US 11,818,289 B2
Detecting scam callers using conversational agent and machine learning systems and methods
Ovidiu Serban, Sammamish, WA (US)
Assigned to T-Mobile USA, Inc., Bellevue, WA (US)
Filed by T-Mobile USA, Inc., Bellevue, WA (US)
Filed on Aug. 29, 2022, as Appl. No. 17/897,792.
Application 17/897,792 is a continuation of application No. 17/372,337, filed on Jul. 9, 2021, granted, now 11,463,582.
Prior Publication US 2023/0008822 A1, Jan. 12, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. H04M 3/42 (2006.01); H04M 3/493 (2006.01); G06N 20/00 (2019.01); H04M 3/436 (2006.01)
CPC H04M 3/493 (2013.01) [G06N 20/00 (2019.01); H04M 3/436 (2013.01); H04M 2203/6027 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method to detect a scam call using a conversational agent, the method comprising:
training a first model to detect scam-call indicia in call audio,
wherein the first model is trained using a training dataset that includes data associated with known scam calls;
training a second model to simulate speech of a user;
generating a conversational agent from the first model and the second model,
wherein the conversational agent is configured to, in a given phone call to a mobile device associated with the user, simulate speech of the user and detect the scam-call indicia in one or more responses to the simulated speech during the given phone call; and
in response to receiving a particular phone call, using the conversational agent to process the particular phone call.