公开(公告)号：WO2022216375A1

公开(公告)日：2022-10-13

申请号：PCT/US2022/017429

申请日：2022-02-23

Applicant: NEC LABORATORIES AMERICA, INC.

Inventor： NATSUMEDA, Masanao ,  CHENG, Wei ,  MIZOGUCHI, Takehiko ,  CHEN, Haifeng

IPC: G06N3/04 ,  G06N3/08

Abstract: Methods and systems for training a neural network include training models for respective sensor groups in a cyber-physical system. Combinations of sensor groups and operational modes are sampled. A combination model is trained for each of the sampled combinations. A best combination model is determined based on performance measured during training. The best combination model is fine-tuned.

公开(公告)号：WO2022076402A1

公开(公告)日：2022-04-14

申请号：PCT/US2021/053558

申请日：2021-10-05

Applicant: NEC LABORATORIES AMERICA, INC

Inventor： TANG, Luan ,  CHENG, Wei ,  CHEN, Haifeng ,  CHEN, Zhengzhang ,  REN, Yuxiang

IPC: B60W50/02 ,  B60W30/08 ,  B60W10/20 ,  B60W40/107 ,  B60W10/18 ,  H04L65/40

Abstract: A method for vehicle fault detection is provided. The method includes training (810), by a cloud module controlled by a processor device, an entity-shared modular and a shared modular connection controller. The entity- shared modular stores common knowledge for a transfer scope, and is formed from a set of sub- networks which are dynamically assembled for different target entities of a vehicle by the shared modular connection controller. The method further includes training (820), by an edge module controlled by another processor device, an entity-specific decoder and an entity-specific connection controller. The entity-specific decoder is for filtering entity- specific information from the common knowledge in the entity- shared modular by dynamically assembling the set of sub-networks in a manner decided by the entity specific connection controller.

公开(公告)号：WO2021242585A1

公开(公告)日：2021-12-02

申请号：PCT/US2021/033107

申请日：2021-05-19

Applicant: NEC LABORATORIES AMERICA, INC.

Inventor： YU, Wenchao ,  CHEN, Haifeng ,  CHENG, Wei

IPC: G06N20/00

Abstract: A method for learning prototypical options for interpretable imitation learning is presented. The method includes initializing (701) options by bottleneck state discovery, each of the options presented by an instance of trajectories generated by experts, applying (703) segmentation embedding learning to extract features to represent current states in segmentations by dividing the trajectories into a set of segmentations, learning (705) prototypical options for each segment of the set of segmentations to mimic expert policies by minimizing loss of a policy and projecting prototypes to the current states, training (707) option policy with imitation learning techniques to learn a conditional policy, generating (709) interpretable policies by comparing the current states in the segmentations to one or more prototypical option embeddings, and taking (711) an action based on the interpretable policies generated.

公开(公告)号：WO2021162839A1

公开(公告)日：2021-08-19

申请号：PCT/US2021/014339

申请日：2021-01-21

Applicant: NEC LABORATORIES AMERICA, INC.

Inventor： CHENG, Wei ,  CHEN, Haifeng ,  YU, Wenchao

IPC: H04N21/466 ,  G06N3/08 ,  G06N3/04 ,  G06F16/9537 ,  G06F16/9535

Abstract: A method for employing a graph enhanced attention network for explainable point-of-interest (POI) recommendation (GEAPR) is presented. The method includes interpreting (801) POI prediction in an end-to-end fashion by adopting an adaptive neural network, learning (803) user representations by aggregating a plurality of factors, the plurality of factors including structural context, neighbor impact, user attributes, and geolocation influence, and quantifying (805) each of the plurality of factors by numeric values as feature salience indicators.

公开(公告)号：WO2021154838A1

公开(公告)日：2021-08-05

申请号：PCT/US2021/015262

申请日：2021-01-27

Applicant: NEC LABORATORIES AMERICA, INC.

Inventor： CHENG, Wei ,  CHEN, Haifeng ,  NI, Jingchao ,  XU, Dongkuan

IPC: G06N3/08 ,  G06N3/04 ,  G06N3/063

Abstract: Methods and systems for training a neural network model include training (404) a modular neural network model (100), which has a shared encoder and one or more task-specific decoders, including training one or more policy networks that control connections between the shared encoder and the one or more task-specific decoders in accordance with multiple tasks. A multitask neural network model (120) is trained (404) for the multiple tasks, with an output of the modular neural network model and the multitask neural network model being combined to form a final output.

公开(公告)号：WO2017139147A1

公开(公告)日：2017-08-17

申请号：PCT/US2017/015969

申请日：2017-02-01

Applicant: NEC LABORATORIES AMERICA, INC.

Inventor： CHENG, Wei ,  ZHANG, Kai ,  CHEN, Haifeng ,  JIANG, Guofei

IPC: H04L12/24

CPC classification number: G06F11/079 ,  G06F11/0709 ,  G06F11/0751 ,  G06F2201/805 ,  G06F2201/81 ,  G06N99/005

Abstract: A method is provided for root cause anomaly detection in an invariant network having a plurality of nodes that generate time series data. The method includes modeling anomaly propagation in the network. The method includes reconstructing broken invariant links in an invariant graph based on causal anomaly ranking vectors. Each broken invariant link involves a respective node pair formed from the plurality of nodes such that one of the nodes in the respective node pair has an anomaly. Each causal anomaly ranking vector is for indicating a respective node anomaly status for a given one of the plurality of nodes when paired. The method includes calculating a sparse penalty of the casual anomaly ranking vectors to obtain a set of time-dependent anomaly rankings. The method includes performing temporal smoothing of the set of rankings, and controlling an anomaly-initiating one of the plurality of nodes based on the set of rankings.

