Natural language processing to merge related alert messages for accessibility

• US 10,380,994 B2
• Filed: 10/26/2017
• Issued: 08/13/2019
• Est. Priority Date: 07/08/2017
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Boxwell, Stephen A, Brake, Kyle M, Frost, Keith G, Vernier, Stanley J
• Primary Examiner(s): Akki, Munear T
• Application Number:
  US15/795,205
• Publication Number:
  US 20190013007A1
• Time in Patent Office: 656 Days
• CPC Class Codes: G06F 40/211 Syntactic parsing, e.g. based on context-free grammar [CFG] or unification grammars
• CPC Class Codes: G06Q 30/0601 Electronic shopping [e-shopping]
• CPC Class Codes: G10L 13/07 Concatenation rules

Abstract

A method for merging incoming alerts for accessibility is described. Two input alerts intended for presentation by a screen reader are received. If the two input alerts have arrived with a specified time interval, the two input alerts are combined into an output alert. The output alert is sent to a screen reader for presentation.

Claims

1. A method in a data processing system comprising a processor and a memory, the memory comprising instructions executed by the processor for merging incoming alerts for accessibility, the method comprising:

receiving two input alerts intended for presentation by a screen reader;

determining, by the processor, whether the two input alerts arrived within a specified time interval;

responsive to determining that the two input alerts arrived within the specified time interval, determining whether the two input alerts are similar; and

responsive to determining that the two input alerts are not similar, combining the two input alerts in a concatenation operation to produce an output alert, wherein the output alert is a concatenated alert; and

sending the output alert to the screen reader for presentation.

2. The method as recited in claim 1, wherein the similarity of the two input alerts is determined by comparing dependency trees of the two input alerts.

3. The method as recited in claim 2, wherein a similarity rule used for determining whether two dependency trees are similar, each of the dependency trees for a respective one of two input alerts if the root nodes of the two dependency trees are similar and a predetermined number of corresponding sub-nodes of the two dependency trees are similar, then the two dependency trees are similar.

4. The method as recited in claim 2, wherein an equivalency rule is used for determining whether dependency trees of the two input alerts are equivalent, wherein the equivalency rule is that a root node of each tree is the same or equivalent and a specified number or percentage of the sub-nodes are the same or equivalent.

5. The method as recited in claim 1, further comprising resetting a timer after a second input alert of the two input alerts is received to determine whether a third input alert arrived within the specified time interval of the second input alert.

6. The method as recited in claim 1, wherein the specified time interval is a function of a time interval required to present a first input alert.

7. The method as recited in claim 1, wherein a similarity rule is used to determine whether corresponding elements from respective input alerts are similar.

8. The method as recited in claim 1, wherein an equivalency rule is used for determining whether dependency trees of the two input alerts are equivalent.

9. The method as recited in claim 8, wherein the equivalency rule is that a root node of each dependency tree is the same or equivalent and a specified number or percentage of a set of sub-nodes of each dependency tree are the same or equivalent with corresponding sub-nodes of the other dependency tree.

10. The method as recited in claim 2, wherein the specified time interval is a function of a time interval required to present a first input alert.

11. The method as recited in claim 2, wherein a similarity rule is used to determine whether corresponding elements from respective input alerts are similar.