公开(公告)号：US10381009B2

公开(公告)日：2019-08-13

申请号：US15818231

申请日：2017-11-20

Applicant: PINDROP SECURITY, INC.

Inventor： Elie Khoury ,  Matthew Garland

IPC: G06N3/04 ,  G06N3/08 ,  G10L15/16 ,  G10L17/02 ,  G10L17/04 ,  G10L17/08 ,  G10L17/18 ,  G10L17/22

Abstract: The present invention is directed to a deep neural network (DNN) having a triplet network architecture, which is suitable to perform speaker recognition. In particular, the DNN includes three feed-forward neural networks, which are trained according to a batch process utilizing a cohort set of negative training samples. After each batch of training samples is processed, the DNN may be trained according to a loss function, e.g., utilizing a cosine measure of similarity between respective samples, along with positive and negative margins, to provide a robust representation of voiceprints.

公开(公告)号：US10375248B2

公开(公告)日：2019-08-06

申请号：US16026498

申请日：2018-07-03

Applicant: PINDROP SECURITY, INC.

Inventor： Payas Gupta ,  David Dewey

IPC: H04M3/42 ,  H04M3/54 ,  H04W4/50 ,  H04L29/06 ,  H04L29/08 ,  H04M3/436 ,  H04W12/06

Abstract: The invention may verify calls to a telephone device by activating call forwarding to redirect calls for the telephone device to a prescribed destination; receiving a message from a server verifying the call; deactivating call forwarding to receive the call; and reactivating call forwarding when the call is concluded. In another embodiment, the invention may, in response to a telephone device initiating a call to a second telephone device installed with a particular application or software, transmit a message to a server causing it to instruct the second telephone device to deactivate call forwarding. In yet another embodiment, the invention may cause a server to receive a message from a prescribed location indicating that a call was received via call forwarding, and in response to the message, transmit an instruction to the intended recipient to deactivate the call forwarding if the call is verified as legitimate.

公开(公告)号：US10325601B2

公开(公告)日：2019-06-18

申请号：US15709290

申请日：2017-09-19

Applicant: PINDROP SECURITY, INC.

Inventor： Elie Khoury ,  Matthew Garland

IPC: G10L17/00 ,  G06N7/00 ,  G10L15/07 ,  G10L15/26 ,  G10L17/04 ,  H04M1/27 ,  G10L17/24 ,  G10L15/19 ,  G10L17/08

Abstract: Utterances of at least two speakers in a speech signal may be distinguished and the associated speaker identified by use of diarization together with automatic speech recognition of identifying words and phrases commonly in the speech signal. The diarization process clusters turns of the conversation while recognized special form phrases and entity names identify the speakers. A trained probabilistic model deduces which entity name(s) correspond to the clusters.

公开(公告)号：US20190141184A1

公开(公告)日：2019-05-09

申请号：US16200379

申请日：2018-11-26

Applicant: PINDROP SECURITY, INC.

Inventor： Lance DOUGLAS

IPC: H04M3/22 ,  H04M3/436

Abstract: Aspects of the invention determining a threat score of a call traversing a telecommunications network by leveraging the signaling used to originate, propagate and terminate the call. Outer-edge data utilized to originate the call may be analyzed against historical, or third party real-time data to determine the propensity of calls originating from those facilities to be categorized as a threat. Storing the outer edge data before the call is sent over the communications network permits such data to be preserved and not subjected to manipulations during traversal of the communications network. This allows identification of threat attempts based on the outer edge data from origination facilities, thereby allowing isolation of a compromised network facility that may or may not be known to be compromised by its respective network owner. Other aspects utilize inner edge data from an intermediate node of the communications network which may be analyzed against other inner edge data from other intermediate nodes and/or outer edge data.

公开(公告)号：US10257591B2

公开(公告)日：2019-04-09

申请号：US15600625

申请日：2017-05-19

Applicant: PINDROP SECURITY, INC.

Inventor： Nick Gaubitch ,  Scott Strong ,  John Cornwell ,  Hassan Kingravi ,  David Dewey

IPC: H04M1/56 ,  H04M15/06 ,  H04Q3/70 ,  G10L25/51 ,  H04L25/02 ,  H04M3/493 ,  H04M3/22 ,  H04M7/12 ,  H04Q1/45

Abstract: Systems, methods, and computer-readable media for call classification and for training a model for call classification, an example method comprising: receiving DTMF information from a plurality of calls; determining, for each of the calls, a feature vector including statistics based on DTMF information such as DTMF residual signal comprising channel noise and additive noise; training a model for classification; comparing a new call feature vector to the model; predicting a device type and geographic location based on the comparison of the new call feature vector to the model; classifying the call as spoofed or genuine; and authenticating a call or altering an IVR call flow.

公开(公告)号：US20180254046A1

公开(公告)日：2018-09-06

申请号：US15910387

申请日：2018-03-02

Applicant: PINDROP SECURITY, INC.

