Abstract
Telephone users frequently receive unwanted calls from sales, advertising, or spam callers. Such calls may be made by automated agents or bots of such sophistication that they are often difficult to distinguish from human callers.
This disclosure presents machine-learning techniques that enable differentiation of bots from human callers. Machine-learning models are trained to recognize artifacts that distinguish bot callers. Users can report callers as bots or humans, thereby enabling federated learning of differences between human callers and bots. A suspected bot caller is challenged with audio or visual captchas to further filter out bots. An incoming call that is confirmed as bot-initiated is either not delivered to the user, or the call recipient is alerted that the caller is likely a bot.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Kabeli, Ori; Bar, Nadav; and Schlesinger, Benny, "Distinguishing Bots from Human Callers", Technical Disclosure Commons, (August 30, 2018)
https://www.tdcommons.org/dpubs_series/1480