Distribution of tasks among microservices in a copilot
Recorded sessions of skilled personnel at work are used to train a first machine learning (ML) tool to extract workflow maps. These maps are used to train a second ML tool to emulate at least one workflow. The first ML tool is trained to predict an annotator's output. Either ML tool can be a copilot having a microservice network architecture. Further, complex sets of workflows can be subdivided for efficient support with small ML tools for (1) workflow map extraction and (2) emulation. Based on segment demarcation accompanying a recording, segments are assigned to specialized ML extraction tools for workflow map extraction. Extracted workflow maps are used to train workflow emulator(s). Similarly, specialized ML emulation tools can emulate respective tasks. A distribution microservice can identify a task to be emulated and can invoke the appropriate specialized ML tool. Similar microservice network architectures support both applications.
1 . One or more non-transitory computer-readable media storing instructions executable by one or more hardware processors, wherein the instructions include:
a plurality of modules which, when executed, implement respective microservices among a weakly connected network of the microservices, which is configured as a copilot to one or more client applications;
wherein the network of microservices includes a distribution microservice and a plurality of task microservices, each task microservice comprising a respective trained machine learning (ML) tool;
wherein the task microservices, including at least distinct first and second task microservices, are trained to perform respective tasks, the tasks including at least distinct first and second tasks; and
wherein the distribution microservice is configured to:
receive input from a given one of the client applications indicating a particular task to be performed;
identify the particular task as one among the respective tasks;
in a first case for which the particular task is identified as the first task, forward the particular task toward the first task microservice, causing the first task microservice to perform the first task; and
in a second case for which the particular task is identified as the second task, forward the particular task toward the second task microservice, causing the second task microservice to perform the second task.
2 . The one or more non-transitory computer-readable media of claim 1 , wherein the instructions further comprise:
a training module configured to train, upon execution, the first task microservice using a plurality of training data records, each training data record comprising a workflow map corresponding to the first task, wherein the workflow map comprises a description of one or more procedures for accomplishing the first task.
3 . The one or more non-transitory computer-readable media of claim 2 , wherein the workflow map is an input workflow map, the first task is extraction of an output workflow map from a recorded session, and the each training data record further comprises:
the recorded session of one or more experts performing at least the first task.
4 . The one or more non-transitory computer-readable media of claim 2 , wherein the first task is emulation of a workflow associated with the workflow map.
5 . The one or more non-transitory computer-readable media of claim 1 , wherein the distribution microservice is another trained ML tool.
6 . The one or more non-transitory computer-readable media of claim 5 , wherein the instructions further comprise:
a training module configured to train, upon execution, the distribution microservice using a plurality of training data records, each training data record comprising:
a recorded session of one or more experts performing one or more among the tasks; and
an annotator's output identifying, for each of one or more segments of the recorded session, a respective task among the tasks.
7 . The one or more non-transitory computer-readable media of claim 1 , wherein the copilot further comprises an expansion microservice, a retrieval microservice, a core microservice, and an evaluation microservice.
8 . The one or more non-transitory computer-readable media of claim 1 , wherein a training corpus comprises distinct first and second segments of each of a plurality of interview scripts, and the instructions further include:
one or more first training modules configured to train, upon execution, the first task microservice using the first segments as training data records; and
one or more second training modules configured to train, upon execution, the second task microservice using the second segments as training data records.
9 . The one or more non-transitory computer-readable media of claim 1 , wherein a training corpus comprises distinct first and second segments of each of a plurality of recorded work sessions, and the instructions further include:
one or more first training modules configured to train, upon execution, the first task microservice using the first segments as training data records; and
one or more second training modules configured to train, upon execution, the second task microservice using the second segments as training data records.
10 . The one or more non-transitory computer-readable media of claim 1 , wherein the first task microservice is configured to transmit a result of performing the first task toward a receiver external to the copilot.
11 . The one or more non-transitory computer-readable media of claim 10 , wherein the receiver is a first receiver and the second task microservice, distinct from the first task microservice, is configured to transmit a result of performing the second task toward a second receiver within the copilot.
12 . A system comprising:
one or more hardware processors with memory coupled thereto; and
computer readable media storing instructions executable by the one or more hardware processors, wherein the instructions include:
a plurality of modules which, when executed, implement respective microservices among a weakly connected network of the microservices, which is configured as a copilot to one or more client applications;
wherein the network of microservices includes a distribution microservice and a plurality of task microservices, each task microservice comprising a respective trained machine learning (ML) tool;
wherein the task microservices, including at least distinct first and second task microservices, are trained to perform respective tasks, the tasks including at least distinct first and second tasks; and
wherein the distribution microservice is configured to:
receive input from a given one of the client applications indicating a particular task to be performed;
identify the particular task as one among the respective tasks;
in a first case for which the particular task is identified as the first task, forward the particular task toward the first task microservice, causing the first task microservice to perform the first task; and
in a second case for which the particular task is identified as the second task, forward the particular task toward the second task microservice, causing the second task microservice to perform the second task.
13 . The system of claim 12 , wherein the instructions further comprise:
a training module configured to train, upon execution, the first task microservice using a plurality of training data records, each training data record comprising a workflow map corresponding to the first task, wherein the workflow map comprises a description of one or more procedures for accomplishing the first task.
14 . The system of claim 13 , wherein the workflow map is an input workflow map, the first task is extraction of an output workflow map from a recorded session, and the each training data record further comprises:
the recorded session of one or more experts performing at least the first task.
15 . The system of claim 13 , wherein the first task is emulation of a workflow associated with the workflow map.
16 . The system of claim 12 , wherein the distribution microservice is another trained ML tool.
17 . The system of claim 16 , wherein the instructions further comprise:
a training module configured to train, upon execution, the distribution microservice using a plurality of training data records, each training data record comprising:
a recorded session of one or more experts performing one or more among the tasks; and
an annotator's output identifying, for each of one or more segments of the recorded session, a respective task among the tasks.
18 . The system of claim 12 , wherein a training corpus comprises distinct first and second segments of each of a plurality of interview scripts, and the instructions further include:
one or more first training modules configured to train, upon execution, the first task microservice using the first segments as training data records; and
one or more second training modules configured to train, upon execution, the second task microservice using the second segments as training data records.
19 . The system of claim 12 , wherein a training corpus comprises distinct first and second segments of each of a plurality of recorded work sessions, and the instructions further include:
one or more first training modules configured to train, upon execution, the first task microservice using the first segments as training data records; and
one or more second training modules configured to train, upon execution, the second task microservice using the second segments as training data records.
20 . The system of claim 12 , wherein the first task microservice is configured to transmit a result of performing the first task toward a receiver external to the copilot.