# US 2021/0232936 A1 · Automatically extracting feature engineering knowledge from execution traces

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2021/0232936 A1
- Document ID: US20210232936A1
- Status: Published application
- Kind: A1
- Application number: 16/752,191
- Filed: 2020-01-24
- Published: 2021-07-29
- Earliest claimed: 2020-01-24
- Assignees: International Business Machines Corporation
- Applicants: International Business Machines Corporation
- Inventors: Takaaki Tateishi
- Web page: https://patentgrep.com/patent/US20210232936A1

## CPC

- G06F 8/433: {Dependency analysis; Data or control flow analysis}
- G06F 16/9024: {Graphs; Linked lists (G06F16/9027 takes precedence)}
- G06N 5/022: Knowledge engineering; Knowledge acquisition
- G06N 5/04: Inference or reasoning models
- G06N 20/00: Machine learning
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

A method is provided for creating a directed graph. The method obtains an execution trace having a sequence of invocations of table transformations. Each invocation is classified as a call-with-object invocation or call-with-column invocation. The call-with-object invocation indicates that an invoked function has object IDs as at least one of input and output information. The call-with-column invocation indicates that the invoked function has column names as the at least one of input and output information. The method processes the trace to build the graph. Each node indicates an object or a column and each edge indicates data flow. The method performs, in response to all invocations being processed, an automated graph reduction process by eliminating given nodes indicating a respective object such that that every two nodes that connect with each other, via a respective one of the given nodes as an intermediate node, can keep the connection therebetween.

## Claims

### Claim 1

A computer-implemented method for creating and reducing a directed graph from an execution trace, comprising:
obtaining, by a hardware processor, the execution trace having a sequence of invocations of table transformations, each of the invocations being classified as a call-with-object invocation or as a call-with-column invocation, the call-with-object invocation indicating that an invoked function has object IDs as at least one of input and output information and the call-with-column invocation indicating that the invoked function has column names as the at least one of input and output information;
processing, by the hardware processor, the execution trace to build the directed graph, wherein each node in the directed graph indicates an object or a column and each edge indicates data flow; and
performing, by the hardware processor in response to all the invocations being processed, an automated graph reduction process on the directed graph by eliminating from the directed graph given nodes indicating a respective object such that that every two nodes that connect with each other, via a respective one of the given nodes as an intermediate node, can keep the connection therebetween.

### Claim 2 (depends on 1)

The computer-implemented method of claim 1, wherein the column names had by the invoked function as the at least one of input and output information comprise column names taken by the invoked function as the input information and column names newly created by the invoked function as the output information.

### Claim 3 (depends on 2)

The computer-implemented method of claim 2, wherein the column names newly created comprise new features, and the method further comprises applying the new features in a machine learning pipeline to generate a prediction.

### Claim 4 (depends on 1)

The computer-implemented method of claim 1, in responsive to a particular invocation being processed being a call-with-column invocation, said processing step comprises (i) creating a new node labeled with a column and the input and output columns, (ii) connecting the new node with another node, and (iii) labelling an edge between the new node and the other node with a function name.

### Claim 5 (depends on 1)

The computer-implemented method of claim 1, in responsive to a particular invocation, from among the invocations being processed, being a call-with-object invocation, said processing step comprises (i) creating a new node labeled with a particular object and the input and output object IDs, (ii) connecting the new node with another node, and (iii) labelling an edge between the new node and the other node with a function name.

### Claim 6 (depends on 1)

The computer-implemented method of claim 1, wherein the method further comprises:
using the directed graph to generate a prediction of an impending collision responsive to a new column being created for the directed graph by the invoked function that is directed to an obstacle-avoidance action; and
performing the obstacle-avoidance action on a physical mobile device implicated by the impending collision.

### Claim 7 (depends on 1)

The computer-implemented method of claim 1, wherein nodes in the directed graph have a type selected from the group consisting of (i) a get type, (ii) a set type, (iii) a call-with-objects type, and (iv) a call-with-columns type.

### Claim 8 (depends on 7)

The computer-implemented method of claim 7, wherein the get type gets a column object.

