Slide 1

Slide 1 text

渋滞長を 予測 信号 制 御 交通管制セン タ ー カ ーナビ ・ 地図アプリ 経 路 誘導

Slide 2

Slide 2 text

• • • 1 https://www.mlit.go.jp/hakusyo/mlit/r01/hakusho/r02/html/n2611000.html https://deepmind.google/discover/blog/graphcast-ai-model-for-faster-and-more-accurate-global-weather-forecasting/

Slide 3

Slide 3 text

• • 2 https://shintosei.metro.tokyo.lg.jp/leading-project/leading-project-08/ https://www8.cao.go.jp/cstp/society5_0/bosai.html https://www3.nhk.or.jp/news/html/20210105/k10012798521000.html

Slide 4

Slide 4 text

• • • 3

Slide 5

Slide 5 text

4

Slide 6

Slide 6 text

• • • 5

Slide 7

Slide 7 text

• • 6

Slide 8

Slide 8 text

• • X Z Y X Z G Y

Slide 9

Slide 9 text

• X Z • 8 1 2 4 3 1 2 3 1 2 4 3

Slide 10

Slide 10 text

• • 9 1 2 3 4 1 2 3 4

Slide 11

Slide 11 text

• • • 10 Koh Takeuchi, Ryo Nishida, Hisashi Kashima, Masaki Onishi. ” Causal Effect Estimation on Hierarchical Spatial Graph Data. ” In Proceedings of the 29th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD), 2023.

Slide 12

Slide 12 text

11 , , , , , , , 2016, 31 , 6 , p. AG-I_1-10, 2017

Slide 13

Slide 13 text

• • 12

Slide 14

Slide 14 text

13

Slide 15

Slide 15 text

• • • • • 14 M. Barth and K. Boriboonsomsin, Real-World Carbon Dioxide Impacts of Traffic Congestion. Transportation Research Record 2058, 1 (2008).

Slide 16

Slide 16 text

• • 15 Ryu Shirakami, Toshiya Kitahara, Koh Takeuchi, Hisashi Kashima. “QTNet: Theory-based Queue Length Prediction for Urban Traffic. “ In Proceedings of the 29th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD), 2023. この車群が到着するころに 青になるよう制御 待ち行列台数 流入交通量

Slide 17

Slide 17 text

16

Slide 18

Slide 18 text

17 𝐶𝑎𝑝𝑎𝑐𝑖𝑡𝑦 𝐷𝑒𝑚𝑎𝑛𝑑 直近の状態から渋滞の発生 を予測するのは難しい

Slide 19

Slide 19 text

• • 18 Ryu Shirakami, Toshiya Kitahara, Koh Takeuchi, Hisashi Kashima. “QTNet: Theory-based Queue Length Prediction for Urban Traffic. “ In Proceedings of the 29th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD), 2023. • • • • • •

Slide 20

Slide 20 text

19 Ryu Shirakami, Toshiya Kitahara, Koh Takeuchi, Hisashi Kashima. “QTNet: Theory-based Queue Length Prediction for Urban Traffic. “ In Proceedings of the 29th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD), 2023. 𝑃𝑇𝐹, 𝑆𝑇𝐹 STGNN QT-layer Encoders Decoders 渋滞長 平均速度 流入交通量 補正項 平均速度 流入交通量 Inputs: 直近T’タイムスタンプの交通データ Outputs: 将来Tタイムスタンプ先までの交通データ 渋滞長

Slide 21

Slide 21 text

• • • • 20 Weiwei Jiang, Jiayun Luo, “Graph neural network for traffic forecasting: A survey”, Expert Systems with Applications, Vol. 207, 2022, 1 2 3 4 1 2 3 4 1 2 3 4

Slide 22

Slide 22 text

• 交通工学で知られる式 • 定常状態の待ち行列理論の式である「Littleの法則」に基づく • 「定常状態」なので時間発展は記述できない 21 道路リンク 道路リンク長 L 渋滞長 l 流入交通量 Q 自由流速度 v0 平均速度 v ρ0 , a: ハイパーパラメタ

Slide 23

Slide 23 text

• • 22 Ryu Shirakami, Toshiya Kitahara, Koh Takeuchi, Hisashi Kashima. “QTNet: Theory-based Queue Length Prediction for Urban Traffic. “ In Proceedings of the 29th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD), 2023.

Slide 24

Slide 24 text

• ➔ 23 Ryu Shirakami, Toshiya Kitahara, Koh Takeuchi, Hisashi Kashima. “QTNet: Theory-based Queue Length Prediction for Urban Traffic. “ In Proceedings of the 29th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD), 2023. 交通量 渋滞長 QT-layer STGNN QTNN 速度 流入交通量の増加によって 渋滞長が伸びそうだ

Slide 25

Slide 25 text

• 24 渋滞長を 予測 信号 制 御 交通管制セン タ ー カ ーナビ ・ 地図アプリ 経 路 誘導

Slide 26

Slide 26 text

25

Slide 27

Slide 27 text

• • • 26

Slide 28

Slide 28 text

• 27 渡り鳥※1,2の飛行軌跡データ Koh Takeuchi, Masaaki Imaizumi, Shunsuke Kanda, Yasuo Tabei, Keisuke Fujii, Ken Yoda, Masakazu Ishihata, Takuya Maekawa. Fréchet Kernel for Trajectory Data Analysis, Proceedings of the 29th International Conference on Advances in Geographic Information Systems (SIGSPATIAL), pp. 221-22, 2021

Slide 29

Slide 29 text

• • 28 Keisuke Fujii, Koh Takeuchi, Atsushi Kuribayashi, Naoya Takeishi, Yoshinobu Kawahara, Kazuya Takeda. “Estimating counterfactual treatment outcomes over time in multi-vehicle simulation,” In Proceedings of the 30th International Conference on Advances in Geographic Information Systems (SIGSPATIAL), 2022.

