In an approach to generating advice for supply chain disruptions, one or more computer processors receive a query associated with a supply chain disruption. The one or more computer processors retrieve data corresponding to the supply chain disruption. Based on the retrieve data, the one or more computer processors determine one or more solutions to the supply chain disruption. The one or more computer processors display the one or more determined solutions. The one or more computer processors receive a selection of one of the one or more determined solutions. The one or more computer processors detect one or more patterns associated with the selected solution.

A method may include generating, based a context-free grammar, a sample forming a corpus. The context-free grammar may include production rules for replacing a first nonterminal symbol with a second nonterminal symbol and/or a terminal symbol. The sample may be generated by rewriting recursively a first text string to form a second text string associated with the sample. The first text string may be rewritten by applying the production rules to replace nonterminal symbols included in the first text string until no nonterminal symbols remain in the first text string. A machine learning model may be trained, based on the corpus, to process a natural language. Related methods and articles of manufacture are also disclosed.

A system and method includes receiving a first context update for an application. The method determines whether one or more of multiple rules has been satisfied in view of the first context update. The multiple rules include a multiple conditions and are associated with multiple contextual situations. The determining includes minimizing a number of the multiple conditions to be evaluated to determine whether a particular rule of the multiple rules has been satisfied. The method responsive to determining a first rule of the multiple rules has been satisfied, identifies a first contextual situation of the multiple contextual situations that is associated with the first rule. The method also determines at least one action that is associated with the first contextual situation.

According to an embodiment, a learning device includes one or more hardware processors configured to: acquire a current state of a device; learn a reinforcement learning model, and determine a first action of the device on the basis of the current state and the reinforcement learning model; determine a second action of the device on the basis of the current state and a first rule; and select one of the first action and the second action as a third action to be output to the device according to a progress of learning of the reinforcement learning model.

A system, method and computer program product automating time management, includes an automated time management framework using an AI engine to for making trade offs among tasks as unexpected events occur. Each day is divided into blocks of time called Skeds. Once a Sked has begun, scheduled tasks and tasks that have been manually or automatically added to the Sked are treated as a collection. It is determined if the Sked is balanced based on total available labor minutes compared time to perform uncompleted tasks. If not, a score is calculated for each uncompleted task based on its ratings, including priority, movability, optionality, difficulty, and/or unpleasantness of the task. The uncompleted tasks are ranked using the scores, and abandoned or moved to later Skeds based on the ranking until the Sked is balanced. The balancing process is repeated until the Sked is balanced and, if not, until an end thereof.

A method, apparatus, and system provide the ability to optimize execution of an application. An application is acquired. The application includes functions, and each function has a corresponding feature flag that determines whether the corresponding function is executed. Execution conditions of execution of the application are monitored at run-time (in a machine learning module). The machine learning module recognizes a pattern relating to the execution conditions to determine a stress relating to the execution of the application. During execution of the application, the machine learning module toggles the feature flags based on the pattern and the stress such that the corresponding functions do not execute.

A system, method and computer program product automating time management, includes an automated time management framework using an AI engine to for making trade offs among tasks as unexpected events occur. Each day is divided into blocks of time called Skeds. Once a Sked has begun, scheduled tasks and tasks that have been manually or automatically added to the Sked are treated as a collection. It is determined if the Sked is balanced based on total available labor minutes compared time to perform uncompleted tasks. If not, a score is calculated for each uncompleted task based on its ratings, including priority, movability, optionality, difficulty, and/or unpleasantness of the task. The uncompleted tasks are ranked using the scores, and abandoned or moved to later Skeds based on the ranking until the Sked is balanced. The balancing process is repeated until the Sked is balanced and, if not, until an end thereof.

An apparatus that searches an input stream having a sequence of N-bit wide data words for a pattern using a plurality of small FSMs is disclosed. The apparatus includes a plurality of sub-word FSMs and a combiner. Each sub-word FSM has an input word size less than N-bits. Each FSM processes a corresponding segment of the N-bit words and generates a match output indicative of a possible match to the pattern when one of the input words to that FSM is received and that FSM moves to a predetermined match state. The combiner receives the match outputs from all of the sub-word FSMs and generates a pattern match output if all of the sub-word FSMs indicate a match to the pattern. The pattern is a variable pattern. In one embodiment, the FSMs are single bit FSMs.

An apparatus that searches an input stream having a sequence of N-bit wide data words for a pattern using a plurality of small FSMs is disclosed. The apparatus includes a plurality of sub-word FSMs and a combiner. Each sub-word FSM has an input word size less than N-bits. Each FSM processes a corresponding segment of the N-bit words and generates a match output indicative of a possible match to the pattern when one of the input words to that FSM is received and that FSM moves to a predetermined match state. The combiner receives the match outputs from all of the sub-word FSMs and generates a pattern match output if all of the sub-word FSMs indicate a match to the pattern. The pattern is a variable pattern. In one embodiment, the FSMs are single bit FSMs.

The technology disclosed herein provides a summary of a predicted timeline for an event that has yet to occur. In a particular implementation, a method provides identifying a first event that has yet to occur. The method further provides identifying first data objects from a plurality of data objects obtained from a plurality of information sources. The first data objects include information pertinent to the first event. The method also provides extracting first time information relevant to the first event from the first data objects, determining a confidence level for each portion of the first time information, and generating a summary of the first time information based on the confidence level for each portion of the first time information.