Preprints
Nariki Tanaka, Hiroshi Kera, Kazuhiko Kawamoto,
Fourier Analysis on Robustness of Graph Convolutional Neural Networks for Skeleton-based Action Recognition,
https://arxiv.org/abs/2301.08874
Go Nan, Hiroshi Kera, Kazuhiko Kawamoto,
Improving Accuracy of Zero-Shot Action Recognition with Handcrafted Features,
https://arxiv.org/abs/2301.08874
Ryosuke Masuya, Yuichi Ike, Hiroshi Kera,
Vanishing Component Analysis with Contrastive Normalization,
https://arxiv.org/abs/2210.03349
Kosuke Sumiyasu, Kazuhiko Kawamoto, Hiroshi Kera,
Game-Theoretic Understanding of Misclassification,
https://arxiv.org/abs/2210.03349
Wataru Okamoto, Hiroshi Kera, and Kazuhiko Kawamoto
Reinforcement Learning with Adaptive Curriculum Dynamics Randomization for Fault-Tolerant Robot Control,
https://arxiv.org/pdf/2111.10005.pdf
Hiroshi Kera and Yoshihiko Hasegawa
Monomial-agnostic computation of vanishing ideals,
https://arxiv.org/abs/2101.00243 (Update on Mar. 13, 2022)
(Code: https://github.com/HiroshiKERA/monomial-agnostic-vanishing-ideal)
Soichiro Kumano, Hiroshi Kera, and Toshihiko Yamasaki
Sparse Fooling Images: Fooling Machine Perception through Unrecognizable Images,
https://arxiv.org/abs/2012.03843
Journal Papers
Takeshi Haga, Hiroshi Kera, and Kazuhiko Kawamoto.
Sequential Variational Autoencoder with Adversarial Classifier for Video Disentanglement,
Sensors 23, no. 5: 2515, 2023 [IF2022 = 3.84]
Kazuma Fujii, Hiroshi Kera, and Kazuhiko Kawamoto
Adversarially Trained Object Detector for Unsupervised Domain Adaptation,
IEEE Access, vol 10, pp. 59534–59543, 2022 [IF2021 = 3.58]
Kevin Richard G. Operiano, Wanchalerm Pora, Hitoshi Iba, and Hiroshi Kera
Evolving Architectures with Gradient Misalignment toward Low Adversarial Transferability,
IEEE Access vol. 9, pp. 164379–164393, 2021 [IF2020 = 3.37]
Hiroshi Kera and Yoshihiko Hasegawa
Spurious Vanishing Problem in Approximate Vanishing Ideal,
IEEE Access vol 7, pp.178961–178976, 2019 [IF2018 = 4.64]
Hiroshi Kera and Yoshihiko Hasegawa,
Noise-Tolerant Algebraic Method for Reconstruction of Nonlinear Dynamical Systems,
Nonlinear Dynamics, vol. 85, pp.675-692, 2016 [IF2015 = 3.42]
Conference Papers
International
Elias Wirth, Hiroshi Kera, Sebastian Pokutta
Approximate Vanishing Ideal Computations at Scale,
The Eleventh International Conference on Learning Representations (ICLR), 2023.
Aoshi Kawaguchi, Hiroshi Kera, Toshihiko Yamasaki,
Epoch-Wise Double Descent Triggered by Learning a Single Sample
In Proceedings of Asia Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC), to appear, 2022.
Chun Yang Tan, Kazuhiko Kawamoto, Hiroshi Kera,
Adversarial amplitude swap towards robust image classifiers,
European Conference on Computer Vision Workshop on Adversarial Robustness in the Real World (ECCV AROW), to appear, 2022. (Short paper)
https://arxiv.org/abs/2203.07138
Takaaki Azakami, Hiroshi Kera, Kazuhiko Kawamoto,
Adversarial Body Shape Search for Legged Robots,
In Proceedings of International Conference on Systems, Man, and Cybernetics (SMC 2022), pp. 682-687, Oct. 2022.
https://arxiv.org/abs/2203.07138
Takuto Otomo, Hiroshi Kera, Kazuhiko Kawamoto,
Adversarial joint attacks on legged robots,
In Proceedings of International Conference on Systems, Man, and Cybernetics (SMC 2022), pp. 676-681, Oct. 2022.
