Abstract: A method of completing a task, the task being of a given type of task. The method includes receiving, by a server, an indication of a first result of the task having been completed by a human assessor, executing by the server a machine learning algorithm (MLA) to complete the task by the MLA to generate a second result of the task, determining, by the server, a confidence level parameter indicative of a probability of the first result being correct, and determining, by the server, whether the probability indicated by the determined confidence level parameter exceeds a pre-defined threshold probability. In response to determining that the probability indicated by the determined confidence level parameter exceeds the pre-defined threshold probability, the server processes the task as having been completed with the first result.

Abstract: Methods and servers for training a Machine Learning Algorithm (MLA) for translation of text are disclosed. The MLA has been trained using a first plurality of string pairs. The first plurality of string pairs has a natural proportion of string pairs of each context. The MLA is biased to generate the given parallel string as a translation of the respective string occurred in the main context. The method includes determining a second plurality of string pairs comprising a controlled proportion of string pairs of each context. The second plurality of string pairs are associated with labels indicative of the respective contexts. The method comprises re-training the MLA using the second plurality of string pairs and the respective labels. The MLA is re-trained to determine a given context of a given string and generate a respective parallel string as a translation having considered the given context.

Abstract: There is disclosed a device for cooling a server rack within a server room having a first chamber and a second chamber, the device comprising: a main chilling unit; a heat exchanging unit operatively coupled to the main chilling unit; the main chilling unit including a housing for housing: an evaporator; a first airlock device to removably secure the main chilling unit to the server rack from a side of the first chamber; a first fan configured to force air from the first chamber to the second chamber; the heat exchanging unit including a housing for housing: a condenser in fluid communication with the evaporator, the condenser being configured to transmit heat from the liquid coolant to air surrounding the condenser; a second airlock device configured to removably secure the heat exchanging unit to the second chamber; a second fan configured to blow the heated air into the second chamber.

Abstract: A method of processing a search query, the method comprising receiving a search query; determining whether the search query is stored in an offline query index by comparing the search query against one or more offline search queries stored therein; responsive to determining that the search query matches a given offline search query, determining an offline SERP template pointer associated with the given offline search query; determining, based on the offline SERP template pointer, an associated offline SERP template stored within an offline SERP template index; retrieving a list of one or more ordered element pointers comprising the associated offline SERP template; retrieving, for each one or more ordered element pointers comprising the list, a respective SERP element, from the at least one SERP elements index; generating an offline SERP by compiling the retrieved SERP elements; and displaying the offline SERP on the electronic device.

Abstract: A computer-implemented method of and a system for generating a prediction model and determining an accuracy parameter of a trained decision tree prediction model. The method comprises accessing the trained decision tree prediction model having been generated at least partially based on a set of training objects; generating a subset of random parameters of interest; associating the subset of random parameters of interests with a given leaf; determining a leaf accuracy parameter based on (i) the parameters of interest associated with the given leaf and (ii) the subset of random parameters of interest of the given leaf; and determining the accuracy parameter of the trained decision tree prediction model based on the determined leaf accuracy parameter for each of the leafs of the decision tree.

Abstract: There is disclosed a method and system for determining users submitting abnormal organization ratings. The method comprises retrieving organization rating activity. The method comprises generating a graph of the organization rating activity and determining a complete subgraph in the graph. The method then comprises determining whether the ratings for the subgraph satisfy a threshold distribution, and after determining that the ratings do satisfy the threshold distribution, storing an indication that the users who submitted those ratings are associated with abnormal organization ratings.

Abstract: A tray for accommodating electronic components is disclosed. The tray has a body that defines a storage space for accepting electronic components in a side-by-side configuration. The body also defines a first airflow channel for guiding airflow from the storage space towards the outside of the body. The body has air inlets and air outlets for providing fluid communication with the storage space. The component bodies of the electronic components are configured to at least partially define a second airflow channel in the storage space for guiding, in use, the airflow from the air inlets to the air outlets, and has been selected for increasing a surface area of the first electronic component being in contact, in use, with the airflow in the storage space.

Abstract: There is disclosed a method for identifying malicious activity, the method being executable by a supervisory electronic device. The method comprises accessing, a log to retrieve usage information associated with at least a portion of the plurality of electronic devices; analyzing the usage information to identify a subset of electronic devices; analyzing a list of network resources accessed; executing, by the supervisory electronic device a polling robot, the polling robot configured to: transmit to each of the list of network resources a ping message, the ping message having a first pre-determined format having been generated based on the malicious activity; analyzing, the response message; responsive to the response message having a second pre-determined format, the second pre-determined format having been identified based on the pre-determined type of malicious activity: determining that an associated network resource having generated the response message is associated with malicious activity.

Abstract: A method and system for ranking digital object based on an objective characteristic associated therewith are provided. The method comprising: generating a set of digital objects based on a user request, the set of digital objects being rankable according to an objective characteristic thereof; receiving a filter request from the user, the filter request being based on a secondary characteristic of digital objects in the set of digital objects; determining object parameters for the digital objects in the set of digital objects, a given object parameter being indicative of a likelihood that an inclusion of a respective digital object in a re-ranked set of digital objects will increase a quality metric of the re-ranked set of digital objects; selecting digital objects based on object parameters; ranking digital objects based on respective values of the secondary characteristic, thereby generating the re-ranked set of digital objects.