Inventor： Elie KHOURY ,  Parav NAGARSHETH ,  Kailash PATIL ,  Matthew GARLAND

IPC: G10L17/02 ,  G10L17/04 ,  G10L17/06 ,  G10L17/18 ,  G10L19/02 ,  G10L25/24

Abstract: An automated speaker verification (ASV) system incorporates a first deep neural network to extract deep acoustic features, such as deep CQCC features, from a received voice sample. The deep acoustic features are processed by a second deep neural network that classifies the deep acoustic features according to a determined likelihood of including a spoofing condition. A binary classifier then classifies the voice sample as being genuine or spoofed.

公开(公告)号：US20170372725A1

公开(公告)日：2017-12-28

申请号：US15610378

申请日：2017-05-31

Applicant: PINDROP SECURITY, INC.

Inventor： Elie KHOURY ,  Matthew GARLAND

IPC: G10L25/45 ,  G10L25/78 ,  G10L25/27

CPC classification number: G10L25/45 ,  G10L25/27 ,  G10L25/51 ,  G10L25/78

Abstract: Methods, systems, and apparatuses for audio event detection, where the determination of a type of sound data is made at the cluster level rather than at the frame level. The techniques provided are thus more robust to the local behavior of features of an audio signal or audio recording. The audio event detection is performed by using Gaussian mixture models (GMMs) to classify each cluster or by extracting an i-vector from each cluster. Each cluster may be classified based on an i-vector classification using a support vector machine or probabilistic linear discriminant analysis. The audio event detection significantly reduces potential smoothing error and avoids any dependency on accurate window-size tuning. Segmentation may be performed using a generalized likelihood ratio and a Bayesian information criterion, and the segments may be clustered using hierarchical agglomerative clustering. Audio frames may be clustered using K-means and GMMs.

公开(公告)号：US09824692B1

公开(公告)日：2017-11-21

申请号：US15262748

申请日：2016-09-12

Applicant: PINDROP SECURITY, INC.

Inventor： Elie Khoury ,  Matthew Garland

IPC: G10L17/18 ,  G10L15/16 ,  G10L17/04 ,  G10L17/08 ,  G10L17/02 ,  G10L17/22 ,  G06N3/04 ,  G06N3/08

CPC classification number: G10L17/08 ,  G06N3/04 ,  G06N3/08 ,  G10L15/16 ,  G10L17/02 ,  G10L17/04 ,  G10L17/18 ,  G10L17/22

Abstract: The present invention is directed to a deep neural network (DNN) having a triplet network architecture, which is suitable to perform speaker recognition. In particular, the DNN includes three feed-forward neural networks, which are trained according to a batch process utilizing a cohort set of negative training samples. After each batch of training samples is processed, the DNN may be trained according to a loss function, e.g., utilizing a cosine measure of similarity between respective samples, along with positive and negative margins, to provide a robust representation of voiceprints.

公开(公告)号：US12301753B2

公开(公告)日：2025-05-13

申请号：US17378567

申请日：2021-07-16

Applicant: Pindrop Security, Inc.

Inventor： MohammedAli Merchant

IPC: H04M3/00 ,  H04L9/32 ,  H04L12/66 ,  H04M3/42 ,  H04M5/00

Abstract: Embodiments described herein provide for authenticating callers to call centers using browser fingerprinting. A call center system or third-party analytics system includes a server that transmits notifications to a caller device that includes an interactive hyperlink or URL to a particular webpage. When a browser of the caller device navigates to the webpage, the server captures certain types of information about the caller device and generates a browser fingerprint for the caller device. The browser fingerprint is compared against a database of registered browser fingerprints to verify that the caller device of the current call is the registered, expected caller device. The server transmits the notification via any number of communication channels and protocols, such as text messages (e.g., SMS messages, MMS messages), emails, and push notifications associated with client-side software, among others.

公开(公告)号：US20250124945A1

公开(公告)日：2025-04-17

申请号：US18989690

申请日：2024-12-20

Applicant: PINDROP SECURITY, INC.

Inventor： Hrishikesh RAO ,  Kedar PHATAK ,  Elie KHOURY

IPC: G10L25/60 ,  G06N20/20 ,  G10L15/02

Abstract: Embodiments described herein provide for a machine-learning architecture for modeling quality measures for enrollment signals. Modeling these enrollment signals enables the machine-learning architecture to identify deviations from expected or ideal enrollment signal in future test phase calls. These differences can be used to generate quality measures for the various audio descriptors or characteristics of audio signals. The quality measures can then be fused at the score-level with the speaker recognition's embedding comparisons for verifying the speaker. Fusing the quality measures with the similarity scoring essentially calibrates the speaker recognition's outputs based on the realities of what is actually expected for the enrolled caller and what was actually observed for the current inbound caller.