Abstract: An approach is provided for machine learning-based registration of imagery with different perspectives. The approach, for example, involves retrieving a first training image and a second training image. The first training image depicts a geographic area from a first perspective and the second training image depicts the geographic area from a second perspective. The approach also involves initiating a labeling of one or more ground truth correspondence masks between the first training image and the second training image. The one or more ground truth correspondence masks denote an image region of the first training image that matches a corresponding image region of the second training image or vice versa. The approach further involves using the one or more ground truth correspondence masks to train a machine learning model to determine one or more predicted correspondence masks between a first input image and a second input image.

Abstract: An approach is provided for progressive training of long-lived, evolving machine learning architectures. The approach involves, for example, determining alternative paths for the evolution of the machine learning model from a first architecture to a second architecture. The approach also involves determining one or more migration step alternatives in the alternative paths. The migration steps, for instance, include architecture options for the evolution of the machine learning model. The approach further involves processing data using the options to determine respective model performance data. The approach further involves selecting a migration step from the one or more migration step alternatives based on the respective model performance data to control a rate of migration steps over a rate of training in the evolution of the machine learning model. The approach further involves initiating a deployment the selected migration step to the machine learning model.

Abstract: A method, apparatus and computer program product are provided for learning to generate maps from raw geospatial observations from sensors traveling within an environment.

Abstract: A method, apparatus and computer program product are provided for learning to generate maps from raw geospatial observations from sensors traveling within an environment.

Abstract: A method, apparatus and computer program product are provided for learning to generate maps from raw geospatial observations from sensors traveling within an environment.

Abstract: A method, apparatus and computer program product are provided for learning to generate maps from raw geospatial observations from sensors traveling within an environment. Methods may include: receiving a plurality of sequences of geospatial observations from discrete trajectories; refining a summary entity set representation through iteration over discrete trajectories; determining a drive offset for each of the discrete trajectories based on a comparison of the summary entity set representation to a drive entity set for each of the discrete trajectories; aligning the discrete trajectories to generate aligned geospatial observations based, at least in part, on the drive offset for a respective discrete trajectory; concatenating the aligned geospatial observations; processing the concatenated, aligned geospatial observations using at least one Set Transformer; and generating, from the at least one Set Transformer, map geometries including objects from the geospatial observations.

Abstract: A method, apparatus and computer program product are provided to incentivize crowd sourcing of data by identifying and compensating content contributors based on a value of the content to training a neural network. Methods may include: receiving a request; processing the request using a machine learning model to generate a response to the request; based on the processing of the request using the machine learning model, identifying training data contributing to the response to the request; identifying one or more data contributors as providing the identified training data contributing to the response to the request; and providing a response to the request and an indication of the one or more data contributors.

Abstract: An approach is provided for deep learning of sparse spatial data functions. The approach involves, for instance, creating a sort convolutional neural network (SortCNN) layer comprising a multi-head cross-attention layer and one or more convolutional neural network (CNN) layers. At least one attention head of the multi-head cross-attention layer is associated with at least one linear projection matrix that is trained to arrange and quantize an unsorted set of input entities (e.g., sparse data) along an axis of a query/key space into a soft sorted set of the input entities based on inducing points in the query/key space. The approach also involves projecting the soft sorted entities from the multi-head cross-attention layer through the CNN layers. The CNN layers learn one or more functions based on integrating information from the soft sorted entities as arranged and quantized by the at least one linear projection matrix.

Abstract: A method, apparatus and computer program product are provided for learning to generate maps from raw geospatial observations from sensors traveling within an environment. Methods may include: receiving a plurality of sequences of geospatial observations from discrete trajectories; aligning the trajectories to generate aligned geospatial observations; concatenating the aligned geospatial observations; processing the concatenated, aligned geospatial observations using one or more Set Transformers; generating, from the at least one Set Transformer, map geometries including objects from the geospatial observations; and providing at least one of navigational assistance or at least semi-autonomous vehicle control based on the map geometries. According to some embodiments, aligning the trajectories includes applying a geospatial offset for one or more of the trajectories.

