arxivGitHubあり2026-06-05Constructing VAE Latent Spaces with Prescribed TopologyVariational autoencoders (VAEs) learn low-dimensional latent representations of high-dimensional data. When th品質予測/異常検知生成AIVAE画像用途技術検証・論文読解補助難易度HardコストHigh★→