Message platform for autonomous entities
A computer-implemented method can receive a message sent from a source entity, perform first pre-processing operations for verifying validity of the message, perform second pre-processing operations for determining a category of the message, extract metadata from the message, generate an enriched message comprising the metadata and the determined category, perform post-processing operations for classifying the enriched message into one of a plurality of event types, broadcast the enriched message to a message broker, and routing, by the message broker, the enriched message to one or more target entities registered an event type into which the message is classified.
1. A computer-implemented method for improved entity communication automation in a heterogenous computing environment where a plurality of entities resides in a distributed network, the method comprising:
receiving a message sent from a source entity in the heterogenous computing environment, wherein the message comprises a plurality of original attributes and values paired with the plurality of original attributes, wherein the plurality of original attributes indicates one or more operating conditions of the source entity;
performing, by a pre-processing engine of a message broker, first pre-processing operations comprising:
verifying validity of the message; and
deriving one or more additional attributes of the message based on evaluating the plurality of original attributes and their paired values using one or more predefined derivative rules;
generating, by the pre-processing engine of the message broker, an enriched message, wherein generating the enriched message comprises:
adding the one or more additional attributes to the message;
performing second pre -processing operations for determining a topic of the message based on measuring similarity between the message and a plurality of topic clusters and a non-topic cluster that were generated based on a text corpus; and
adding to the message a new attribute paired with the topic of the message;
performing, by a post-processing engine of the message broker, post-processing operations, wherein the post-processing operations comprise classifying the enriched message into an event type selected from a plurality of event types;
broadcasting the enriched message by the message broker; and
routing, by the message broker, the enriched message to one or more target entities in the heterogenous computing environment, wherein the one or more target entities have registered the event type.
2. The method of claim 1 , wherein verifying validity of the message comprises checking compliance of the message against one or more predefined validity rules.
3. The method of claim 1 , further comprising generating the plurality of topic clusters from the text corpus using latent Dirchlet analysis and creating the non-topic cluster different from the plurality of topic clusters, wherein the determined topic corresponds to one of the topic clusters or the non-topic cluster.
4. The method of claim 1 , further comprising generating a warning responsive to determining the determined topic corresponds to the non -topic cluster.
5. The method of claim 1 , wherein determining the topic of the message comprises calculating similarity indexes between the message and the plurality of topic clusters and the non-topic cluster using contrastive learning.
6. The method of claim 5 , wherein determining the topic of the message further comprises selecting a topic cluster or non-topic cluster having the maximum similarity index as the determined topic.
7. The method of claim 1 , further comprising training a classifier using previously received messages and classified event types corresponding to the received messages.
8. A computing system for improved entity communication automation in a heterogenous computing environment where a plurality of entities resides in a distributed network, the computing system comprising:
memory;
one or more hardware processors coupled to the memory; and
one or more computer readable storage media storing instructions that, when loaded into the memory, cause the one or more hardware processors to perform operations comprising:
receiving a message sent from a source entity in the heterogenous computing environment, wherein the message comprises a plurality of original attributes and values paired with the plurality of original attributes, wherein the plurality of original attributes indicates one or more operating conditions of the source entity;
performing, by a pre-processing engine of a message broker, first pre-processing operations comprising:
verifying validity of the message; and
deriving one or more additional attributes of the message based on evaluating the plurality of original attributes and their paired values using one or more predefined derivative rules;
generating, by the pre-processing engine of the message broker, an enriched message, wherein generating the enriched message comprises:
adding the one or more additional attributes to the message;
performing second pre -processing operations for determining a topic of the message based on measuring similarity between the message and a plurality of topic clusters and a non-topic cluster that were generated based on a text corpus; and
adding to the message a new attribute paired with the topic of the message;
performing, by a post-processing engine of the message broker, post-processing operations, wherein the post-processing operations comprise classifying the enriched message into an event type selected from a plurality of event types;
broadcasting the enriched message by the message broker; and
routing, by the message broker, the enriched message to one or more target entities in the heterogenous computing environment, wherein the one or more target entities have registered the event type.
9. The system of claim 8 , wherein verifying validity of the message comprises checking compliance of the message against one or more predefined validity rules.
10. The system of claim 8 , wherein the operations further comprise generating the plurality of topic clusters from the text corpus using latent Dirchlet analysis and creating the non-topic cluster different from the plurality of topic clusters, wherein the determined topic corresponds to one of the topic clusters or the non-topic cluster.
11. The system of claim 8 , wherein the operations further comprise generating a warning responsive to determining the determined topic corresponds to the non-topic cluster.
12. The system of claim 8 , wherein determining the topic of the message comprises calculating similarity indexes between the message and the plurality of topic clusters and the non-topic cluster using contrastive learning.
13. The system of claim 12 , wherein determining the topic of the message further comprises selecting a topic cluster or non-topic cluster having the maximum similarity index as the determined topic.
14. The system of claim 8 , wherein the operations further comprise training a classifier using previously received messages and classified event types corresponding to the received messages.
15. One or more non-transitory computer-readable media having encoded thereon computer-executable instructions causing one or more processors to perform a method for improved entity communication automation in a heterogenous computing environment where a plurality of entities resides in a distributed network, the method comprising:
receiving a message sent from a source entity in the heterogenous computing environment, wherein the message comprises a plurality of original attributes and values paired with the plurality of original attributes, wherein the plurality of original attributes indicates one or more operating conditions of the source entity;
performing, by a pre-processing engine of a message broker, first pre-processing operations comprising:
verifying validity of the message; and
deriving one or more additional attributes of the message based on evaluating the plurality of original attributes and their paired values using one or more predefined derivative rules;
generating, by the pre-processing engine of the message broker, an enriched message, wherein generating the enriched message comprises:
adding the one or more additional attributes to the message;
performing second pre -processing operations for determining a topic of the message based on measuring similarity between the message and a plurality of topic clusters and a non-topic cluster that were generated based on a text corpus; and
adding to the message a new attribute paired with the topic of the message;
performing, by a post-processing engine of the message broker, post-processing operations, wherein the post-processing operations comprise classifying the enriched message into an event type selected from a plurality of event types;
broadcasting the enriched message by the message broker; and
routing, by the message broker, the enriched message to one or more target entities in the heterogenous computing environment, wherein the one or more target entities have registered the event type,
wherein performing first pre-processing operations comprises:
checking compliance of the message against one or more predefined validity rules; and
generating a warning responsive to finding that the message violates one of the predefined validity rules.
16. The one or more non-transitory computer-readable media of claim 15 , wherein the method further comprises generating the plurality of topic clusters from the text corpus using latent Dirchlet analysis and creating the non-topic cluster different from the plurality of topic clusters, wherein the determined topic corresponds to one of the topic clusters or the non-topic cluster.
17. The one or more non-transitory computer-readable media of claim 15 , wherein determining the topic of the message comprises calculating similarity indexes between the message and the plurality of clusters and the non-topic cluster using contrastive learning.
18. The one or more non-transitory computer-readable media of claim 15 , wherein determining the topic of the message further comprises selecting a topic cluster or non-topic cluster having the maximum similarity index as the determined topic.