Slide 30

Slide 30 text

• • • • • • • • • • 29 渋滞長を 予測 信号 制 御 交通管制セン タ ー カ ーナビ ・ 地図アプリ 経 路 誘導

Slide 31

Slide 31 text

30

Slide 32

Slide 32 text

• • • 31 Koh Takeuchi, Hisashi Kashima, and Naonori Ueda. “Autoregressive Tensor Factorization for Spatio-temporal Predictions,” Proceedings of IEEE International Conference on Data Mining (ICDM), pp. 1105-1110, 2017.

Slide 33

Slide 33 text

• • • • • • 32 • • • •

Slide 34

Slide 34 text

• 33 A crash course on graph theory: https://speakerdeck.com/tasusu/a-crash-course-on-graph-theory : https://www.jstage.jst.go.jp/browse/jjsai/36/4/_contents/-char/ja : https://www.jstage.jst.go.jp/article/jjsai/38/2/38_158/_article/-char/ja/

Slide 35

Slide 35 text

• 34 https://www.mlit.go.jp/hakusyo/mlit/r01/hakusho/r02/html/n2611000.html https://www.tokyometro.jp/library_in/en/subwaymap/pdf/rosen_en_1803.pdf https://www-nature-com.kyoto-u.idm.oclc.org/articles/s41586-019-1352-7

Slide 36

Slide 36 text

• • • 35 1 2 3 4 1 2 3 4

Slide 37

Slide 37 text

• • • 36 1 2 3 4 1 2 3 4 1 2 3 4

Slide 38

Slide 38 text

• • • • 37 Weiwei Jiang, Jiayun Luo, “Graph neural network for traffic forecasting: A survey”, Expert Systems with Applications, Vol. 207, 2022, 1 2 3 4 1 2 3 4 1 2 3 4

Slide 39

Slide 39 text

• • 38 Zonghan Wu et al. "Graph WaveNet for Deep Spatial-Temporal Graph Modeling,” In Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence (IJCAI), 2019. Lei Bai et al. “Adaptive Graph Convolutional Recurrent Network for Traffic Forecasting,” In Proceediongs of NeurIPS, 2020. 1 2 3 4 1 2 3 4

Slide 40

Slide 40 text

• • • • 39 39 X C Y p(C | X)

Slide 41

Slide 41 text

• • • 40

Slide 42

Slide 42 text

• X Z Y • • 41 Koh Takeuchi, Ryo Nishida, Hisashi Kashima, Masaki Onishi. ” Grab the reins of crowds: Estimating the effects of crowd movement guidance using causal inference. ” Proceedings of the 20th International Conference on Autonomous Agents and MultiAgent Systems (AAMAS), 2021.

Slide 43

Slide 43 text

42 Xn ∈ 0,1 𝑀 𝑛 Ck ∈ 0,1 𝐾 An G Yt (t=1, …, T ) Zn An Xn N G Ck K Yt T

Slide 44

Slide 44 text

• • • 43

Slide 45

Slide 45 text

• •

Slide 46

Slide 46 text

• • • 45

Slide 47

Slide 47 text

• • 46 Koh Takeuchi, Ryo Nishida, Hisashi Kashima, Masaki Onishi. ” Causal Effect Estimation on Hierarchical Spatial Graph Data. ” In Proceedings of the 29th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD), 2023.

Slide 48

Slide 48 text

• • 47

Slide 49

Slide 49 text

階層的な空間グラフデータからの介入効果推定 • 48 Koh Takeuchi, Ryo Nishida, Hisashi Kashima, Masaki Onishi. ” Causal Effect Estimation on Hierarchical Spatial Graph Data. ” In Proceedings of the 29th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD), 2023.

Slide 50

Slide 50 text

49

Slide 51

Slide 51 text

• • • 50 F. D. Johansson et al. “Learning Representations for Counterfactual Inference,” ICML2016. U. Shalit et al. "Estimating individual treatment effect: generalization bounds and algorithms." ICML, 2017. X C f Φ g Y

Slide 52

Slide 52 text

• • 51 F. D. Johansson et al. “Learning Representations for Counterfactual Inference,” ICML2016. U. Shalit et al. "Estimating individual treatment effect: generalization bounds and algorithms." ICML, 2017. X C f Φ g Y IPM 正則化モデルの損失関数は、 の期待値の上限と証明

Slide 53

Slide 53 text

• 反事実回帰モデル • 共変量 から抽出した特徴量 の分布が介入と • 正則化モデルの損失関数が、 の期待値の上限となると証明 52 F. D. Johansson et al. “Learning Representations for Counterfactual Inference,” ICML2016. U. Shalit et al. "Estimating individual treatment effect: generalization bounds and algorithms." ICML, 2017. X C f Φ g Y IPM

Slide 54

Slide 54 text

• • •

Slide 55

Slide 55 text

• • • Z X ? Y p(Z|X) D D f

Slide 56

Slide 56 text

• Queueing-Theory layer (QT-Layer) • QT-Layer 3 55 Ryu Shirakami, Toshiya Kitahara, Koh Takeuchi, Hisashi Kashima. “QTNet: Theory-based Queue Length Prediction for Urban Traffic. “ In Proceedings of the 29th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD), 2023.