公开(公告)号：US12135951B2

公开(公告)日：2024-11-05

申请号：US17582464

申请日：2022-01-24

Applicant: NEC Laboratories America, Inc.

Inventor： Xuchao Zhang ,  Bo Zong ,  Haifeng Chen ,  Yanchi Liu

IPC: G06F40/58 ,  G06F40/284 ,  G06F40/30

Abstract: Systems and methods are provided for Cross-lingual Transfer Interpretation (CTI). The method includes receiving text corpus data including premise-hypothesis pairs with a relationship label in a source language, and conducting a source to target language translation. The method further includes performing a feature importance extraction, where an integrated gradient is applied to assign an importance score to each input feature, and performing a cross-lingual feature alignment, where tokens in the source language are aligned with tokens in the target language for both the premise and the hypothesis based on semantic similarity. The method further includes performing a qualitative analysis, where the importance score of each token can be compared between the source language and the target language according to a feature alignment result.

公开(公告)号：US20240355102A1

公开(公告)日：2024-10-24

申请号：US18609097

申请日：2024-03-19

Applicant: NEC Laboratories America, Inc.

Inventor： Sparsh Garg ,  Samuel Schulter

IPC: G06V10/774 ,  G06V10/26 ,  G06V10/82 ,  G06V20/54 ,  G06V20/62

CPC classification number: G06V10/7753 ,  G06V10/26 ,  G06V10/82 ,  G06V20/54 ,  G06V20/625 ,  G06V2201/08

Abstract: Systems and methods for traffic violation prediction. The systems and methods include obtaining a plurality of bounding boxes of road scene categories from an input dataset by employing a pre-trained detection model. A plurality of pseudo-labels of road scene categories for the plurality of bounding boxes can be obtained by employing the pre-trained detection model. A labeled dataset can be obtained by filtering the input dataset for images having the plurality of pseudo-labels and the plurality of bounding boxes. A traffic violation prediction model can be trained with both unlabeled and labeled dataset including the road scene categories obtained from the pre-trained detection model to predict simultaneous traffic violations of one or more riders in a road scene.

公开(公告)号：US20240355090A1

公开(公告)日：2024-10-24

申请号：US18639534

申请日：2024-04-18

Applicant: NEC Laboratories America, Inc.

Inventor： Vijay Kumar Baikampady Gopalkrishna ,  Samuel Schulter ,  Xiang Yu ,  Manmohan Chandraker

IPC: G06V10/74 ,  G06F16/532 ,  G06F40/186 ,  G06V10/774 ,  G06V20/60 ,  G06V20/70

CPC classification number: G06V10/761 ,  G06F16/532 ,  G06V10/774 ,  G06V20/60 ,  G06V20/70 ,  G06F40/186

Abstract: Systems and methods are provided for matching one or more images using conditional similarity pseudo-labels, including analyzing an unlabeled dataset of images, accessing a foundational vision-language model trained on a plurality of image-text pairs, and defining a set of attributes each comprising multiple possible values for generating pseudo-labels based on notions of similarity (NoS). Text prompts are generated for each attribute value using a prompt template and encoding the text prompts using a text encoder of the foundational model. Each image in the dataset of images is processed through a vision encoder of the foundational model to obtain visual features, the visual features are compared against encoded text prompts to assign a pseudo-label for each attribute for each image, and a conditional similarity network (CSN) is trained with the pseudo-labeled images to generate a conditional similarity model.

公开(公告)号：US20240354953A1

公开(公告)日：2024-10-24

申请号：US18616983

申请日：2024-03-26

Applicant: NEC Laboratories America, Inc.

Inventor： Eric Cosatto

IPC: G06T7/00 ,  G01N1/30

CPC classification number: G06T7/0014 ,  G01N1/30 ,  G06T2207/20081 ,  G06T2207/20084

Abstract: Methods and systems for training a model include performing color deconvolution on a set of training images, stained according to a first staining process, to generate channels that correspond to dyes used in the first staining process and dyes used in a second staining process. A channel is selected corresponds to a dye used in the second staining process. A machine learning model is trained, using the selected channel of the set of training images, to function with images stained according to the first staining process and images stained according to the second staining process.

公开(公告)号：US20240354642A1

公开(公告)日：2024-10-24

申请号：US18616930

申请日：2024-03-26

Applicant: NEC Laboratories America, Inc.

Inventor： Vijay Kumar Baikampady Gopalkrishna ,  Manmohan Chandraker

IPC: G06N20/00

CPC classification number: G06N20/00

Abstract: Methods and systems for fine-tuning a model include generating a label space for a target domain. Text pseudo-labels are generated for images in an unlabeled dataset from the target domain based on the label space using a pre-trained vision language model. The pre-trained vision language model is fine-tuned for the target domain using the images with the text pseudo-labels.