https://arxiv.org/abs/2205.10098
Soichiro Kumano, Hiroshi Kera, and Toshihiko Yamasaki
Superclass Adversarial Attack,
International Conference on Machine Learning Workshop on Adversarial Machine Learning Frontiers (ICML Workshop 2022), to appear, Jul. 2022 (arXiv: https://arxiv.org/abs/2205.14629)
Hiroshi Kera
Border basis computation with gradient-weighted normalization,
In Proceedings of the International Symposium on Symbolic and Algebraic Computation (ISSAC 2022), pp. 225–234, Jul. 2022 (arXiv: https://arxiv.org/abs/2101.00401)
Nariki Tanaka, Hiroshi Kera, and Kazuhiko Kawamoto
Adversarial Bone Length Attack on Action Recognition,
In Proceedings of the Thirty-sixth AAAI Conference on Artificial Intelligence (AAAI 2022), pp. 2335–2343, Feb. 2022 (acceptance rate < 15%) (arXiv: https://arxiv.org/abs/2109.05830)
Hiroshi Kera and Yoshihiko Hasegawa.
Gradient Boosts the Approximate Vanishing Ideal,
In Proceedings of the Thirty-fourth AAAI Conference on Artificial Intelligence (AAAI 2020), pp.4428–4435, Feb. 2020 (acceptance rate < 21%) (arXiv: https://arxiv.org/abs/1911.04174)
Hiroshi Kera and Yoshihiko Hasegawa.
Approximate Vanishing Ideal via Data Knotting,
In Proceedings of the Thirty-Second AAAI Conference on Artificial Intelligence (AAAI 2018), pp. 3399–3406, Feb. 2018 (acceptance rate < 25%) (arXiv: https://arxiv.org/abs/1801.09367)
Yifei Huang, Minjie Cai, Hiroshi Kera, Ryo Yonetani, Keita Higuchi, and Yoichi Sato, Temporal Localization and Spatial Segmentation of Joint Attention in Multiple First-Person Videos,
International Workshop on Egocentric Perception, Interaction, and Computing (EPIC 2017) in conjunction with ICCV 2017, pp. 2313–2321, Oct. 2017.
Hiroshi Kera, Ryo Yonetani, Keita Higuchi, and Yoichi Sato
Discovering Objects of Joint Attention via First-Person Sensing,
In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshop on Egocentric Vision, pp.7-15, Jul. 2016
Hiroshi Kera and Hitoshi Iba
Vanishing Ideal Genetic Programming,
In Proceedings of the IEEE Congress on Evolutionary Computation (CEC 2016), pp.5018-5025, Jul. 2016 (Oral presentation)
Domestic
Soichiro Kumano, Hiroshi Kera, Toshihiko Yamasaki
Probabilistic Approach towards Theoretical Understanding for Adversarial Training
信学技報, 画像工学研究会 (IE), vol.122, no.385, IE2022-76, pp.95-100, Feb.21-22, 2023．（IE賞）
川口碧, 計良宥志, 山崎俊彦
データ内の難易度差と汎化性能への影響
信学技報, 画像工学研究会 (IE), vol.122, no.385, IE2022-74, pp.83-88, Feb.21-22, 2023
芳賀 壮，計良宥志，川本一彦，
動画像の解きほぐしに向けた敵対的補助分類器の効果検証，
信学技報, Vol. 122, No. 181, PRMU2022-21, pp. 67-71, 2022年9月. （月間ベストプレゼンテーション賞）
Soichiro Kumano, Hiroshi Kera, Toshihiko Yamasaki
Superclass Adversarial Attack,
第25回画像の認識・理解シンポジウム (MIRU), 2022
川口 碧，計良宥志，山崎俊彦
単一サンプル学習におけるDouble Descent現象の解析,
第25回画像の認識・理解シンポジウム (MIRU), 2022
芳賀壮，計良宥志，川本一彦，
動画像の解きほぐしに向けた補助分類器の効果検証，
第25回画像の認識・理解シンポジウム，2022.
Chun Yang Tan, Kazuhiko Kawamoto, and Hiroshi Kera
Adversarial amplitude swap towards robust image classifiers,
The 25th Meeting on Image Recognition and Understanding, 2022.
七海友康，川本一彦，計良宥志，
変分オートエンコーダを用いたテールサンプル生成，
第25回画像の認識・理解シンポジウム，2022.
住安宏介，川本一彦，計良宥志，
相互作用値を用いた敵対的転移性の評価，
第25回画像の認識・理解シンポジウム，2022.
