Variational Autoencoders | A Quest After Perspectiveshttps://iphysresearch.github.io/blog/tag/variational-autoencoders/Variational AutoencodersHugo Blox Builder (https://hugoblox.com)en-usTue, 25 Aug 2026 00:00:00 +0000https://iphysresearch.github.io/blog/media/sharing.pngVariational Autoencodershttps://iphysresearch.github.io/blog/tag/variational-autoencoders/Conditional variational autoencoders for cosmological model discrimination and anomaly detection in cosmic microwave background power spectrahttps://iphysresearch.github.io/blog/mypublication/2026_cmb_cvae/Tue, 25 Aug 2026 00:00:00 +0000https://iphysresearch.github.io/blog/mypublication/2026_cmb_cvae/<p>The learned latent space is examined for its relationship to cosmological parameters and their degeneracies. Reconstruction and representation-based tests explore how the model responds to spectra outside its reference cosmology.</p> <h2 id="sources">Sources</h2> <ul> <li><a href="https://link.aps.org/doi/10.1103/b4kr-srbk" target="_blank" rel="noopener">Publisher record</a></li> <li><a href="https://arxiv.org/abs/2510.27086" target="_blank" rel="noopener">arXiv abstract and submission history</a></li> <li><a href="https://api.crossref.org/works/10.1103/b4kr-srbk" target="_blank" rel="noopener">Publisher metadata via Crossref</a></li> </ul>Accelerating Bayesian sampling for massive black hole binaries with prior constraints from conditional variational autoencoderhttps://iphysresearch.github.io/blog/mypublication/2025_cvae_bayesian_sampling/Fri, 30 May 2025 00:00:00 +0000https://iphysresearch.github.io/blog/mypublication/2025_cvae_bayesian_sampling/<p>The learned model provides a rapid initial estimate, followed by conventional sampling. This combination is evaluated as a way to reduce sampling effort while retaining Bayesian precision.</p> <h2 id="sources">Sources</h2> <ul> <li><a href="https://link.aps.org/doi/10.1103/8r8b-hckx" target="_blank" rel="noopener">Publisher record</a></li> <li><a href="https://arxiv.org/abs/2502.09266" target="_blank" rel="noopener">arXiv abstract and submission history</a></li> <li><a href="https://api.crossref.org/works/10.1103/8r8b-hckx" target="_blank" rel="noopener">Publisher metadata via Crossref</a></li> </ul>