開催報告

ACML-TAB2019開催報告

投稿日:
  • ACML2019 workshop on Machine Learning for Trajectory, Activity, and Behavior (ACML-TAB), in conjunction with The 11th Asian Conference on Machine Learning (ACML2019), Nagoya, Japan
  • ウェブサイト:http://acml-tab2019.animal-behavior-challenge.org, http://acml-tab2019.mystrikingly.com/
  • 日程:Sunday 17th, November 2019
  • 場所:WINC AICHI, Room 1103 (11th floor) or orals, 6th floor for posters
  • 招待講演:2件
  • 口頭発表:4件
  • ポスター発表:18件

Timetable

Place: Room 1103 (11th floor); see ACML2019 program for other info

10:00-10:05: Opening

10:05-10:35: Invited talk 1: Hisashi Murakami, Self-organization of human crowds and animal groups driven by inherent noise

10:35-10:45: break

10:45-11:45: Oral session

11:45-11:50: break

11:50-12:20: Invited talk 2: Eijiro Takeuchi, Ways for mobile robot to go to destination

12:30-13:30: Lunch (6th Floor)

16:00-18:00: Poster session (6th floor) with coffee break

 

Invited Talks

Dr. Hisashi Murakami (Research Center for Advanced Science and Technology, The University of Tokyo, Japan)

Title: Self-organization of human crowds and animal groups driven by inherent noise

Abstract: Collective animal behavior is a paradigmatic example of self-organization in living systems. Although individual noisy movements seem to collapse the global order, they can be compatible with and rather facilitate dynamic collective behavior of the whole group. Here, through experiments and video tracking, we show that such movements in both human crowds and fish schools appear as a scale-free movement strategy called Lévy walk, which is considered an optimal strategy when searching unpredictably distributed resources. This suggests that by seeking clear passages through a group and/or interacting with various neighbors, a noise would be generated inherently in collective groups, which facilitate efficient transition to the group-level behavior, activating self-organization of the whole group.

 

 

Dr. Eijiro Takeuchi (Graduate School of Informatics, Nagoya University, Japan)

Title: Ways for mobile robot to go to destination
Abstract: Nowadays, autonomous vehicles are actively developed and steadily approaching commercialization. Such autonomous vehicles realize point to point navigation using highly integrated mobile robot technologies. This presentation introduces fundamental components of mobile robots such as localization, recognition, planning, and control, and how to realize point to point navigation using these functions. Almost these methods are designed by physical model. On the other hand, recently, so many deep learning solutions for navigation problems are proposed. This presentation discusses relationship between recent deep learning solutions for autonomous navigation and traditional navigation functions of mobile robots.

 

Oral session 10:45-11:45 (15min each including QA)

3 Duy Nguyen Le Vo, Takuto Sakuma, Taiju Ishiyama, Hiroki Toda, Kazuya Arai, Masayuki Karasuyama, Yuta Okubo, Masayuki Sunaga, Yasuo Tabei and Ichiro Takeuchi. Statistical Significance of Discriminative Sub-trajectory
9 Hao Niu, Kei Yonekawa, Mori Kurokawa, Shinya Wada and Kiyohito Yoshihara. Transferable Representation Learning for Human Activity Recognition in Smart Homes
11 Yiming Tian, Takuya Maekawa, Joseph Korpela, Daichi Amagata, Takahiro Hara, Sakiko Matsumoto and Ken Yoda. Preliminary investigation of co-occurrence rule extraction for sex-specific behavior of Streaked Shearwater
14 Sandeep Nayak, Kazunori Ohno and Satoshi Tadokoro. Autonomous Navigation Guidance for Human through wearable light Stimuli

 

Poster session 16:00-18:00

1 Keisuke Fujii, Naoya Takeishi and Yoshinobu Kawahara. Interpretable classification of complex collective motions using graph dynamic mode decomposition
2 Kazushi Tsutsui, Keisuke Fujii and Kazuya Takeda. Data-driven modeling of locomotor behaviors in game-based chase and escape interactions
3 Duy Nguyen Le Vo, Takuto Sakuma, Taiju Ishiyama, Hiroki Toda, Kazuya Arai, Masayuki Karasuyama, Yuta Okubo, Masayuki Sunaga, Yasuo Tabei and Ichiro Takeuchi. Statistical Significance of Discriminative Sub-trajectory
4 Ryota Tomonaka, Toru Tamaki, Bisser Raytchev, Kazufumi Kaneda and Ken Yoda. On Trajectory Interpolation using LSTM
5 Kazuki Fujimori, Bisser Raytchev, Kazufumi Kaneda, Emyo Fujioka, Shizuko Hiryu and Toru Tamaki. Position estimation using multi-channel audio signals
6 Chentao Wen and Kotaro Kimura. Analyzing whole brain activities of a worm using data-driven models
7 Ryota Shimizu, Takahiro Uchiya and Ichi Takumi. Congestion Mitigation Verification using a Theme Park Guide Schedule
8 Yukihiro Achiha, Tsubasa Hirakawa, Takayoshi Yamashita and Hironobu Fujiyoshi. Flow Histogram-based Recurrent Neural Network for Visual Odometry Estimation
9 Hao Niu, Kei Yonekawa, Mori Kurokawa, Shinya Wada and Kiyohito Yoshihara. Transferable Representation Learning for Human Activity Recognition in Smart Homes
10 Shiho Koyama, Yuichi Mizutani and Ken Yoda. Unraveling the foraging strategies of breeding seabirds by combining trajectory, activity, and physiology
11 Yiming Tian, Takuya Maekawa, Joseph Korpela, Daichi Amagata, Takahiro Hara, Sakiko Matsumoto and Ken Yoda. Preliminary investigation of co-occurrence rule extraction for sex-specific behavior of Streaked Shearwater
12 Javier Zazo, Melanie F. Pradier and Santiago Zazo. Distributed Non-Convex Least Squares Localization Problem
13 Tiwat Larpvisuttisaroj and Koichi Hashimoto. Deep Neural Network for Estimating Bat Pose
14 Sandeep Nayak, Kazunori Ohno and Satoshi Tadokoro. Autonomous Navigation Guidance for Human through wearable light Stimuli
15 Shintaro Takayama, Masao Kuwahara, Yosuke Kawasaki, Shogo Umeda and Koichi Hashimoto. Vehicle anormality evaluation using probe trajectory data
16 Yasutaka Furusho and Kazushi Ikeda. Generation and Visualization of Tennis Swing Motion by Conditional Variational RNN with Hidden Markov Model
17 Shinsuke Kajioka, Takuto Sakuma and Ichiro Takeuchi. Comparative Sequential Pattern Mining of Human Trajectory Obtained from BLE Beacons Collected by Smartphones

18 Ranulfo Bezerra, Kazunori Ohno, Thomas Westfechtel and Satoshi Tadokoro. Pedestrian Flow Estimation Using Sparse Observation from Autonomous Vehicles