least one bot agent of the plurality of bot agents is configured to receive a request from a user in natural language. The system may further include a plurality of digital resources including one or more of a software program, a service, a web service and a dataset. Each digital resource of the plurality of digital resources may be configured to communicate with a dedicated bot agent of the plurality of bot agents. Also, each bot agent may be configured to interact with its associated digital resource via an application programming interface (API) of the associated digital resource and translate between the natural language and a language of the associated digital resource.

Abstract: A method may include obtaining first images, each image including an object, and determining a set of one or more visual cues for each. The method may include selecting a common visual cue of the one or more visual cues that is common to each set of one or more visual cues determined for each corresponding image and determining a correlation between a location of the common visual cue in each image of the first images and a location of the object in each image of the first images. The method may include obtaining a second image of an environment and identifying the common visual cue in the second image. The method may include determining a placement location for the object in the environment based on the correlation and a location of the common visual cue in the second image.

Abstract: A method may include hardwiring: a first dynamic input of M slices in a section of a sliced architecture to receive a main data sample; and a second dynamic input of each of X and Y slices to respectively receive a first or second delayed data sample, X, Y being subsets of M. A slice current may be multiplied with: the data sample in each of A of the M slices; and the first delayed data sample in each of B of the X slices. The method may also include summing: outputs of the A slices to obtain a weighted output current of the data sample; outputs of the B slices to obtain a weighted output current of the first delayed data sample; and the weighted output currents of the main data sample and of the first delayed data sample to obtain a net weighted output current of the section.

Abstract: A method may include obtaining current sensor information that corresponds to one or more sensors of a wheelchair and that indicates a current state of the wheelchair and its surrounding environment. The method may further include obtaining current situation information that indicates a current situation of a wheelchair and its corresponding user. In addition, the method may include determining current solution information that corresponds to a current solution to help navigate the current situation. The current solution information may be determined based on one or more of the following: the current situation information and the current sensor information. The method may additionally include implementing the current solution based on the current solution information.

Abstract: A method may include obtaining current sensor information that corresponds to one or more sensors of a wheelchair and that indicates a current state of the wheelchair and its surrounding environment. The method may further include obtaining current situation information that indicates a current situation of a wheelchair and its corresponding user. In addition, the method may include determining current solution information that corresponds to a current solution to help navigate the current situation. The current solution information may be determined based on one or more of the following: the current situation information and the current sensor information. The method may additionally include implementing the current solution based on the current solution information.