An Exploratory Study of AI System Risk Assessment from the Lens of Data Distribution and Uncertainty

no code implementations • 13 Dec 2022 • Zhijie Wang, Yuheng Huang, Lei Ma, Haruki Yokoyama, Susumu Tokumoto, Kazuki Munakata

More importantly, it also lacks systematic investigation on how to perform the risk assessment for AI systems from unit level to system level.

Verifying Attention Robustness of Deep Neural Networks against Semantic Perturbations

1 code implementation • 13 Jul 2022 • Satoshi Munakata, Caterina Urban, Haruki Yokoyama, Koji Yamamoto, Kazuki Munakata

Semantic perturbations can significantly change the saliency-map.

valid

NeuRecover: Regression-Controlled Repair of Deep Neural Networks with Training History

no code implementations • 1 Mar 2022 • Shogo Tokui, Susumu Tokumoto, Akihito Yoshii, Fuyuki Ishikawa, Takao Nakagawa, Kazuki Munakata, Shinji Kikuchi

Search-based repair techniques for DNNs have potentials to tackle this challenge by enabling localized updates only on "responsible parameters" inside the DNN.

regression

AugmentedCode: Examining the Effects of Natural Language Resources in Code Retrieval Models

1 code implementation • 16 Oct 2021 • Mehdi Bahrami, N. C. Shrikanth, Yuji Mizobuchi, Lei Liu, Masahiro Fukuyori, Wei-Peng Chen, Kazuki Munakata

Code retrieval is allowing software engineers to search codes through a natural language query, which relies on both natural language processing and software engineering techniques.

Retrieval