Alert button
Picture for Tsubasa Takahashi

Tsubasa Takahashi

Alert button

Understanding Likelihood of Normalizing Flow and Image Complexity through the Lens of Out-of-Distribution Detection

Feb 16, 2024
Genki Osada, Tsubasa Takahashi, Takashi Nishide

Viaarxiv icon

Scaling Private Deep Learning with Low-Rank and Sparse Gradients

Jul 06, 2022
Ryuichi Ito, Seng Pei Liew, Tsubasa Takahashi, Yuya Sasaki, Makoto Onizuka

Figure 1 for Scaling Private Deep Learning with Low-Rank and Sparse Gradients
Figure 2 for Scaling Private Deep Learning with Low-Rank and Sparse Gradients
Figure 3 for Scaling Private Deep Learning with Low-Rank and Sparse Gradients
Figure 4 for Scaling Private Deep Learning with Low-Rank and Sparse Gradients
Viaarxiv icon

Shuffle Gaussian Mechanism for Differential Privacy

Jul 04, 2022
Seng Pei Liew, Tsubasa Takahashi

Figure 1 for Shuffle Gaussian Mechanism for Differential Privacy
Figure 2 for Shuffle Gaussian Mechanism for Differential Privacy
Viaarxiv icon

Shuffled Check-in: Privacy Amplification towards Practical Distributed Learning

Jun 07, 2022
Seng Pei Liew, Satoshi Hasegawa, Tsubasa Takahashi

Figure 1 for Shuffled Check-in: Privacy Amplification towards Practical Distributed Learning
Figure 2 for Shuffled Check-in: Privacy Amplification towards Practical Distributed Learning
Figure 3 for Shuffled Check-in: Privacy Amplification towards Practical Distributed Learning
Figure 4 for Shuffled Check-in: Privacy Amplification towards Practical Distributed Learning
Viaarxiv icon

Network Shuffling: Privacy Amplification via Random Walks

Apr 08, 2022
Seng Pei Liew, Tsubasa Takahashi, Shun Takagi, Fumiyuki Kato, Yang Cao, Masatoshi Yoshikawa

Figure 1 for Network Shuffling: Privacy Amplification via Random Walks
Figure 2 for Network Shuffling: Privacy Amplification via Random Walks
Figure 3 for Network Shuffling: Privacy Amplification via Random Walks
Figure 4 for Network Shuffling: Privacy Amplification via Random Walks
Viaarxiv icon

PEARL: Data Synthesis via Private Embeddings and Adversarial Reconstruction Learning

Jun 08, 2021
Seng Pei Liew, Tsubasa Takahashi, Michihiko Ueno

Figure 1 for PEARL: Data Synthesis via Private Embeddings and Adversarial Reconstruction Learning
Figure 2 for PEARL: Data Synthesis via Private Embeddings and Adversarial Reconstruction Learning
Figure 3 for PEARL: Data Synthesis via Private Embeddings and Adversarial Reconstruction Learning
Figure 4 for PEARL: Data Synthesis via Private Embeddings and Adversarial Reconstruction Learning
Viaarxiv icon

FaceLeaks: Inference Attacks against Transfer Learning Models via Black-box Queries

Oct 27, 2020
Seng Pei Liew, Tsubasa Takahashi

Figure 1 for FaceLeaks: Inference Attacks against Transfer Learning Models via Black-box Queries
Figure 2 for FaceLeaks: Inference Attacks against Transfer Learning Models via Black-box Queries
Figure 3 for FaceLeaks: Inference Attacks against Transfer Learning Models via Black-box Queries
Figure 4 for FaceLeaks: Inference Attacks against Transfer Learning Models via Black-box Queries
Viaarxiv icon

P3GM: Private High-Dimensional Data Release via Privacy Preserving Phased Generative Model

Jun 22, 2020
Shun Takagi, Tsubasa Takahashi, Yang Cao, Masatoshi Yoshikawa

Figure 1 for P3GM: Private High-Dimensional Data Release via Privacy Preserving Phased Generative Model
Figure 2 for P3GM: Private High-Dimensional Data Release via Privacy Preserving Phased Generative Model
Figure 3 for P3GM: Private High-Dimensional Data Release via Privacy Preserving Phased Generative Model
Figure 4 for P3GM: Private High-Dimensional Data Release via Privacy Preserving Phased Generative Model
Viaarxiv icon

Differentially Private Variational Autoencoders with Term-wise Gradient Aggregation

Jun 19, 2020
Tsubasa Takahashi, Shun Takagi, Hajime Ono, Tatsuya Komatsu

Figure 1 for Differentially Private Variational Autoencoders with Term-wise Gradient Aggregation
Figure 2 for Differentially Private Variational Autoencoders with Term-wise Gradient Aggregation
Viaarxiv icon