Institution profile

Fujitsu Limited

Industry researchasia · jp
Official website
Research library109linked papers
Opportunities0open roles
Selected work

Representative Papers

TULiP: Test-time Uncertainty Estimation via Linearization and Weight Perturbation

May 22, 2025

This work addresses the secure deployment of deep learning models in open-world settings, focusing on reliable uncertainty estimation for out-of-distribution (OOD) samples at test time. We propose the first theoretical framework grounded in linearized training dynamics, deriving a differentiable, posterior-style upper bound on predictive uncertainty under weight perturbations—without requiring retraining. Our method integrates training dynamic modeling, stochastic weight perturbation sampling, and prediction ensembling, accompanied by rigorous error-bound analysis. Evaluated on large-scale image-based OOD benchmarks, it achieves state-of-the-art performance, particularly improving detection accuracy for near-OOD samples. The approach bridges theoretical interpretability with practical efficiency, offering both provable guarantees and computational scalability for real-world deployment.

1 citations1 influentialRead paper

Manifold Induced Biases for Zero-shot and Few-shot Detection of Generated Images

Apr 21, 2025

Zero- and few-shot detection of AI-generated images suffers from theoretical underpinning gaps and performance bottlenecks due to reliance on forged training data. Method: This paper proposes a novel paradigm that requires no synthetic training images. It theoretically models intrinsic biases in generative content by characterizing the curvature and gradient structure of the implicit probability manifold. We further design a suite of techniques grounded in pre-trained diffusion models: (i) score-based curvature approximation, (ii) gradient bias quantification, and (iii) a Mixture-of-Experts (MoE) adaptation mechanism. Results: Our method achieves state-of-the-art performance across 20 mainstream generative models, significantly outperforming prior zero- and few-shot detectors in accuracy. It demonstrates strong generalization across unseen generators, rigorous theoretical interpretability, and practical deployability—bridging foundational analysis with real-world applicability.

1 citationsRead paper

An Experience Report on Regression-Free Repair of Deep Neural Network Model

Mar 01, 2023IEEE International Conference on Software Analysis, Evolution, and Reengineering

In high-reliability industrial systems—such as automotive image security applications—updating deep neural network (DNN) models often induces performance regression on critical classes, jeopardizing safety and reliability. Method: This paper proposes NeuRecover, the first category-specific, zero-regression neural network repair framework. It introduces a customized objective function and a class-level constrained optimization mechanism, validated empirically on real-world automotive image data. Contribution/Results: Applied to Fujitsu’s production automotive model update task, NeuRecover achieves strict zero regression for safety-critical classes—i.e., no performance degradation—while preserving primary task accuracy. The work systematically identifies practical challenges in zero-regression repair within industrial settings and establishes a viable, deployable technical paradigm for trustworthy AI model evolution.

1 citationsRead paper
Recent publications

Latest Papers