公开(公告)号：US11983625B2

公开(公告)日：2024-05-14

申请号：US16911100

申请日：2020-06-24

Applicant: Intel Corporation

Inventor： Nilesh Ahuja ,  Ignacio J. Alvarez ,  Ranganath Krishnan ,  Ibrahima J. Ndiour ,  Mahesh Subedar ,  Omesh Tickoo

IPC: G06N3/08 ,  G05B13/02 ,  G06F18/2431 ,  G06F18/25 ,  G06N5/046 ,  G06N7/01

CPC classification number: G06N3/08 ,  G05B13/026 ,  G05B13/027 ,  G06F18/2431 ,  G06F18/251 ,  G06N5/046 ,  G06N7/01

Abstract: Techniques are disclosed for using neural network architectures to estimate predictive uncertainty measures, which quantify how much trust should be placed in the deep neural network (DNN) results. The techniques include measuring reliable uncertainty scores for a neural network, which are widely used in perception and decision-making tasks in automated driving. The uncertainty measurements are made with respect to both model uncertainty and data uncertainty, and may implement Bayesian neural networks or other types of neural networks.

公开(公告)号：US20230298322A1

公开(公告)日：2023-09-21

申请号：US18325436

申请日：2023-05-30

Applicant: Intel Corporation

Inventor： Ibrahima Ndiour ,  Nilesh Ahuja ,  Ranganath Krishnan ,  Mahesh Subedar ,  Omesh Tickoo ,  Ergin Genc

IPC: G06V10/77 ,  G06V10/82 ,  G06V10/80

CPC classification number: G06V10/7715 ,  G06V10/82 ,  G06V10/80

Abstract: Features extracted from one or more layers of a trained deep neural network (DNN) are used to detect out-of-distribution (OOD) data, such as anomalies. An OOD detection process includes transforming a feature output from a layer of the DNN from a relatively high-dimensional feature space to a lower-dimensional space, and then performing a reverse transformation back to the higher-dimensional feature space, resulting in a reconstructed feature. A feature reconstruction error is calculated based on a difference between the reconstructed feature and the original feature output from the DNN. The OOD detection process may further include calculating a score based on the feature reconstruction error and generating a visual representation of the feature reconstruction error.

公开(公告)号：US11375352B2

公开(公告)日：2022-06-28

申请号：US16828986

申请日：2020-03-25

Applicant: INTEL CORPORATION

Inventor： Richard Dorrance ,  Ignacio Alvarez ,  Deepak Dasalukunte ,  S M Iftekharul Alam ,  Sridhar Sharma ,  Kathiravetpillai Sivanesan ,  David Israel Gonzalez Aguirre ,  Ranganath Krishnan ,  Satish Jha

IPC: H04W4/46 ,  G08G1/01 ,  H04W72/04

Abstract: Vehicle navigation control systems in autonomous driving rely on the accuracy of maps which include features about a vehicle's environment so that a vehicle may safely navigate through its surrounding area. Accordingly, this disclosure provides methods and devices which implement mechanisms for communicating features observed about a vehicle's environment for use in updating maps so as to provide vehicles with accurate and “real-time” features of its surroundings while taking network resources, such as available frequency-time resources, into consideration.

公开(公告)号：US11314258B2

公开(公告)日：2022-04-26

申请号：US16727955

申请日：2019-12-27

Applicant: Intel Corporation

Inventor： David Gomez Gutierrez ,  Ranganath Krishnan ,  Javier Felip Leon ,  Nilesh Ahuja ,  Ibrahima Ndiour

IPC: G05D1/02 ,  B60W50/02 ,  B60W30/095 ,  B60W50/00 ,  B60W40/02

Abstract: A safety system for a vehicle may include one or more processors configured to determine uncertainty data indicating uncertainty in one or more predictions from a driving model during operation of a vehicle; change or update one or more of the driving model parameters to one or more changed or updated driving model parameters based on the determined uncertainty data; and provide the one or more changed or updated driving model parameters to a control system of the vehicle for controlling the vehicle to operate in accordance with the driving model including the one or more changed or updated driving model parameters.