公开(公告)号：US20240304329A1

公开(公告)日：2024-09-12

申请号：US18591838

申请日：2024-02-29

Applicant: NEC Laboratories America, Inc.

Inventor： Wei Cheng ,  Haifeng Chen ,  Xujiang Zhao ,  Xianjun Yang

IPC: G16H50/20 ,  G06F40/40 ,  G16H20/00

CPC classification number: G16H50/20 ,  G06F40/40 ,  G16H20/00

Abstract: Methods and systems for prompt tuning include training a tuning function to set prompt position, prompt length, or prompt pool based on a language processing task. The tuning function is applied to an input query to generate a combined input, with prompt text having the prompt length, being selected according to the prompt pool, and being added to the input query at the prompt position. The combined input is applied to a language model.

公开(公告)号：US20240249614A1

公开(公告)日：2024-07-25

申请号：US18417769

申请日：2024-01-19

Applicant: NEC Laboratories America, Inc.

Inventor： Shaobo HAN ,  Yuheng CHEN ,  Ming-Fang HUANG ,  Tingfeng LI ,  Ting WANG

IPC: G08G1/01

CPC classification number: G08G1/0116 ,  G08G1/0112

Abstract: Disclosed are vehicle-infrastructure interaction systems and methods employing a distributed fiber optic sensing (DFOS) system operating with pre-deployed fiber-optic telecommunication cables buried alongside/proximate to highways/roadways which provide 24/7 continuous information stream of vehicle traffic at multiple sites; only require a single optical sensor cable that senses/monitors multiple locations of interest and multiple lanes of traffic; the single optical sensor cable measures multiple related information (multi-parameters) about a vehicle, including driving speed, wheelbase, number of axles, tire pressure, and others, that can be used to derive secondary information such as weight-in-motion; and overall information about a fleet of vehicles, such as traffic congestion or traffic-cargo volume. Different from merely traffic counts, our approach can provide the count grouped by vehicle-types and cargo weights. Precise measurements are facilitated by high temporal sampling rates of the distributed acoustic sensing and a dedicated peak finding algorithm for extracting the timing information reliably.

公开(公告)号：US20240248228A1

公开(公告)日：2024-07-25

申请号：US18417812

申请日：2024-01-19

Applicant: NEC Laboratories America, Inc.

Inventor： Ming-Fang HUANG ,  Milad SALEMI

IPC: G01V1/30 ,  G01V1/22 ,  G01V1/28

CPC classification number: G01V1/307 ,  G01V1/226 ,  G01V1/288 ,  G01V2210/43 ,  G01V2210/65 ,  G01V2210/72

Abstract: Disclosed are systems and methods that estimate machine distance to an optical fiber cable from sensing data collected using distributed fiber optic sensing (DFOS). Specialized hardware, DFOS that uses optical sensor fiber as a continuous spatial sensor along with a real time Artificial Intelligence (AI) processing unit, that detects threats within a proximity of buried fiber optic cable and a determines a moving direction of the threats such that it can effectively mitigate and contain the threats before damage to the buried fiber optic cable occurs. Advantageously, the system according to the present disclosure does not require any prior location knowledge or surveying prior to performing its monitoring.

公开(公告)号：US20240247974A1

公开(公告)日：2024-07-25

申请号：US18417798

申请日：2024-01-19

Applicant: NEC Laboratories America, Inc.

Inventor： Junqiang HU ,  Yue-Kai HUANG

IPC: G01H9/00 ,  G01D5/353

CPC classification number: G01H9/004 ,  G01D5/35361

Abstract: Disclosed is a forward phase method using regular narrow line width CW laser, to cover the acoustic band with reduced processing speed, while tolerant laser frequency drift. A narrow linewidth CW laser is used to launch its power into an optical fiber at a transmitter side. At a receiver side, another narrow linewidth CW laser is used to coherently detect the received signal. The detected signal, which includes both X and Y polarizations, each having in-phase and quadrature to represent a “complex” channel, are connected to an ADC's inputs. Signal processing following the ADC inputs and extracts the phase change of the acoustic band.

公开(公告)号：US20240212865A1

公开(公告)日：2024-06-27

申请号：US18539506

申请日：2023-12-14

Applicant: NEC Laboratories America, Inc.

Inventor： Wenchao Yu ,  Wei Cheng ,  Haifeng Chen

IPC: G16H50/70 ,  G16H10/60 ,  G16H20/00 ,  G16H50/20

CPC classification number: G16H50/70 ,  G16H10/60 ,  G16H20/00 ,  G16H50/20

Abstract: Methods and systems for training a healthcare treatment machine learning model include segmenting a patient trajectory, which includes a sequence of patient states and treatment actions. A machine learning model is trained based on segments of the patient trajectory, including a prototype layer that learns prototype vectors representing respective classes of trajectory segments and an imitation learning layer that learns a policy to select a treatment action based on an input state and a skill embedding.