我覺得這半年生活的不是很有規律,研究上除了延續Causality的做法,在RL的使用上沒有太多的長進,我覺得主要問題有幾個:

  • 1) 對於VAE Beyasien Inference 的數學模型掌握不夠

  • 2) 這一年來Pytorch為主的架構盛行,但是我之前使用還是以TF為主,雖然會用Torch,並沒有很系統性的學習上Pytorch

  • 3) 對於Adversarial Examples, Deep Reinforcement Learning 的State of the Art 把握度被沒有到徹底精熟

  • 4) 畢業需要針對 IEEE Trans. Extension

但我覺得主要問題還是,做人需要每天檢討一下自己,瞭解一下自己的不足,因為通常繞不過去的缺點也是那麼幾個,但是人總是喜歡拖著,當作是掩耳盜鈴,這樣的生活態度如果多了,習慣了,真的是一件事很可惜的事情,還記得原本那個想把事情都做完,才放心,愉快的睡覺的自己的嗎?了解成長的喜悅,少花點時間在社群軟體上,少在乎虛榮和鏡中他人,固定多花點時間多陪陪家人,親密的人,給自己點信心,就從今天做起吧!

  • July 大事記 7th - Hardrive Backup 9:00 am - Done.

12th - IEEE/ACM NoCS Notification - Got Rejected.

To do:

  • 系統性學習 Pytorch (Done. I took CS 7643 Deep Learning from Prof. Dhruv Batra and got highest grade.)

  • 系統性學習 Dowhy Package & Causality Analysis

  • Baseline Running ARM RL

  • Check timeline for Nov. & Hitachi (Done. But stay too short owing the semester.)

  • Check Funding Source for Spring Semester (Done. Appreciate Prof. Lee for my TA and RA.)

  1. Drafting IEEE Comm Magazine (Done. but re-submission now

  2. Initiate ACCV Patent (Done.)

  3. House Moving (after 7/15 - Done. thank my roomate Jen)

  4. Deep Speech Textbook Study

  5. Udacity Differential Privacy

August

5th to 8th: Li Deng to Georgia Tech, prepare resume for internship

September

  • 6th: AAAI ddl
  • 22nd to 25th: HK Visa, IEEE ICIP
  • 27th: ICLR ddl

Dec

  • 7th: NeurIPS
  • 14th: NeurIPS