### Claim 9 (depends on 7)

The computer-implemented method of claim 7, wherein the set type adds a column object to a table object.

### Claim 10 (depends on 7)

The computer-implemented method of claim 7, wherein the call-with-objects type calls the invoked function with input objects and obtains output objects.

### Claim 11 (depends on 7)

The computer-implemented method of claim 7, wherein the call-with-columns type calls the invoked function with input columns and obtains new output columns type.

### Claim 12 (depends on 1)

The computer-implemented method of claim 1, further comprising feeding the directed graph into a program analysis pipeline to generate a prediction.

### Claim 13

A computer program product for creating and reducing a directed graph from an execution trace, the computer program product comprising a non-transitory computer readable storage medium having program instructions embodied therewith, the program instructions executable by a computer to cause the computer to perform a method comprising:
obtaining, by a hardware processor, the execution trace having a sequence of invocations of table transformations, each of the invocations being classified as a call-with-object invocation or as a call-with-column invocation, the call-with-object invocation indicating that an invoked function has object IDs as at least one of input and output information and the call-with-column invocation indicating that the invoked function has column names as the at least one of input and output information;
processing, by the hardware processor, the execution trace to build the directed graph, wherein each node in the directed graph indicates an object or a column and each edge indicates data flow; and
performing, by the hardware processor in response to all the invocations being processed, an automated graph reduction process on the directed graph by eliminating from the directed graph given nodes indicating a respective object such that that every two nodes that connect with each other, via a respective one of the given nodes as an intermediate node, can keep the connection therebetween.

### Claim 14 (depends on 13)

The computer program product of claim 13, wherein the column names had by the invoked function as the at least one of input and output information comprise column names taken by the invoked function as the input information and column names newly created by the invoked function as the output information.

### Claim 15 (depends on 14)

The computer program product of claim 14, wherein the column names newly created comprise new features, and the method further comprises applying the new features in a machine learning pipeline to generate a prediction.

### Claim 16 (depends on 13)

The computer program product of claim 13, in responsive to a particular invocation being processed being a call-with-column invocation, said processing step comprises (i) creating a new node labeled with a column and the input and output columns, (ii) connecting the new node with another node, and (iii) labelling an edge between the new node and the other node with a function name.

### Claim 17 (depends on 13)

The computer program product of claim 13, in responsive to a particular invocation, from among the invocations being processed, being a call-with-object invocation, said processing step comprises (i) creating a new node labeled with a particular object and the input and output object IDs, (ii) connecting the new node with another node, and (iii) labelling an edge between the new node and the other node with a function name.

### Claim 18 (depends on 13)

The computer program product of claim 13, wherein the method further comprises:
using the directed graph to generate a prediction of an impending collision responsive to a new column being created for the directed graph by the invoked function that is directed to an obstacle-avoidance action; and
performing the obstacle-avoidance action on a physical mobile device implicated by the impending collision.

### Claim 19 (depends on 13)

The computer program product of claim 13, wherein nodes in the directed graph have a type selected from the group consisting of (i) a get type, (ii) a set type, (iii) a call-with-objects type, and (iv) a call-with-columns type.

### Claim 20

A computer processing system for determining command-to-process correspondence, comprising:
a memory device including program code stored thereon;
a hardware processor, operatively coupled to the memory device, and configured to run the program code stored on the memory device to
obtain the execution trace having a sequence of invocations of table transformations, each of the invocations being classified as a call-with-object invocation or as a call-with-column invocation, the call-with-object invocation indicating that an invoked function has object IDs as at least one of input and output information and the call-with-column invocation indicating that the invoked function has column names as the at least one of input and output information;
process the execution trace to build the directed graph, wherein each node in the directed graph indicates an object or a column and each edge indicates data flow; and
perform, in response to all the invocations being processed, an automated graph reduction process on the directed graph by eliminating from the directed graph given nodes indicating a respective object such that that every two nodes that connect with each other, via a respective one of the given nodes as an intermediate node, can keep the connection therebetween.
