公开(公告)号：WO2021195845A1

公开(公告)日：2021-10-07

申请号：PCT/CN2020/082103

申请日：2020-03-30

Applicant: ORANGE ,  WANG, Xiaoyu Richard ,  ZHENG, Tao ,  WANG, Xin

Inventor： WANG, Xiaoyu Richard ,  ZHENG, Tao ,  WANG, Xin

IPC: G06N3/02 ,  G06N3/0454 ,  G06N3/08

Abstract: Methods and systems to train artificial intelligence modules protect privacy of sensitive data by virtue of the fact that the data source (1) extracts blocks of partial data from the source item of content (11) and distributes the extracted partial data to a plurality of processing nodes (5) in an intermediate processing system (2). The processing nodes each perform an initial portion of the training process, for example by performing a convolution of the partial data, to produce a partial model (PM). The partial models (PM) are transmitted to a merging module (10) which amalgamates them and completes the training process to generate a global model (GM) for the AI task.

公开(公告)号：WO2022109771A1

公开(公告)日：2022-06-02

申请号：PCT/CN2020/131068

申请日：2020-11-24

Applicant: ORANGE ,  ZHENG, Tao ,  WANG, Xiaoyu Richard ,  WANG, Xin

Inventor： ZHENG, Tao ,  WANG, Xiaoyu Richard ,  WANG, Xin

IPC: H04N7/15 ,  H04N17/00 ,  G06F3/14

Abstract: Methods and systems to monitor, at a source device (1), the rendering at a destination device (2) of content received by the destination device from the source device. In one use case, a source device monitors the fidelity of screen-sharing by a remote device. At the destination device (2), a metric is generated characterising the content rendered at the destination device. Using this metric, the similarity between the content rendered at the destination device (2) and source content transmitted by the source device (1) is estimated. An indication of the estimated similarity is then displayed on a display (11) associated with the source device.