2022/03/11

• Attendee: Zhiyuan, Xiangyu, Yang
• Meeting Summary:
  • use package ‘pyechart’ in python to visualize the cumulative record amount in different h3 hexagon with data from XIAN
  • implement LSTM on the processed data to predict the future record in each 5min time piece till 5 times
  • prepare the PPT for opening report

• To-Do
  • find the historical data construction used in literature，survey on how to jointly use long/short term data. We can use fusion or residue methods
  • find the metrics to measure the prediction other than MSE, such as MAPE
  • for solution of the data sparsity and erratic, we can make a hybrid model.
    • First, we set a threshold to filter out those districts of historical average less than the threshold, and adopt a traditional statistical model
    • Second, we can use our multi-task DNN to predict other places.
    • At last, we can hybrid these two situations together
  • find out the challenge and remedial algorithms

2022/03/09

• Attendee: Zhiyuan, Xiangyu, Yang
• Meeting Summary:
  • Zhiyuan: Share the LSTM+GCN overfitting problems, Yang gives some advice:
    • 1. Do more data mining in dataset such as visualize the prediction results and see the correlation about the closeness points in the graph.
  • Xiangyu: Show the results about META-MLP and MLP in UberNYC. Yang gives advice:
    • 1. Divide the task into weekdays and weekends looks like a cross-domain problem. In cross-domain problems, the pre-training method is often better than meta-learning, share the article about the cross-domain.
    • 2. Change the backbone network such as GCN to see the results.
  • Yang: Communicate with the Gaode group once every two weeks.

• To Do:
  • Zhiyuan and Yuanbo:
    • 1. Visualize the graph model prediction data and compare it with the ground truth using the hot map.
    • 2. Pre-training the LSTM then add the results into the GCN.
  • Xiangyu:
    • 1. Construct the few shot tasks: remove some weekends data to test the Meta-MLP and Pretraining.
    • 2. Add the baseline that uses the traditional time-series prediction method.
    • 3. Add more complex backbone models like GCN

2022/03/02

• Attendee: Zhiyuan and Yuanbo, Xiangyu, Yang
• Meeting Summary:
  • Discuss NYC taxi dataset(https://data.cityofnewyork.us/Transportation/NYC-Taxi-Zones/d3c5-ddgc): discuss the details about the datasets, pay attention to whether the pattern of ride-hailing data and taxi data is different.
  • Share related multi-task ride-hailing prediction papers.
  • Assigning pre-processing on dataset and trim the into an 8-days long slide window for region-based outflow prediction
  • Constructing a graph based on the closeness relationship among regions
  • Constructing a network based on a LSTM following with a GCN for predicting, but the loss function doesn’t converge.
  • Discuss the training of the base model. How to construct the graph construction of GCN and how to select the centroid? Is given by dataset. The training of GCN is no convergence.
  • Discuss the MLP+META compare with the pretraining method, try more metrics(MAE e.t.) and the task division method
  • Bayesian Graph Model: Read the paper “A Variational Bayesian Approach for Fast Adaptive Air Pollution Prediction”https://ieeexplore.ieee.org/document/9672075 Meta-learning under a Bayesian framework.PPT:https://cloud.tsinghua.edu.cn/f/0b7a4786c75c48108f72/

• TODO:
  • Drew the GCN network structure and have a discussion later.
  • Zhiyuan & Yuanbo:
    • We should do some explorative analysis on the data
    • Using heat map to observe if there exists convincing neighborhood similarity on the dataset
    • Using heat map to observe if GCN does smooth the discrepancy around neighborhoods and is it reasonable based on the ground truth
    • Analyze the pattern of the dataset and the detailed distribution of specific areas
    • Try to change the training session of the LSTM&GCN (can training solely on LSTM in the first few epochs)
  • Xiangyu:
    • Deduce how to variation in the bayesian meta-learning.
    • Try more methods in the task division methods in META-MLP.

2022/01/05

• Attendee: Yuanbo, Yang
• Meeting Summary:
  • Yuanbo reading: “MultiPath: Multiple Probabilistic Anchor Trajectory Hypotheses for Behavior Prediction”https://arxiv.org/abs/1910.05449v1
  • Discussion about how to produce pathlet by NN or numerical method.https://ieeexplore.ieee.org/document/7293682
  • Discussion about how to predict trajectory using pathlet.

• TODO:
  • Yuanbo: deeper understanding of paper: figure out how the authors predict the trajectory and how they construct the Input representation.
  • Do some research on how to produce pathlet by NN and predict trajectory using pathlet.

2022/01/12

• Attendee: Yuanbo, Yang, Zhiyuan
• Meeting Summary:
  • Discussion about how to produce pathlet by NN using differentiable loss function.
  • Yuanbo propose one method which matches content and index between trajectories and candidate pathlets using slide windows.
  • The detailed info can be referred to this file: zhiyuan_2022_1_12

• TODO:
  • Yuanbo & Zhiyuan: Reimplement the proposed algorithm and do someexperiments on synthetic dataset, like fixed-pattern strings.