Abstract: An Artificial Intelligence (AI)-based innovation data processing system receives at least one query word related to a category. Information material including textual and non-textual data is retrieved from a plurality of data sources using the at least one query word. The information material is tokenized and parsed using a dependency parser for entity recognition, building entity relationships and for generating knowledge graphs. The output of the dependency parser is accessed by a trained classifier for obtaining respective confidence levels for each of the sentences in the textual data. The confidence levels are compared to a predetermined threshold confidence level for determining if the sentences include references to innovations. In addition, trends in the innovations are determined and responses to user queries are generated based on one or more of knowledge graphs and the trends.

Abstract: A machine learning (ML) based automated search system receives an entity information document including a plurality of entities and identifies new entities that are similar to the plurality of entities which are not included in the entity information document via automated searches. Entity intelligence reports are generated for the plurality of entities which are further used to extract search terms. The search terms are used for executing automatic searches for documents with relevant portions. The documents are further analyzed to identify other, new entities which are not included in the entity information document. Entity intelligence reports including information regarding the new entities are also generated. Significant attributes are determined for the entities. The significant attributes are further employed in ranking the new entities so that top ranked new entities are identified. Alerts can be generated based on the information regarding the new entities.

Abstract: An Artificial Intelligence (AI)-based innovation data processing system receives at least one query word related to a category. Information material including textual and non-textual data is retrieved from a plurality of data sources using the at least one query word. The information material is tokenized and parsed using a dependency parser for entity recognition, building entity relationships and for generating knowledge graphs. The output of the dependency parser is accessed by a trained classifier for obtaining respective confidence levels for each of the sentences in the textual data. The confidence levels are compared to a predetermined threshold confidence level for determining if the sentences include references to innovations. In addition, trends in the innovations are determined and responses to user queries are generated based on one or more of knowledge graphs and the trends.

Abstract: A system for categorizing words into clusters includes a receiver to receive a set of sentences formed by a plurality of words. The set of sentences is indicative of interaction of a user with a virtual assistant. A categorizer categorizes the plurality of words into a first set of clusters by using a first clustering technique, and categorizes the plurality of words into a second set of clusters by using a second clustering technique. A detector detects words that appear in similar clusters after categorization by the first clustering technique and the second clustering technique. Similarity of clusters is based on a nature of words forming the clusters. A generator generates a confidence score for each of the plurality of words based on the detection. The confidence score of a word is indicative of accuracy of the categorization of the word.

Abstract: A machine learning (ML) based automated search system receives an entity information document including a plurality of entities and identifies new entities that are similar to the plurality of entities which are not included in the entity information document via automated searches. Entity intelligence reports are generated for the plurality of entities which are further used to extract search terms. The search terms are used for executing automatic searches for documents with relevant portions. The documents are further analyzed to identify other, new entities which are not included in the entity information document. Entity intelligence reports including information regarding the new entities are also generated. Significant attributes are determined for the entities. The significant attributes are further employed in ranking the new entities so that top ranked new entities are identified. Alerts can be generated based on the information regarding the new entities.

Abstract: A system for categorizing words into clusters includes a receiver to receive a set of sentences formed by a plurality of words. The set of sentences is indicative of interaction of a user with a virtual assistant. A categorizer categorizes the plurality of words into a first set of clusters by using a first clustering technique, and categorizes the plurality of words into a second set of clusters by using a second clustering technique. A detector detects words that appear in similar clusters after categorization by the first clustering technique and the second clustering technique. Similarity of clusters is based on a nature of words forming the clusters. A generator generates a confidence score for each of the plurality of words based on the detection. The confidence score of a word is indicative of accuracy of the categorization of the word.

Abstract: Systems and methods for determining critical thresholds on a number of events (k) and a window length (t) for properly defining a burst of events in a data stream. A new coverage metric Ck,t is defined and used in the determination, where the coverage metric Ck,t is defined for a particular pair (k,t) as a fraction, with the numerator defined a number of events that occur within some (k,t)-bursty window and the denominator defined as the total number of events (n) that occurred along the entire time span being analyzed. Coverage metric Ck,t is monotonic non-increasing in k and monotonic non-decreasing in t, allowing for a divide-and-conquer search strategy to be used to find the critical threshold pairs (k*, t*).

Abstract: Systems and methods for determining critical thresholds on a number of events (k) and a window length (t) for properly defining a burst of events in a data stream. A new coverage metric Ck,t is defined and used in the determination, where the coverage metric Ck,t is defined for a particular pair (k,t) as a fraction, with the numerator defined a number of events that occur within some (k,t)-bursty window and the denominator defined as the total number of events (n) that occurred along the entire time span being analyzed. Coverage metric Ck,t is monotonic non-increasing in k and monotonic non-decreasing in t, allowing for a divide-and-conquer search strategy to be used to find the critical threshold pairs (k*, t*).