公开(公告)号：US20210117792A1

公开(公告)日：2021-04-22

申请号：US17132858

申请日：2020-12-23

Applicant: Intel Corporation

Inventor： Nilesh Ahuja ,  Mahesh Subedar ,  Ranganath Krishnan ,  Ibrahima Ndiour ,  Omesh Tickoo

IPC: G06N3/08 ,  G06N3/04 ,  G06F7/58

Abstract: Methods, apparatus, systems and articles of manufacture are disclosed to facilitate continuous learning. An example apparatus includes a trainer to train a first Bayesian neural network (BNN) and a second BNN, the first BNN associated with a first weight distribution and the second BNN associated with a second weight distribution. The example apparatus includes a weight determiner to determine a first sampling weight associated with the first BNN and a second sampling weight associated with the second BNN. The example apparatus includes a network sampler to sample at least one of the first weight distribution or the second weight distribution based on a pseudo-random number, the first sampling weight, and the second sampling weight. The example apparatus includes an inference controller to generate an ensemble weight distribution based on the sample.

公开(公告)号：US20210117760A1

公开(公告)日：2021-04-22

申请号：US17133072

申请日：2020-12-23

Applicant: Intel Corporation

Inventor： Ranganath Krishnan ,  Omesh Tickoo ,  Nilesh Ahuja ,  Ibrahima Ndiour ,  Mahesh Subedar

IPC: G06N3/04 ,  G06N3/08 ,  G06K9/62

Abstract: Methods, systems, and apparatus to obtain well-calibrated uncertainty in probabilistic deep neural networks are disclosed. An example apparatus includes a loss function determiner to determine a differentiable accuracy versus uncertainty loss function for a machine learning model, a training controller to train the machine learning model, the training including performing an uncertainty calibration of the machine learning model using the loss function, and a post-hoc calibrator to optimize the loss function using temperature scaling to improve the uncertainty calibration of the trained machine learning model under distributional shift.

公开(公告)号：US10217292B2

公开(公告)日：2019-02-26

申请号：US15089173

申请日：2016-04-01

Applicant: Intel Corporation

Inventor： Ignacio J. Alvarez ,  Ranganath Krishnan

IPC: G06T19/20 ,  G06T17/20 ,  H04L29/06

Abstract: According to various embodiments, devices, methods, and computer-readable media for reconstructing a 3D scene are described. A server device, sensor devices, and client devices may interoperate to reconstruct a 3D scene sensed by the sensor devices. The server device may generate one or more models for objects in the scene, including the identification of dynamic and/or static objects. The sensor devices may, provide model data updates based on these generated models, such that only delta changes in the scene may be provided, in addition to raw sensor data. Models may utilize semantic knowledge, such as knowledge of the venue or identity of one or more persons in the scene, to further facilitate model generation and updating. Other embodiments may be described and/or claimed.

公开(公告)号：US20170148224A1

公开(公告)日：2017-05-25

申请号：US15089173

申请日：2016-04-01

Applicant: Intel Corporation

Inventor： Ignacio J. Alvarez ,  Ranganath Krishnan

IPC: G06T19/20 ,  H04L29/06 ,  G06T17/20

CPC classification number: G06T19/20 ,  G06T17/00 ,  G06T17/20 ,  G06T2200/16 ,  G06T2210/08 ,  G06T2219/024 ,  H04L67/42

Abstract: According to various embodiments, devices, methods, and computer-readable media for reconstructing a 3D scene are described. A server device, sensor devices, and client devices may interoperate to reconstruct a 3D scene sensed by the sensor devices. The server device may generate one or more models for objects in the scene, including the identification of dynamic and/or static objects. The sensor devices may, provide model data updates based on these generated models, such that only delta changes in the scene may be provided, in addition to raw sensor data. Models may utilize semantic knowledge, such as knowledge of the venue or identity of one or more persons in the scene, to further facilitate model generation and updating. Other embodiments may be described and/or claimed.