Abstract: Methods, systems and computer program products are described to improve machine learning (ML) model-based classification of data items by identifying and removing inaccurate training data. Inaccurate training samples may be identified, for example, based on excessive variance in vector space between a training sample and a mean of category training samples, and based on a variance between an assigned category and a predicted category for a training sample. Suspect or erroneous samples may be selectively removed based on, for example, vector space variance and/or prediction confidence level. As a result, ML model accuracy may be improved by training on a more accurate revised training set. ML model accuracy may (e.g., also) be improved, for example, by identifying and removing suspect categories with excessive (e.g., weighted) vector space variance. Suspect categories may be retained or revised. Users may (e.g., also) specify a prediction confidence level and/or coverage (e.g., to control accuracy).

Abstract: Methods, systems and computer program products are described to improve machine learning (ML) model-based classification of data items by identifying and removing inaccurate training data. Inaccurate training samples may be identified, for example, based on excessive variance in vector space between a training sample and a mean of category training samples, and based on a variance between an assigned category and a predicted category for a training sample. Suspect or erroneous samples may be selectively removed based on, for example, vector space variance and/or prediction confidence level. As a result, ML model accuracy may be improved by training on a more accurate revised training set. ML model accuracy may (e.g., also) be improved, for example, by identifying and removing suspect categories with excessive (e.g., weighted) vector space variance. Suspect categories may be retained or revised. Users may (e.g., also) specify a prediction confidence level and/or coverage (e.g., to control accuracy).

Abstract: Methods, systems and computer program products are described to improve machine learning (ML) model-based classification of data items by identifying and removing inaccurate training data. Inaccurate training samples may be identified, for example, based on excessive variance in vector space between a training sample and a mean of category training samples, and based on a variance between an assigned category and a predicted category for a training sample. Suspect or erroneous samples may be selectively removed based on, for example, vector space variance and/or prediction confidence level. As a result, ML model accuracy may be improved by training on a more accurate revised training set. ML model accuracy may (e.g., also) be improved, for example, by identifying and removing suspect categories with excessive (e.g., weighted) vector space variance. Suspect categories may be retained or revised. Users may (e.g., also) specify a prediction confidence level and/or coverage (e.g., to control accuracy).

Abstract: Embodiments described herein are directed to improving machine learning (ML) model-based techniques for automatically labeling data items based on identifying and resolving labels that are problematic. An ML model may be trained to predict labels for any given data item. The ML model may be validated to determine a confusion metric with respect to each distinct pair of labels predicted by the ML model. Each confusion metric indicates how a particular label is being mistaken for another particular label. The confusion metrics are analyzed to determine whether any of the ML model-generated labels are problematic (e.g., a label conflicts with another label, a label that is rarely predicted, a label that is incorrectly predicted, etc.). Steps for resolving the problematic labels are implemented, and the ML model is retrained based on the resolution steps. By doing so, the ML model generates a more accurate label for a data item.

Abstract: Embodiments described herein are directed to generating a machine learning (ML) model. A plurality of vectors are accessed, each vector of the plurality of vectors including a first set of features associated with a corresponding data item. A second set of features is identified by expanding the first set of features. A ML model is trained using vectors including the expanded set of features, and it is determined that an accuracy of the ML model trained using the vectors increased. A third set of features is identified by determining a measure of importance for different subsets of features in the second set and replacing subsets having a low measure of importance with new features. A ML model is trained using vectors that include the third set, and it is determined that an accuracy of the model increased due to the replacing.

Abstract: Embodiments described herein are directed to generating a machine learning (ML) model. A plurality of vectors are accessed, each vector of the plurality of vectors including a first set of features associated with a corresponding data item. A second set of features is identified by expanding the first set of features. A ML model is trained using vectors including the expanded set of features, and it is determined that an accuracy of the ML model trained using the vectors increased. A third set of features is identified by determining a measure of importance for different subsets of features in the second set and replacing subsets having a low measure of importance with new features. A ML model is trained using vectors that include the third set, and it is determined that an accuracy of the model increased due to the replacing.

Abstract: An approach is provided for large scale vehicle routing. The approach involves, for example, receiving a plurality of plans, wherein a plan of the plurality of plans assigns a vehicle, a driver of the vehicle, or a combination thereof a set of rides to traverse. The approach also involves clustering the plurality of plans into one or more clusters based on a proximity measure. The proximity measure indicates a proximity of a first plan of the plurality of plans to a second plan of a plurality of plans. The approach further involves, for each cluster of the one or more clusters, separately computing a solution to a multiple vehicle routing problem for the set of rides in said each cluster.

Abstract: A method of determining a value for an apnea-hypopnea index (AHI) for a person, the method comprising: recording a voice track of a person; extracting features from the voice track that characterize the voice track; and processing the features to determine an AHI.

Abstract: Various aspects of a method, a system, and a computer program product are disclosed herein in accordance with at least one example embodiment for spatial clustering of a plurality of cells of a geographical location. The method may include generation of cell clusters based on multiple criteria. The method may further include reception of mobility data from one or more external devices associated with the plurality of cells. The mobility data may be further processed to extract observations for generation of feature vectors from the observations. The feature vectors may be normalized to generate one or more clusters of a set of cells of the location. Cluster data associated with the clustering of the plurality of cells may be further used in one or more location based applications.

