Bot-based data collection for detecting phone solicitations

Patent metadata

Field	Value
Publication number	US-10657463-B2
Application number	US-201916455593-A
Country	US
Kind code	B2
Filing date	Jun 27, 2019
Priority date	Feb 17, 2017
Publication date	May 19, 2020
Grant date	May 19, 2020

Abstract

One embodiment provides a method comprising answering one or more incoming phone calls received at one or more pre-specified phone numbers utilizing a bot. The bot is configured to engage in a conversation with a caller initiating an incoming phone call utilizing a voice recording that impersonates a human being. The method further comprises recording each conversation the bot engages in, and classifying each recorded conversation as one of poison data or truthful training data based on content of the recorded conversation and one or more learned detection models for detecting poisoned data.

First claim

What is claimed is:

1. A method comprising:
detecting an unsolicited phone call based on a real-time analysis of a conversation between a caller who initiated the unsolicited phone call and a bot, wherein the bot is configured to select a voice recording impersonating an individual belonging to a particular demographic targeted by the caller, and dynamically converse with the caller utilizing the selected voice recording; and
blocking the unsolicited phone call.

2. The method of claim 1 , wherein the bot is configured to dynamically converse with the caller based on a conversation template including one or more learned dialogue patterns.

3. The method of claim 1 , wherein the selected voice recording is selected from multiple voice recordings, and the bot is configured to dynamically converse with the caller utilizing the selected voice recording to increase a likelihood the caller perceives the bot as the individual belonging to the particular demographic targeted by the caller.

4. The method of claim 1 , wherein detecting an unsolicited phone call comprises:
analyzing the conversation in real-time by applying a learned detection model to the conversation to detect one or more attributes that indicate the unsolicited phone call is a phone scam or a telemarketing call.

5. The method of claim 4 , wherein applying a learned detection model comprises:
applying a classifier from a set of classifiers to the conversation.

6. The method of claim 5 , further comprising:
rotating which of the set of classifiers is applied to each conversation analyzed to reduce a likelihood of a phone scammer or telemarketer circumventing at least one classifier of the set of classifiers.

7. A system comprising:
at least one processor; and
a non-transitory processor-readable memory device storing instructions that when executed by the at least one processor causes the at least one processor to perform operations including:
detecting an unsolicited phone call based on a real-time analysis of a conversation between a caller who initiated the unsolicited phone call and a bot, wherein the bot is configured to select a voice recording impersonating an individual belonging to a particular demographic targeted by the caller, and dynamically converse with the caller utilizing the selected voice recording; and
blocking the unsolicited phone call.

8. The system of claim 7 , wherein the bot is configured to dynamically converse with the caller based on a conversation template including one or more learned dialogue patterns.

9. The system of claim 7 , wherein the selected voice recording is selected from multiple voice recordings, and the bot is configured to dynamically converse with the caller utilizing the selected voice recording to increase a likelihood the caller perceives the bot as the individual belonging to the particular demographic targeted by the caller.

10. The system of claim 7 , wherein detecting an unsolicited phone call comprises:
analyzing the conversation in real-time by applying a learned detection model to the conversation to detect one or more attributes that indicate the unsolicited phone call is a phone scam or a telemarketing call.

11. The system of claim 10 , wherein applying a learned detection model comprises:
applying a classifier from a set of classifiers to the conversation.

12. The system of claim 11 , wherein the operations further include:
rotating which of the set of classifiers is applied to each conversation analyzed to reduce a likelihood of a phone scammer or telemarketer circumventing at least one classifier of the set of classifiers.

13. A computer program product comprising a computer-readable hardware storage medium having program code embodied therewith, the program code being executable by a computer to implement a method comprising:
detecting an unsolicited phone call based on a real-time analysis of a conversation between a caller who initiated the unsolicited phone call and a bot, wherein the bot is configured to select a voice recording impersonating an individual belonging to a particular demographic targeted by the caller, and dynamically converse with the caller utilizing the selected voice recording; and
blocking the unsolicited phone call.

14. The computer program product of claim 13 , wherein the bot is configured to dynamically converse with the caller based on a conversation template including one or more learned dialogue patterns.

15. The computer program product of claim 13 , wherein the selected voice recording is selected from multiple voice recordings, and the bot is configured to dynamically converse with the caller utilizing the selected voice recording to increase a likelihood the caller perceives the bot as the individual belonging to the particular demographic targeted by the caller.

16. The computer program product of claim 13 , wherein detecting an unsolicited phone call comprises:
analyzing the conversation in real-time by applying a learned detection model to the conversation to detect one or more attributes that indicate the unsolicited phone call is a phone scam or a telemarketing call.

17. The computer program product of claim 16 , wherein applying a learned detection model comprises:
applying a classifier from a set of classifiers to the conversation.

18. The computer program product of claim 17 , wherein the method further comprises:
rotating which of the set of classifiers is applied to each conversation analyzed to reduce a likelihood of a phone scammer or telemarketer circumventing at least one classifier of the set of classifiers.

Assignees

• IBM

Inventors

• Baracaldo Angel Nathalie
• Chowdhary Pawan R
• Ludwig Heiko H
• Moore Robert J
• Nakamura Taiga

Classifications

• G10L15/1822

  Parsing for meaning understanding · CPC title

• H04M2201/41

  using speaker recognition · CPC title

• G10L17/06

  Decision making techniques; Pattern matching strategies · CPC title

• G10L25/48

  specially adapted for particular use · CPC title

• H04M2203/6027

  Fraud preventions · CPC title