Processing

Please wait...

Settings

Settings

Goto Application

1. WO2022003509 - COMPUTER-IMPLEMENTED METHOD FOR THE GENERATION OF A MATHEMATICAL MODEL WITH REDUCED COMPUTATIONAL COMPLEXITY

Publication Number WO/2022/003509
Publication Date 06.01.2022
International Application No. PCT/IB2021/055646
International Filing Date 25.06.2021
IPC
G06N 3/04 2006.1
GPHYSICS
06COMPUTING; CALCULATING OR COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
04Architecture, e.g. interconnection topology
G06N 3/08 2006.1
GPHYSICS
06COMPUTING; CALCULATING OR COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
08Learning methods
CPC
G06N 3/0454
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
04Architectures, e.g. interconnection topology
0454using a combination of multiple neural nets
G06N 3/084
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
08Learning methods
084Back-propagation
Applicants
  • POLITECNICO DI MILANO [IT]/[IT]
Inventors
  • DEDE', Luca
  • QUARTERONI, Alfio
  • REGAZZONI, Francesco
Agents
  • GRANA, Daniele
Priority Data
10202000001561929.06.2020IT
Publication Language English (en)
Filing Language Italian (IT)
Designated States
Title
(EN) COMPUTER-IMPLEMENTED METHOD FOR THE GENERATION OF A MATHEMATICAL MODEL WITH REDUCED COMPUTATIONAL COMPLEXITY
(FR) PROCÉDÉ MIS EN ŒUVRE PAR ORDINATEUR DE GÉNÉRATION D'UN MODÈLE MATHÉMATIQUE À COMPLEXITÉ DE CALCUL RÉDUITE
Abstract
(EN) The computer- implemented method (M) for the generation of a mathematical model with reduced computational complexity comprises at least the following steps: receiving at input a dataset comprising a plurality of input- output pairs relating to a phenomenon under consideration (step 1); receiving at input a plurality of functions ( Jphys, i ) relating to principles (1,..., Np) governing the phenomenon under consideration (step 2); determining a first term (Jdata) starting from the dataset comprising a plurality of input- output pairs (step 3); etermining a second term (Jphys) starting from the functions ( JPhys,i ) relating to principles (1,..., Np) (step 4); generating a mathematical model by means of at least one artificial neural network (f, g) trained on the basis of a loss function obtained starting from the first term (Jdata) and the second term (Jphys) (step 5).
(FR) L'invention concerne un procédé (M) mis en œuvre par ordinateur de génération d'un modèle mathématique à complexité de calcul réduite, comprenant au moins les étapes suivantes consistant à : recevoir, au niveau d'une entrée, un ensemble de données comprenant une pluralité de paires entrée-sortie relatives à un phénomène considéré (étape 1) ; recevoir, au niveau d'une entrée, une pluralité de fonctions (Jphys, i ) relative à des principes (1,..., Np) gouvernant le phénomène considéré (étape 2) ; déterminer un premier terme (Jdata) en commençant à partir de l'ensemble de données comprenant une pluralité de paires entrée-sortie (étape 3) ; déterminer un second terme (Jphys) en commençant à partir des fonctions (JPhys, i) relatives aux principes (1,..., Np) (étape 4) ; générer un modèle mathématique au moyen d'au moins un réseau neuronal artificiel (f, g) entraîné sur la base d'une fonction de perte obtenue en commençant à partir du premier terme (Jdata) et du second terme (Jphys) (étape 5).
Latest bibliographic data on file with the International Bureau