Abstract: Methods, systems and computer program products are described to improve machine learning (ML) model-based classification of data items by identifying and removing inaccurate training data. Inaccurate training samples may be identified, for example, based on excessive variance in vector space between a training sample and a mean of category training samples, and based on a variance between an assigned category and a predicted category for a training sample. Suspect or erroneous samples may be selectively removed based on, for example, vector space variance and/or prediction confidence level. As a result, ML model accuracy may be improved by training on a more accurate revised training set. ML model accuracy may (e.g., also) be improved, for example, by identifying and removing suspect categories with excessive (e.g., weighted) vector space variance. Suspect categories may be retained or revised. Users may (e.g., also) specify a prediction confidence level and/or coverage (e.g., to control accuracy).

Abstract: Embodiments described herein are directed to generating a machine learning (ML) model. A plurality of vectors are accessed, each vector of the plurality of vectors including a first set of features associated with a corresponding data item. A second set of features is identified by expanding the first set of features. A ML model is trained using vectors including the expanded set of features, and it is determined that an accuracy of the ML model trained using the vectors increased. A third set of features is identified by determining a measure of importance for different subsets of features in the second set and replacing subsets having a low measure of importance with new features. A ML model is trained using vectors that include the third set, and it is determined that an accuracy of the model increased due to the replacing.

Abstract: An approach is provided for large scale vehicle routing. The approach involves, for example, receiving a plurality of plans, wherein a plan of the plurality of plans assigns a vehicle, a driver of the vehicle, or a combination thereof a set of rides to traverse. The approach also involves clustering the plurality of plans into one or more clusters based on a proximity measure. The proximity measure indicates a proximity of a first plan of the plurality of plans to a second plan of a plurality of plans. The approach further involves, for each cluster of the one or more clusters, separately computing a solution to a multiple vehicle routing problem for the set of rides in said each cluster.

Abstract: Embodiments described herein are directed to improving machine learning (ML) model-based techniques for automatically labeling data items based on identifying and resolving labels that are problematic. An ML model may be trained to predict labels for any given data item. The ML model may be validated to determine a confusion metric with respect to each distinct pair of labels predicted by the ML model. Each confusion metric indicates how a particular label is being mistaken for another particular label. The confusion metrics are analyzed to determine whether any of the ML model-generated labels are problematic (e.g., a label conflicts with another label, a label that is rarely predicted, a label that is incorrectly predicted, etc.). Steps for resolving the problematic labels are implemented, and the ML model is retrained based on the resolution steps. By doing so, the ML model generates a more accurate label for a data item.

Abstract: The present invention provides a method, apparatus, and computer program product for generating synthetic demand data of vehicle rides corresponding to a first location. The method comprises obtaining historic demand data of vehicle rides corresponding to a second location. The historic demand data comprises at least one first point of interest (POI) associated with at least one of a pick-up event or a drop-off event. The method further comprises retrieving map data associated with the second location and further, determining at least one second POI associated with the first location. The method further comprises generating the synthetic demand data of vehicle rides corresponding to the first location, based on the at least one first POI, the map data, and the at least one second POI.

Abstract: The present invention provides a method, apparatus, and computer program product for generating synthetic demand data of vehicle rides corresponding to a first location. The method comprises obtaining historic demand data of vehicle rides corresponding to a second location. The historic demand data comprises at least one first point of interest (POI) associated with at least one of a pick-up event or a drop-off event. The method further comprises retrieving map data associated with the second location and further, determining at least one second POI associated with the first location. The method further comprises generating the synthetic demand data of vehicle rides corresponding to the first location, based on the at least one first POI, the map data, and the at least one second POI.

Abstract: Various aspects of a method, a system, and a computer program product are disclosed herein in accordance with at least one example embodiment for spatial clustering of a plurality of cells of a geographical location. The method may include generation of cell clusters based on multiple criteria. The method may further include reception of mobility data from one or more external devices associated with the plurality of cells. The mobility data may be further processed to extract observations for generation of feature vectors from the observations. The feature vectors may be normalized to generate one or more clusters of a set of cells of the location. Cluster data associated with the clustering of the plurality of cells may be further used in one or more location based applications.

Abstract: A system and method for searching for an element in speech related documents may include transcribing a set of speech recordings to a set of phoneme strings and including the phoneme strings in a set of phonetic transcriptions. A system and method may reverse-index the phonetic transcriptions according to one or more phonemes such that the one or more phonemes can be used as a search key for searching the phoneme in the phonetic transcriptions. A system and method may transcribe a textual search term into a set of search phoneme strings and use the set of search phoneme strings to search for an element in the set of phonetic transcriptions.

Abstract: A system and method for searching for an element in speech related documents may include transcribing a set of speech recordings to a set of phoneme strings and including the phoneme strings in a set of phonetic transcriptions. A system and method may reverse-index the phonetic transcriptions according to one or more phonemes such that the one or more phonemes can be used as a search key for searching the phoneme in the phonetic transcriptions. A system and method may transcribe a textual search term into a set of search phoneme strings and use the set of search phoneme strings to search for an element in the set of phonetic transcriptions.

Abstract: A method of determining a value for an apnea-hypopnea index (AHI) for a person, the method comprising: recording a voice track of a person; extracting features from the voice track that characterize the voice track; and processing the features to determine an AHI.