MEASURING RELATEDNESS BETWEEN PREDICTION TASKS IN ARTIFICIAL INTELLIGENCE AND CONTINUAL LEARNING SYSTEMS

Patent №

US 11,836,751

Granted

2023-12-05

Filed 2020

Owner

NEC LABORATORIES EUROPE GMBH

+1 more

Lab

AI components

8

ml · nlp · vision · speech · kr · planning · evo · hardware

Assignment

Recorded

Dataset

AIPD

2023_r1 edition

Application

16807202

A method for measuring relatedness between prediction tasks includes receiving data for a first prediction task. The method further includes measuring the relatedness of the first prediction task to at least one previous prediction task as a difference between divergence of conditional probabilities of the tasks. The method can be advantageously applied in artificial intelligence or continual learning systems.

Machine learningNatural languageVisionSpeechKnowledge representationPlanningEvolutionary computationAI hardwareG06Q 30/0206G06F 17/16G06F 17/18G06F 18/217G06F 18/2431G06N 3/08G06N 3/09G06N 5/022+1 more

AI classification

Machine learning1.00
AI hardware1.00
Vision1.00
Natural language1.00
Planning0.99
Knowledge representation0.98
Evolutionary computation0.88
Speech0.65

Ownership

NEC LABORATORIES EUROPE GMBH

assignment · 521170161

NEC CORPORATION

assignment · 653140707

Assignors

YU, SHUJIAN, SHAKER, AMMAR

On an employer assignment, the assignors are typically the inventors.

© 2026 NYSGPT2525 LLC