Providing balanced training data

Two competing machine learning based models, a first model (110) for generating synthetic data (102), and a second model (120) for classifying input data (102, 103) to synthetic data and real data, are trained by training the second model (120) until its accuracy meets a preset rule, and then the fi...

Ausführliche Beschreibung

Gespeichert in:

Bibliographische Detailangaben
Hauptverfasser:	Hojjatinia, Sina, Subramanya, Tejas, Khatri, Vikramajeet, AGGARWAL, Chaitanya, MONSHIZADEH, Mehrnoosh
Format:	Patent
Sprache:	eng ; fin ; swe
Schlagworte:	CALCULATING COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS COMPUTING COUNTING ELECTRIC COMMUNICATION TECHNIQUE ELECTRIC DIGITAL DATA PROCESSING ELECTRICITY PHYSICS TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHICCOMMUNICATION WIRELESS COMMUNICATIONS NETWORKS
Online-Zugang:	Volltext bestellen
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!

Beschreibung
Zusammenfassung:	Two competing machine learning based models, a first model (110) for generating synthetic data (102), and a second model (120) for classifying input data (102, 103) to synthetic data and real data, are trained by training the second model (120) until its accuracy meets a preset rule, and then the first model (110) is trained. After training the first model (110), training of models is repeated until an end criterium is met. Kahta kilpailevaa koneoppimiseen perustuvaa mallia, ensimmäistä mallia (110) synteettisen datan (102) generoimiseksi ja toista mallia (120) syöttödatan (102, 103) luokittelemiseksi synteettiseen dataan ja todelliseen dataan, opetetaan opettamalla toista mallia (120) kunnes sen tarkkuus täyttää ennalta asetetun ehdon, ja sitten opetetaan ensimmäistä mallia (110). Ensimmäisen mallin (110) opettamisen jälkeen mallien opettamista toistetaan kunnes lopettamisehto täyttyy.