Difference between revisions of "Step 2"

From Research management course
Jump to: navigation, search
Line 3: Line 3:
 
== The seminar ==
 
== The seminar ==
 
# [https://forms.gle/gcVn8b82o7ybm5kX7 The warm-up 5-minute test]  
 
# [https://forms.gle/gcVn8b82o7ybm5kX7 The warm-up 5-minute test]  
# Linear models their role in neural networks and expert mixtures
+
# Linear models, and their role in neural networks
 
# Reporting in the academy and the industry
 
# Reporting in the academy and the industry
 
# Plan the project
 
# Plan the project

Revision as of 20:21, 27 September 2024

What is the difference between academic and industrial research projects? It is the focus. Narrow focus boosts the quality of a project and spares time. Applied scientists connect academic and industrial parts in theory and computational experiments. To narrow an industrial project one has to make a clear implementation plan. We discuss basic questions that an analyst and an expert discuss before planning.

The seminar

  1. The warm-up 5-minute test
  2. Linear models, and their role in neural networks
  3. Reporting in the academy and the industry
  4. Plan the project
  5. If someone did homework, we discuss

Resources

Step 2 YouTube video (expected with online version)

Homework

  1. For an industrial project (also known as a computational experiment) description use this template and either in your LaTeX or PDF document write your answers to these questions. The variants of these questions and creativity are yours. But please, keep it within 1–2 pages.
  2. Also, Step 1 homework reminder
  3. Refresh in your memory the matrix decompositions and multilinear models for the next warm-up test, either
    1. look for Singular value decomposition, Principal component analysis, Tensor, Multilinear map, or
    2. do fun-reading, see 4.5 Singular Value Decomposition and 10.5 PCA in High Dimensions, and see L11-4, L11.5

Examples of industrial projects

For your homework, please describe one of your projects as an analyst or expert. Note that industrial projects and scientific computational experiments have many common items in their structures.

  1. Banking credit scoring
  2. Churn prediction
  3. Next year's cash-flow forecasting
  4. Electricity consumption forecasting
  5. Customer demand forecasting
  6. Flood prediction
  7. Ranking
  8. Informational retrieval
  9. Click-through rate prediction
  10. Voting and expert estimations

This list includes some b2c problems. You can pick up yours in this area or others.

Fun

Ask: if someone would like to help with sharing the XLS-progress

Transcript of the video

Appears after the seminar.