Contextual integrity preservation
An example operation includes one or more of receiving, by a data processing node, inference data object from a multi-channel data server over a blockchain, sorting, by the data processing node, longitudinal records contained in the inference data object, linking, by the data processing node, transaction outcomes and inferences data from the inference data object to the sorted longitudinal records, and recording linked data onto a blockchain ledger. The data processing node serves as a validator of data from a robo-advisory using natural language (NL) processing to reduce bias and measure effectiveness of inference from the robo-advisory.
1 . A system, comprising:
a processor of a data processing node; and
a memory comprising machine readable instructions that, when executed by the processor, configure the processor to:
receive data from a plurality of communication channels between a client and the data processing node, wherein
the plurality of communication channels comprises an IVR channel and a chatbot channel between a user associated with the client and the data processing node;
generate a plurality of predicted insights from the received data based on first level insights, second level insights, and an output from an execution of a plurality of machine learning (ML) models on the received data, wherein
the plurality of predicted insights includes the first level insights and the second level insights in a multi-layered architecture, and
generation of the second level insights is based on the first level insights;
receive feedback from the user for the plurality of predicted insights;
link the plurality of predicted insights, the plurality of ML models that generated the plurality of predicted insights, and the feedback from the user to generate a proof chain;
record the proof chain on a blockchain ledger in the memory;
modify variables of an ML model of the plurality of ML models based on data from the proof chain to iteratively train the ML model; and
record an updated proof chain to the blockchain ledger after each iterative training in the iteratively training of the ML model, wherein the updated proof chain comprises details that describe how the ML model was trained.
2 . The system of claim 1 , wherein the processor is further configured to:
link together the received data from the plurality of communication channels with the plurality of predicted insights in the proof chain.
3 . The system of claim 1 , wherein the received data comprises robo-advisory data.
4 . The system of claim 1 , wherein the received data comprises data from a plurality of different types of communication channels.
5 . The system of claim 1 , wherein the processor is further configured to link the plurality of predicted insights and the feedback from the user within the proof chain.
6 . The system of claim 1 , wherein
the processor is further configured to identify a mapping of specific user terms with standard financial terms based on the data, and
the mapping of the specific user terms with the standard financial terms is used for the generation of the second level insights.
7 . A method, comprising:
receiving, by a data processing node, data from a plurality of communication channels between a client and the data processing node, wherein
the plurality of communication channels comprises an IVR channel and a chatbot channel between a user associated with the client and the data processing node;
generating a plurality of predicted insights from the received data based on first level insights, second level insights, and an output from an execution of a plurality of machine learning (ML) models on the received data, wherein
the plurality of predicted insights includes the first level insights and the second level insights in a multi-layered architecture, and
generation of the second level insights is based on the first level insights;
receiving feedback from the user for the plurality of predicted insights;
linking the plurality of predicted insights, the plurality of ML models that generated the plurality of predicted insights, and the feedback from the user to generate a proof chain;
recording the proof chain on a blockchain ledger;
modifying variables of an ML model of the plurality of ML models based on data from the proof chain to iteratively train the ML model; and
recording an updated proof chain to the blockchain ledger after each iterative training in the iteratively training of the ML model, wherein the updated proof chain comprises details describing how the ML model was trained.
8 . The method of claim 7 , further comprising:
linking together the received data from the plurality of communication channels with the plurality of predicted insights in the proof chain.
9 . The method of claim 7 , wherein the received data comprises robo-advisory data.
10 . The method of claim 7 , wherein the received data comprises data from a plurality of different types of communication channels.
11 . The method of claim 7 , further comprising:
linking the plurality of predicted insights and the feedback from the user within the proof chain.
12 . A non-transitory computer-readable medium comprising instructions that, when executed by a processor, cause the processor to perform:
receiving data from a plurality of communication channels between a client and a data processing node, wherein
the plurality of communication channels comprises an IVR channel and a chatbot channel between a user associated with the client and the data processing node;
generating a plurality of predicted insights from the received data based on first level insights, second level insights, and an output from an execution of a plurality of machine learning (ML) models on the received data, wherein
the plurality of predicted insights includes the first level insights and the second level insights in a multi-layered architecture, and
generation of the second level insights is based on the first level insights;
receiving feedback from the user for the plurality of predicted insights;
linking the plurality of predicted insights, the plurality of ML models that generated the plurality of predicted insights, and the feedback from the user to generate a proof chain;
recording the proof chain on a blockchain ledger;
modifying variables of an ML model of the plurality of ML models based on data from the proof chain to iteratively train the ML model; and
recording an updated proof chain to the blockchain ledger after each iterative training in the iteratively training of the ML model, wherein the updated proof chain comprises details describing how the ML model was trained.
13 . The non-transitory computer-readable medium of claim 12 , wherein the instructions further cause the processor to perform:
linking together the received data from the plurality of communication channels with the plurality of predicted insights in the proof chain.
14 . The non-transitory computer-readable medium of claim 12 , wherein the received data comprises robo-advisory data.
15 . The non-transitory computer-readable medium of claim 12 , wherein the received data comprises data from a plurality of different communication channels.
16 . The non-transitory computer-readable medium of claim 7 , wherein the instructions further cause the processor to perform:
linking together the plurality of predicted insights and the feedback from the user within the proof chain.