Abstract translation: 提供了一种用于在具有生成时间序列数据的多个节点的不变网络中进行根本原因异常检测的方法。 该方法包括对网络中的异常传播进行建模。 该方法包括基于因果异常排序向量重建不变图中的断裂不变链接。 每个断开的不变链路涉及由多个节点形成的相应节点对，使得相应节点对中的节点之一具有异常。 每个因果异常排名向量用于在配对时指示多个节点中给定的一个节点的相应节点异常状态。 该方法包括计算偶然异常排名向量的稀疏惩罚以获得一组时间异常排名。 该方法包括执行该组排名的时间平滑，并基于该组排名控制多个节点中的异常发起的一个。

公开(公告)号：WO2023059868A1

公开(公告)日：2023-04-13

申请号：PCT/US2022/046020

申请日：2022-10-07

Applicant: NEC LABORATORIES AMERICA, INC.

Inventor： TANG, LuAn ,  CHEN, Yuncong ,  CHENG, Wei ,  CHEN, Zhengzhang ,  CHEN, Haifeng ,  KOBAYASHI, Yuji ,  REN, Yuxiang

IPC: G06N3/08 ,  G06N3/04 ,  G06N7/00 ,  B60W2420/42 ,  B60W2540/30 ,  B60W40/09 ,  G06F18/25 ,  G06N3/084

Abstract: Systems and methods for data fusion and analysis of vehicle sensor data, including receiving a multiple modality input data stream from a plurality of different types of vehicle sensors, determining latent features by extracting modality-specific features from the input data stream, and aligning a distribution of the latent features of different modalities by feature-level data fusion. Classification probabilities can be determined for the latent features using a fused modality scene classifier. A tree-organized neural network can be trained to determine path probabilities and issue driving pattern judgments, with the tree-organized neural network including a soft tree model and a hard decision leaf. One or more driving pattern judgments can be issued based on a probability of possible driving patterns derived from the modality-specific features.

公开(公告)号：WO2022216599A1

公开(公告)日：2022-10-13

申请号：PCT/US2022/023299

申请日：2022-04-04

Applicant: NEC LABORATORIES AMERICA, INC.

Inventor： CHEN, Yuncong ,  LUMEZANU, Cristian ,  CHENG, Wei ,  MIZOGUCHI, Takehiko ,  NATSUMEDA, Masanao ,  CHEN, Haifeng

IPC: G06F11/30 ,  G06F40/284 ,  G06F40/169 ,  G06N20/00 ,  G06N3/04

Abstract: A method for explaining sensor time series data in natural language is presented. The method includes training (1001) a neural network model with text-annotated time series data, the neural network model including a time series encoder and a text generator, allowing (1003) a human operator to select a time series segment from the text-annotated time series data, the time series segment processed by the time series encoder, outputting (1005), from the time series encoder, a sequence of hidden state vectors, one for each timestep, and generating (1007) readable explanatory texts for the human operator based on the selected time series segment, the readable explanatory texts being a set of comment texts explaining and interpreting the selected time series segment in a plurality of different ways.

公开(公告)号：WO2022109134A1

公开(公告)日：2022-05-27

申请号：PCT/US2021/059888

申请日：2021-11-18

Applicant: NEC LABORATORIES AMERICA, INC.

Inventor： CHENG, Wei ,  CHEN, Haifeng ,  NI, Jingchao ,  LUO, Dongsheng

IPC: G06N3/08 ,  G06N3/04 ,  G06N5/02

Abstract: Systems and methods for augmenting data sets is provided. The systems and methods include feeding an original document (120) into a data augmentation generator (210) to produce one or more augmented documents (220); calculating a contrastive loss (230) between the original document (120) and the one or more augmented documents (220); and using the original document (120) and the one or more augmented documents (220) to train a neural network (1030).

公开(公告)号：WO2021162953A1

公开(公告)日：2021-08-19

申请号：PCT/US2021/016846

申请日：2021-02-05

Applicant: NEC LABORATORIES AMERICA, INC.

Inventor： YU, Wenchao ,  CHENG, Wei ,  CHENG, Wei

IPC: G06N20/00

Abstract: A computer-implemented method is provided for hierarchical multi-agent imitation learning. The method includes learning (510) sub-policies for sub-tasks of a hierarchical multi-agent imitation learning task by imitating expert trajectories of expert demonstrations of the subtasks with guidance from a high-level policy corresponding to the hierarchical multi-agent imitation learning task. The method further includes collecting (520) feedback from the sub-policies relating to updating the high-level-policy with a new observation. The method also includes updating (530) the high-level policy with the new observation responsive to the feedback from the sub-policies. The high-level policy is configured as a contextual multi-arm bandit that sequentially selects k best sub-policies at each of a plurality of time steps based on contextual information derived from the expert demonstrations (510).