no code implementations • 8 Sep 2023 • Hiroki Nakamura, Masashi Okada, Tadahiro Taniguchi
Moreover, experiments on image retrieval using MNIST and PascalVOC showed that the representations of our method can be operated by OR and AND operations.
no code implementations • ICCV 2023 • Hiroki Nakamura, Masashi Okada, Tadahiro Taniguchi
In this study, a novel self-supervised learning (SSL) method is proposed, which considers SSL in terms of variational inference to learn not only representation but also representation uncertainties.
no code implementations • 15 Mar 2022 • Akira Kinose, Masashi Okada, Ryo Okumura, Tadahiro Taniguchi
In this paper, we propose Multi-View Dreaming, a novel reinforcement learning agent for integrated recognition and control from multi-view observations by extending Dreaming.
no code implementations • 1 Mar 2022 • Masashi Okada, Tadahiro Taniguchi
The present paper proposes a novel reinforcement learning method with world models, DreamingV2, a collaborative extension of DreamerV2 and Dreaming.
no code implementations • 29 Jul 2020 • Masashi Okada, Tadahiro Taniguchi
In the present paper, we propose a decoder-free extension of Dreamer, a leading model-based reinforcement learning (MBRL) method from pixels.
no code implementations • 1 Mar 2020 • Masashi Okada, Norio Kosaka, Tadahiro Taniguchi
In this paper, we extend VI-MPC and PaETS, which have been originally introduced in previous literature, to address partially observable cases.
no code implementations • 31 Jan 2020 • Ryo Okumura, Masashi Okada, Tadahiro Taniguchi
We experimentally evaluated the model predictive control performance via imitation learning for continuous control of sparse reward tasks in simulators and compared it with the performance of the existing SRL method.
no code implementations • 16 Sep 2019 • Masashi Okada, Shinji Takenaka, Tadahiro Taniguchi
An important component of SMC, i. e., a proposal distribution, is designed as a probabilistic neural pose predictor, which can propose diverse and plausible hypotheses by incorporating epistemic uncertainty and heteroscedastic aleatoric uncertainty.
no code implementations • 8 Jul 2019 • Masashi Okada, Tadahiro Taniguchi
Probabilistic ensembles with trajectory sampling (PETS) is a leading type of MBRL, which employs Bayesian inference to dynamics modeling and model predictive control (MPC) with stochastic optimization via the cross entropy method (CEM).
no code implementations • 29 Jun 2017 • Masashi Okada, Luca Rigazio, Takenobu Aoshima
We also show that PI-Net is able to learn dynamics and cost models latent in the demonstrations.