 |
|
|
|
|
External Awards INTERACTION 2025 Excellent Paper AwardWinners: Tomu Tominaga, NTT Human Informatics Laboratories; Naomi Yamashita, NTT Communication Science Laboratories (currently, Kyoto University); Takeshi Kurashima, NTT Human Informatics Laboratories Date: March 2, 2025 Organization: Information Processing Society of Japan (IPSJ) For “The Effects of Initial Acceptance Attitudes toward AI Decisions on Perceptions of Algorithmic Recourse.” Published as: T. Tominaga, N. Yamashita, and T. Kurashima, “The Effects of Initial Acceptance Attitudes toward AI Decisions on Perceptions of Algorithmic Recourse,” INTERACTION 2025, Tokyo, Japan, Mar. 2025 (in Japanese). Telecom Interdisciplinary Research Award (Incentive Award)Winners: Yasunori Akagi, NTT Human Informatics Laboratories; Naoki Marumo, NTT Communication Science Laboratories (currently, the University of Tokyo); Takeshi Kurashima, NTT Human Informatics Laboratories Date: March 5, 2025 Organization: The Telecommunications Advancement Foundation For “Analytically Tractable Models for Decision Making under Present Bias.” Published as: Y. Akagi, N. Marumo, and T. Kurashima, “Analytically Tractable Models for Decision Making under Present Bias,” Proc. of the 38th Conference on Artificial Intelligence (AAAI-24), pp. 9441–9450, Vancouver, Canada, Feb. 2024. Telecom System Technology AwardWinners: Hirofumi Sasaki, NTT Network Innovation Laboratories; Yasunori Yagi, NTT Network Innovation Laboratories; Riichi Kudo, NTT Network Innovation Laboratories; Doohwan Lee, NTT Network Innovation Laboratories Date: March 5, 2025 Organization: The Telecommunications Advancement Foundation For “1.58 Tbps OAM Multiplexing Wireless Transmission with Wideband Butler Matrix for Sub-THz Band.” Published as: H. Sasaki, Y. Yagi, R. Kudo, and D. Lee, “1.58 Tbps OAM Multiplexing Wireless Transmission with Wideband Butler Matrix for Sub-THz Band,” IEEE J. Sel. Areas Commun., Vol. 42, No. 6, pp. 1613–1625, 2024. 46th JSAP Outstanding Paper AwardWinners: Motoki Asano, NTT Basic Research Laboratories; Hiroshi Yamaguchi, NTT Basic Research Laboratories; Hajime Okamoto, NTT Basic Research Laboratories Date: March 14, 2025 Organization: The Japan Society of Applied Physics (JSAP) For “Cavity Optomechanical Mass Sensor in Water with Sub-femtogram Resolution.” Published as: M. Asano, H. Yamaguchi, and H. Okamoto, “Cavity Optomechanical Mass Sensor in Water with Sub-femtogram Resolution,” Appl. Phys. Express, Vol. 16, No. 3, 032002, 2023. Papers Published in Technical Journals and Conference ProceedingsExtraction of Equipment from Point Cloud Data for Automation of Heavy EquipmentY. Sakurahara, T. Inoue, Y. Goto, N. Okano, and S. Houman The 2024 IWCS Cable & Connectivity Industry Forum, Providence, Rhode Island, USA, Ocotober 2024. We are conducting research and development to improve worksite workability and safety in telecommunications equipment construction. The heavy machinery used in this work is operated by people on-site, which involves dangerous work. In addition, there is the issue of a decrease in the number of workers due to population decline. As a solution to these problems, we consider the introduction of remote/automatic operations. In this paper, we present the results of a feasibility study for a system that can monitor the work environment in real time using LiDAR (Light Detection and Ranging) and notify workers of hazardous conditions. Test-time Adaptation for Regression by Subspace AlignmentK. Adachi, S. Yamaguchi, A. Kumagai, and T. Hamagami The Thirteenth International Conference on Learning Representations (ICLR 2025), Singapore, April 2025. This paper investigates test-time adaptation (TTA) for regression, where a regression model pre-trained in a source domain is adapted to an unknown target distribution with unlabeled target data. Although regression is one of the fundamental tasks in machine learning, most of the existing TTA methods have classificationspecific designs, which assume that models output class-categorical predictions, whereas regression models typically output only single scalar values. To enable TTA for regression, we adopt a feature alignment approach, which aligns the feature distributions between the source and target domains to mitigate the domain gap. However, we found that naive feature alignment employed in existing TTA methods for classification is ineffective or even worse for regression because the features are distributed in a small subspace and many of the raw feature dimensions have little significance to the output. For an effective feature alignment in TTA for regression, we propose Significant-subspace Alignment (SSA). SSA consists of two components: subspace detection and dimension weighting. Subspace detection finds the feature subspace that is representative and significant to the output. Then, the feature alignment is performed in the subspace during TTA. Meanwhile, dimension weighting raises the importance of the dimensions of the feature subspace that have greater significance to the output. We experimentally show that SSA outperforms various baselines on real-world datasets. |