DATA SCIENCE LAB
INMPA9922-17
Semester: 4
Type: Laboratory
Number of Classes: 0+0+2
Credit: 3
Status: Optional
Assessment: Practical mark
Prerequisites: INMPA0101-17 (Machine learning basics)
Responsible: Dr. András Hajdu

Topics:
To get familiar with the most important challenges, tasks, tools and techniques
related to data science in research and industry. The subject focuses on
solving realistic problems, to directly apply the basic concepts and results.
There is a strong intention to co-operate also with industrial partners within R&D
projects. Challenges is Data Science (Research and Industry). Case Study,
Image Recognition. Case Study, Processing Big Data. Recommender Systems.
Digital Advertisements, Internet Search. Gaming. Fraud and Risk Detection.
Route Planning, Delivery, Optimization. Robotics. Autonomous Driving. Project
Work. Industrial Projects.

Weekly schedule:
1.,  Introduction to data science, and general lines of scientific and technological dirctions. 

2.,  Application of the Python environment for data science, virtual environment tool, repos, machine leraning technologies of data science.

3.,  Setup Python 2 and 3 enviroment with dta science tools

4.,  Case studey help with tensorflow machine learning tool

5.,  Case studey help with keras machine learning tool  
 
6.,  Case studey help with theano machine learning tool 

7.,  Case studey help with pytorh machine learning tool

8.,  Setup Intel environment tools for data science and machine learning

9.,  Case studey help with Intel machine learning tool

10., Optimization possibilities, performance comparsion, compilation from source (Intel caffe)

11., Optimization possibilities, performance comparsion, compilation from source (Intel mkl-dnn)  

12., HPC sollutions, distributed and hybrid deep learning 

13., Hardwares are dedicated for datascience: MIC, NPU, FPGA

14., Hardwares are dedicated for datascience: Intel neural compute stick

Compulsory/Recommended Readings:
- W. McKinney: Python for Data Analysis (1 ed.). O'Reilly Media, Inc. 2012.
- Christopher Bishop: Pattern Recognition and Machine Learning, Springer, 2006.
- D. Conway, J.M. White: Machine Learning for Hackers, O'Reilly Media, Inc., 2012.
- I. Goodfellow, Y. Bengio, A. Courville: Deep Learning, MIT Press, 2016.

Requirements:
It is compulsory to attend the lab. More than three absences will imply to get mark one automatically.
All the student has to undertake a project work agreed by the lecturer. 
The project work will be assigned to a student or a small sized group of the student.
The tasks of the project have to be solved independently based on unique ideas. 
There is time to ask about project work during lab sessions. 
All the students have to present, defend and share the results with the lecturer at the end of the semester. The students get the grade based on the presented results.
If it is relevant, the student can take corrective action, and retake the defend process during the exam period.



Exams/Tests
At the time and place announced on the lab course.
Szóbeli. A szorgalmi időszak végén az előre kiadott projektfeladat során elért eredmények bemutatása, megvédése szóban.

Consultations
Emailben az oktatóval előre egyezetetett időpontban.