Abstract: An approach is provided for biasing machine learning models towards potential risks for controlling vehicles/robots. The approach involves, for example, determining an occluded space that is occluded in sensor data collected from one or more sensors of a vehicle or a robot. The approach also involves generating a sensor space completion that represents the occluded space based on biasing a generation of one or more potential risks to the vehicle or the robot originating from the occluded space. The approach further involves providing the sensor space completion to a system of the vehicle or the robot for generating a control decision, a warning, or a combination thereof.

Abstract: A method, apparatus, and computer program product are provided for using confidential computing to execute code on sensitive data in an encrypted area of an apparatus limiting access to data and code to only their respective owners. Methods may include: generating an outer enclave and at least one inner enclave within the outer enclave; providing an outer enclave key and an inner enclave key to a service provider; providing an inner enclave key to a data provider; receiving, from the data provider, a data retrieval location; processing data from the respective retrieval location at the data provider inner enclave using data provider code to generate data provider processed data; providing the data provider processed data to the service provider inner enclave; and processing the data provider processed data with service provider code to generate resultant data; decrypting the resultant data in the outer enclave.

Abstract: Methods described herein relate to updating pipeline operations for data processing. The method includes receiving pipeline information for at least one of a plurality of pipelines. The pipeline information includes at least one of an input dataset, output dataset, input model, intermediate model, or output model. The method also includes determining one or more of the plurality of pipelines to update based on similarities with the pipeline information received for at least one of the plurality of pipelines. The method further includes updating the one or more of the plurality of pipelines based on the pipeline information received. Updating the pipeline includes updating at least one of the input model, intermediate model, or output model. The method still further includes storing the one or more updated pipelines.

Abstract: An approach is provided for progressive training of long-lived, evolving machine learning architectures. The approach involves, for example, determining alternative paths for the evolution of the machine learning model from a first architecture to a second architecture. The approach also involves determining one or more migration step alternatives in the alternative paths. The migration steps, for instance, include architecture options for the evolution of the machine learning model. The approach further involves processing data using the options to determine respective model performance data. The approach further involves selecting a migration step from the one or more migration step alternatives based on the respective model performance data to control a rate of migration steps over a rate of training in the evolution of the machine learning model. The approach further involves initiating a deployment the selected migration step to the machine learning model.

Abstract: Methods described herein relate to updating pipeline operations for data processing. The method includes receiving pipeline information for at least one of a plurality of pipelines. The pipeline information includes at least one of an input dataset, output dataset, input model, intermediate model, or output model. The method also includes determining one or more of the plurality of pipelines to update based on similarities with the pipeline information received for at least one of the plurality of pipelines. The method further includes updating the one or more of the plurality of pipelines based on the pipeline information received. Updating the pipeline includes updating at least one of the input model, intermediate model, or output model. The method still further includes storing the one or more updated pipelines.

Abstract: A method, apparatus and computer program product are provided to incentivize crowd sourcing of data by identifying and compensating content contributors based on a value of the content to training a neural network. Methods may include: receiving a request for a machine learning model trained from training data received from a plurality of data contributors, where the training data identified a contributor having provided the respective training data; processing the request for the machine learning model to infer a result based on a subset of training data relevant to the request; identifying one or more data contributors that provided the subset of training data relevant to the request; and providing compensation to the one or more data contributors that provided the subset of training data.

Abstract: A method, apparatus and computer program product are provided to incentivize crowd sourcing of data by identifying and compensating content contributors based on a value of the content to training a neural network. Methods may include: receiving a request; processing the request using a machine learning model to generate a response to the request; based on the processing of the request using the machine learning model, identifying training data contributing to the response to the request; identifying one or more data contributors as providing the identified training data contributing to the response to the request; and providing a response to the request and an indication of the one or more data contributors.