Abstract: A method and a system for determining an importance score associated with a digital news item, the importance score for determining push notifications to be provided to a user of an electronic device. The method comprises accessing a digital news item, the digital news item having been electronically published at a news item publication time by a digital news source, the digital news source being accessible via a communication network; determining a source weight of the digital news source; determining the importance score of the digital news item based on the news item publication time and the source weight of the corresponding digital news source; in response to the importance score of the digital news item being above an importance score threshold, generating a push notification based on the digital news item; and transmitting the push notification by the server to the electronic device.

Abstract: A method and a server for generating a meta-feature for ranking documents by a machine learning algorithm (MLA). A past query having been previously submitted on a server is acquired, and a set of past documents having been presented as search results in response to the past query is acquired, where each respective document includes a plurality of features, and respective values for the plurality of features. The meta-feature is generated, where a respective value of the meta-feature for a respective document is based on: a respective value of a given feature of the plurality of features for the respective document, and a value of a parameter associated with the set of past documents. The meta-feature is validated based on its usefulness for ranking future search engine results pages (SERPs). The MLA is then trained to generate the meta-feature for ranking documents in response to a new query.

Abstract: An air flow control method and system for cooling a data center including two server rooms is disclosed. The method includes receiving a first differential pressure value; receiving a second differential pressure value from the second differential pressure gauge; generating a first control signal to adjust the opening of the first damper based on the first differential pressure value; generating a second control signal to adjust the opening of the second damper based on the second differential pressure value.

Abstract: A method and server for estimating interaction data are disclosed. A user and a digital content item that the user has interacted with form an occurred pair. The method includes retrieving (i) interaction data of a first type associated with respective ones of the occurred pairs, and (ii) interaction data of a second type associated with respective ones of the occurred pairs. The first type is distinct from the second type. The method also includes (i) applying a first model to the interaction data of the first type, thereby estimating interaction data of the first type for non-occurred pairs, and (ii) applying a second model to the interaction data of the second type, thereby estimating interaction data of the second type for the non-occurred pairs. The interaction data is estimated such that the non-occurred pairs are associated with respective estimated interaction data of the first type and second type.

Abstract: A method of determining a user-centric traffic estimation error parameter associated with an estimated road traffic condition that is electronically provided to a user of a device. the method comprises: at a first moment in time, acquiring an estimated travel time for a road segment, the estimated travel time having been computed for the first moment in time; responsive to the device approaching the road segment, displaying an application-generated estimated travel time for the road segment, the application-generated estimated travel time being based on a most recently acquired estimated travel time from the server; responsive to the device departing from the road segment, determining an actual travel time for the road segment; computing the user-centric traffic estimation error parameter based on the application-generated estimated travel time and the actual travel time; and transmitting the user-centric traffic estimation error parameter to the server for adjusting a traffic prediction algorithm.

Abstract: A method for processing a request, where the request is of a declaratory-type and is destined from a request source to a distributed database that has a plurality of transaction destination locations (TDLs). The distributed database is part of distributed computer system, and the distributed computer system has a supervisory device. The method is executable by the supervisory device and includes: receiving the request, processing the request to generate a first transaction of a read-type and identifying at least one condition-satisfying TDL, acquiring a lock value for the condition-satisfying TDL, executing the first transaction, processing the request to generate a last transaction of a modification-type for modifying information associated with the at least one TDL, before executing the last transaction, checking a status of the lock, and only executing the last transaction based on a positive outcome. A distributed software system for doing the same is also provided.

Abstract: There is a method and system for translating speech to text, the speech having been received by a client device. A user utterance corresponding to the speech is received. A first predicted text corresponding to the user utterance and a first confidence score corresponding to the first predicted text are determined using a local graph. The user utterance is transmitted to a server. A second predicted text corresponding to the user utterance and a second confidence score corresponding to the second predicted text is received from the server. If the first confidence score is higher than the second confidence score, the first predicted text is output.

Abstract: A method and a system for generating a digital content recommendation. The method comprises receiving a request for the digital content recommendation. Based on the request, a first content item and a second content item responsive to the request are selected, and a relevancy parameter and a completion parameter for each of the first content item and the second content item are determined. Based on the relevancy parameter and the completion parameter, the first content item and the second content item are ranked, and a digital content recommendation is generated based on the ranking of the first content item and the second content item.

Abstract: A method and system for meeting a service level of content item promotion. The method is executable by a server hosting a recommendation service for providing digital content items to a user, the user being one of a plurality of users of the service. Storage communicatively coupled to the server contains data representative of the digital content items, and of the history of interactions of the plurality of users of the service with such digital content items. The method includes upon the server receiving a request for a promotion of a digital content item with a requested service level, calculating a threshold of probability within a distribution of users based on their likelihood that they take action over the digital content item, and providing the digital content item to users within the plurality of users of the service having a probability within the distribution that is above the threshold.