221222 (Thu)

 

7:25	Lab arrival with AEE [1]
~ 8:10	English speaking with Live Academy [2]
~ 8:42	Literature survey
~ 10:00	Reinforcement Learning Study [3]
~ 11:24	Physics-Guided Wake Modeling [4]
~ 12:00	Lunch
~ 13:14	Physics-Guided Wake Modeling [4]
~ 14:03	Python Study [5]
~ 18:10	Manuscript [6]
~ 20:00	Dinner
~ 22:00	Manuscript [6]

 

Study

[1] English listening (All Ears English)

• AEE 1908: 3 Native English Grammar Mistakes to Avoid
• https://www.youtube.com/watch?v=C973mt8G8tg 

[2] English speaking (Youtube - Live Academy)

• https://www.youtube.com/watch?v=fyUAE6koJHE&list=PLIsIUJcT0HIU3r526hLOSkF3zAw7tqyx8&index=13

[3] Reinforcement Learning (Hands-On Machine Learning with Scikit-Learn, Keras & TensorFlow, written by Aurelien Geron, published by O'REILLY)

• Chapter 18: Reinforcement Learning
• 18.1 Learning for reward optimization ~ 18.6 Policy gradient

[5] Python (윤성우의 열혈 파이썬 중급편, written by Sungwoo Yoon, published by Orange Media)

• Story 31: nested function and closure
• Story 32: decorator

 

Research

[4] Physics-Guided Wake Modeling

• Attempt to reduce dimension of QoI
  • Failed since the trend of QoI distribution is so simple --> no worth for PCA
• Simple sensitivity analysis wrt wake parameters vs y-coordinates of peak point
  • In terms of Spearman coefficient, only 2nd wake parameter showed statistically significant negative correlation with y-coord of peak

[6] Writing draft of manuscript

• Working title: Towards Quantifying Calibrated Uncertainty via Deep Ensembles in Missile Performance Regression Tasks
• Revision of the section named "4.2 Assessment of prediction performance"
  • Also, revision of the figure below

• Revision of the section "4.3 Assessment of uncertainty estimation performance"