====== 2021 Project Page ======
===== Transfer learning for battery parameter estimation =====
==== Meeting Notes ====

== 2021/10/6 ==
  * Attendee: Hanbing, Yanru, Yang
  * Meeting summary: 
        * Yanru: on reading: Unsupervised Domain Adaptation for Semantic Segmentation via CBST
        * Hanbing
          * research: optimal transport and domain adaptation problem
          * reading: Joint distribution optimal transportation for domain adaptation; Few-Shot Cross Domain Battery Capacity Estimation
          * slides: {{ :trans_learn:optimal_transport_and_domain_adaptation.pptx |optimal transport and domain adaptation}}
  
  * TODO: 
        * Yanru: make a plan of future work; deeper understanding of papers/ better presentation of work
        * Hanbing: do some research on the application of adversarial network to domain adaptation problem
                     

== 2021/10/13 ==
 
  * Attendee: Hanbing, Yanru, Yang
  * Meeting summary:
        * Yanru: 
          * reading: transferable semi-supervised semantic segmentation; Adversarial Examples for Semantic Segmentation and Object Detection
          * experiment: tried MIoU metric for H-score evaluation; on coding new dataset
        * Hanbing:
          * reading: Adversarial Discriminative Domain Adaptation
          * research: Adversarial network in Domain Adaptation
          * slides: {{ :trans_learn:adversarial_domain_adaptation.pptx |adversarial domain adaptation}}
  * TODO:
        * Yanru: 
          * more relative reading on transfer learning & U-net (also maybe some survey on semantic segmentation & related area); 
          * experiment redesign 
            * target and source setting (H-score more suitable for task transfer)
            * dataset with fewer noise & lower difficulty
            * new calculate method for H-score (devision depending on label instead of pixel location)
        * Hanbing:
          * deeper reading ADDA to find its advantage over other adversarial domain adaptation works
          * modify the ADDA model to deal with battery features

== 2021/10/20 ==

  * Attendee: Hanbing, Yanru, Yang
  * Meeting summary:
        * Yanru: 
          * experiment: H-score trial on task transfer using ADE20K  -- especially good results for 'sky' task
        * Hanbing:
          * experiment: ADDA using battery pulse data 
          * slides:{{ :trans_learn:adda_test.pptx |}}
  * TODO:
        * Yanru:
          * recode the experiment with pytorch
          * cityscape dataset preprocessing
          * change of target data size
          * reading
        * Hanbing:
          * change convolution layer to fully connect or conv1D
          * transform battery data into frequency domain

== 2021/11/10 ==

  * Attendee: Hanbing, Yang
  * Meeting summary:
    * Hanbing:
      * experiment: unsupervised fully connection ADDA using battery pulse data
  * TODO:
    * Hanbing:
      * experiment:
        - source only: divide source dataset into training set and testing set to see if model is overfitting
        - target only: using only target data to get target baseline model
        - transfer learning: finetuning target feature extractor and source classifier useing a few target label

== 2021/11/14 ==
  * Attendee: Yanru, Yang
  * Meeting summary:
    - A brief review of several papers related to spatial-temporal data analyses ({{ :trans_learn:project_reading.pdf |}})
    - Notes: {{:trans_learn:notes1114.png?linkonly|}}
  * TODO: write background part of project plan 

== 2021/11/17 ==
  * Attendee: Hanbing, Yang
  * Meeting summary:
    * Hanbing:
      * experiment:
        - source only and target only on original net and simplified net  
        - add penalty term on target model
  * TODO:
    * Hanbing:
      * experiment:
        - finetunning target model useing a few target label
        - change visual form of result