Deep Learning (5 cr)

Evaluation scale

0-5

Objective

Opiskelija ymmärtää syväoppimisen merkityksen digitalisoituvassa toimintaympäristössä. Opiskelija tietää yleisimmät syväoppimisen menetelmät, osaa soveltaa niitä käytännössä olemassa olevaan dataan sekä tulkita menetelmien tulokset.

Content

- Erilaiset neuroverkot
- Neuroverkkojen arkkitehtuurit ja käyttötarkoitukset (esim. CNN, RNN, LSTM, Autoencoder)
- Avoimen lähdekoodin neuroverkkokirjastoilla työskentely
- Siirretty oppiminen (transfer learning)
- Ennustaminen (prediction) ja luokittelu (classification)
- Sovellusalueet (esim. Konenäkö, NLP)

Assessment criteria, satisfactory (1)

Excellent 5: The student recognizes the advantages of Deep Learning in the era of digitalization. The student knows the most commonly used techniques and is able to critically justify the used techniques in various tasks. The student is able to apply his/her technical know-how in practice and critically assess his/her implementation as well as validate its development.

Very good 4: The student recognizes the advantages of Deep Learning in the era of digitalization. The student knows the most commonly used techniques of Deep Learning and is able to justify versatilely the use of the implemented techniques in various tasks. The student knows how to apply his/her technical know-how in practice and assess his/her implementation critically as well as validate its development.

Good 3: The student is aware of the advantages of Deep Learning in the era of digitalization. The student knows the most commonly used techniques of Deep Learning for various problems. The student is able to apply his/her technical know-how I practice and assess his/her implementation in practice and validate its development.

Satisfactory 2: The student knows the most commonly used techniques in Deep Learning for various problems. He/She knows how to choose the techniques of Deep Learning and apply his/her technical know-how in practice. In addition, the student knows how to assess his/her implementation superficially.

Sufficient 1: The student knows the most commonly used techniques of Deep Learning. He/She is able to apply the most common techniques of Deep Learning. Additionally, the student is able to assess his/her implementation briefly.

Fail 0: The student does not meet the minimum criteria set for the course.

Qualifications

Basics in computing, programming, knowledge and know-how of Python programming language.

Additionally, course in Data preprocessing.