公开(公告)号：US20230281985A1

公开(公告)日：2023-09-07

申请号：US17686273

申请日：2022-03-03

Applicant: NavInfo Europe B.V.

Inventor： Naresh Kumar Gurulingan ,  Elahe Arani ,  Bahram Zonooz

IPC: G06V10/96 ,  G06V10/80 ,  G06V10/82

CPC classification number: G06V10/96 ,  G06V10/80 ,  G06V10/82

Abstract: A deep learning framework in multi-task learning for finding a sharing scheme of representations in the decoder to best curb task interference while benefiting from complementary information sharing. A deep-learning based computer-implemented method for multi-task learning, the method including the step of progressively fusing decoders by grouping tasks stage-by-stage based on a pairwise similarity matrix between learned representations of different task decoders.

公开(公告)号：US20240037455A1

公开(公告)日：2024-02-01

申请号：US17894401

申请日：2022-08-24

Applicant: NavInfo Europe B.V.

Inventor： Naresh Kumar Gurulingan ,  Elahe Arani ,  Bahram Zonooz

IPC: G06N20/10 ,  G06K9/62 ,  G06N3/063 ,  G06N5/02

CPC classification number: G06N20/10 ,  G06K9/6256 ,  G06K9/6215 ,  G06N3/063 ,  G06N5/022

Abstract: A computer-implemented method for multi-task structural learning in artificial neural network in which both the architecture and its parameters are learned simultaneously. The method utilizes two neural operators, namely, neuron creation and neuron removal, to aid in structural learning. The method creates excess neurons by starting from a disparate network for each task. Through the progress of training, corresponding task neurons in a layer pave the way for a specialized group neuron leading to a structural change. In the task learning phase of training, different neurons specialize in different tasks. In the interleaved structural learning phase, locally similar task neurons, before being removed, transfer their knowledge to a newly created group neuron. The training is completed with a final fine-tuning phase where only the multi-task loss is used.