Abstract: The invention provides a process for treating waste water from an industrial process for producing propylene oxide, which comprises subjecting the waste water to a catalytic wet oxidation treatment comprising: feeding a stream comprising the waste water to a reactor; subjecting the stream comprising the waste water to an oxidation treatment in the presence of oxygen and a catalyst resulting in a treated stream; sending at least part of the treated stream to a separator wherein the treated stream is separated into a gas stream and a liquid stream; and recycling part of the liquid stream to the reactor.

Abstract: The invention provides a process for treating waste water from an industrial process for producing propylene oxide, which comprises subjecting the waste water to a catalytic wet oxidation treatment comprising: feeding a stream comprising the waste water to a reactor; subjecting the stream comprising the waste water to an oxidation treatment in the presence of oxygen and a catalyst resulting in a treated stream; sending at least part of the treated stream to a separator wherein the treated stream is separated into a gas stream and a liquid stream; and recycling part of the liquid stream to the reactor.

Abstract: Methods and systems are described for converting a matrix to a sparse vector-based matrix utilizing one or more of a global identifier, a cohort identifier, an n-tuple representation, and a sparse vector. Methods and systems are described for partitioning matrices. Methods and systems are described for managing execution of tasks in a distributed computing environment. Methods and systems are described for positioning data within the distributed computing environment.

Abstract: A method for making entity recommendations to users of a social networking system (SNS) by maximizing a predicted benefit to the SNS through objective value models. The method includes generating a plurality of candidate entities eligible for recommendation, and determining a weight for each candidate entity based upon an output score generated by an objective value model. The objective value model utilizes objectively measurable input values describing the candidate entities and generates larger output scores for those candidate entities that would create larger predicted increases in an objectively measurable amount of benefit for the SNS created by a potential connection being created between the user and the respective candidate entity. One or more entities are selected based upon the determined weights, and an entity recommendation user interface is presented to the user that includes one or more recommendation modules identifying the selected one or more entities.

Abstract: A method for making entity recommendations to users of a social networking system (SNS) by maximizing a predicted benefit to the SNS through objective value models. The method includes generating a plurality of candidate entities eligible for recommendation, and determining a weight for each candidate entity based upon an output score generated by an objective value model. The objective value model utilizes objectively measurable input values describing the candidate entities and generates larger output scores for those candidate entities that would create larger predicted increases in an objectively measurable amount of benefit for the SNS created by a potential connection being created between the user and the respective candidate entity. One or more entities are selected based upon the determined weights, and an entity recommendation user interface is presented to the user that includes one or more recommendation modules identifying the selected one or more entities.