田中成樹, 計良宥志, 川本一彦，
骨の長さの変化のみによる敵対的攻撃，
第25回画像の認識・理解シンポジウム，2022.
木内拓実，計良宥志，川本 一彦，
クラス不均衡下でのドメイン適応における勾配逆伝播と対照学習の比較，
第25回画像の認識・理解シンポジウム，2022.
大友拓門，計良宥志，川本一彦，
歩行ロボット学習における敵対的な関節攻撃，
3Yin2-30，第36回人工知能学会全国大会，2022.
阿座上剛明，計良宥志，川本 一彦，
歩行ロボット学習における敵対的な体型探索，
3Yin2-29，第36回人工知能学会全国大会，2022.
住安 宏介，川本 一彦，計良 宥志
ゲーム理論に基づく敵対的転移性の理解，
情報処理学会研究報告. Vol. 2022-CVIM-230, No.29, pp.1-8, 2022.（卒論セッション 優秀賞）
岡本航昇，計良宥志，川本一彦
耐故障ロボット制御のためのベイズ最適化による適応的ダイナミクスランダム化
DS2ELDiA研究会, 3月 2022
藤井 一磨，計良 宥志，川本一彦
敵対的訓練を用いたドメイン不変な特徴抽出
情報処理学会研究報告. Vol. 2022-CVIM-229, No.35, pp.1-8, 2022.
Kevin Richard Operiano, Wanchalerm Pora, 伊庭斉志，計良宥志
Reducing Transferability using Neuroevolution with Gradient Misalignment,
進化計算シンポジウム2021, 12月 2021
岡本航昇，計良宥志，川本一彦，
耐故障ロボット制御のためのダイナミクスランダム化を用いたカリキュラム強化学習
W-B4-1，インテリジェント・システム・シンポジウム，2021（優秀論文賞）
芳賀壮，計良宥志，川本一彦
変分オートエンコーダを用いた動画像の解きほぐし
第24回画像の認識・理解シンポジウム (MIRU), Jul. 27-30, 2021
木内拓実，計良宥志，川本一彦
ドメイン適応を用いた一人称行動認識
第24回画像の認識・理解シンポジウム (MIRU), Jul. 27-30, 2021
Soichiro Kumano, Hiroshi Kera, Toshihiko Yamasaki
Sparse Fooling Images: Fooling Neural Networks through Unrecognizable Images,
第24回画像の認識・理解シンポジウム (MIRU), Jul. 27-30, 2021（ショートオーラル）
熊野創一郎，計良宥志，山崎俊彦
Sparse Fooling Images: 認識誤りを誘発するピクセル画像
2021年電子情報通信学会総合大会, D-12-12, 情報・システム講演論文集2 , p. 53, online, Mar.9-12, 2021.
計良宥志，長谷川禎彦
Approximate Vanishing Ideal via Data Knotting,
第17回情報科学技術フォーラム (FIT2018), Sep. 2018, （既発表論文枠）
島崎隆，計良宥志，伊庭斉志
CMBL:共分散最大化基底学習器による非線形特徴量抽出,
(CMBL: Covariance Maximized Basis Learner for nonlinear feature extraction),
第4回 人工知能学会 汎用人工知能研究会 (SIG-AGI), 12月 2016
Yifei Huang, Minjie Cai, Hiroshi Kera, Ryo Yonetani, Keita Higuchi, and Yoichi Sato, Spatial-temporal segmentation of joint attention in multiple first-person videos,
In Extended Abstract of Meeting on Image Recognition and Understanding (MIRU), Aug 2017.
Hiroshi Kera, Ryo Yonetani, Keita Higuchi, and Yoichi Sato
Discovering Objects of Shared Attention via First-Person Sensing,
In Extended Abstract of Meeting on Image Recognition and Understanding (MIRU), Aug. 2016
計良宥志，伊庭斉志
消失イデアル遺伝的プログラミング,
進化計算シンポジウム2015, 12月 2015 (IEEE CIS Japan Chapter Young Researcher Award)
Awards
東京大学大学院 情報理工学系研究科 電子情報学専攻 電子情報学専攻長賞，2020
IEEE CIS Japan Chapter Young Researcher Award (Evolutionary Computation), 2015 (http://www.ieee-jp.org/section/tokyo/chapter/CIS-11/index.php?TopPage-J%2FYRA-J%2F2015)
東京大学工学部電子情報工学科 優秀卒業論文賞，2015
Hiroshi Kera 