Research map / Probabilistic and scientific modeling / Variational and generative sampling
Additive Decoders for Latent Variables Identification and Cartesian-Product Extrapolation
Sebastien Lachapelle, Divyat Mahajan, Ioannis Mitliagkas, Simon Lacoste-Julien
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use additive decoder instead of non-additive (ablation) | decoder for latent identification · ScalarLatents | LMS_Spear: 70.6 -> 91.5 | Improved |
| use additive decoder instead of non-additive (ablation) | decoder for latent identification · ScalarLatents out of support | LMS_Spear_OOS: 73.7 -> 89.5 | Improved |
| use additive decoder instead of non-additive (ablation) | decoder for latent identification · ScalarLatents out of support | RMSE_OOS: .18 -> .11 | Improved |
| use additive decoder instead of non-additive (ablation) | decoder for latent identification · ScalarLatents | RMSE: .06 -> .06 | No clear effect |
| use additive decoder instead of non-additive (ablation) | decoder for latent identification · BlockLatents independent z | LMS_Tree: 53.9 -> 92.2 | Improved |
| use additive decoder instead of non-additive (ablation) | decoder for latent identification · BlockLatents dependent z | LMS_Tree: 78.1 -> 99.9 | Improved |
| use additive decoder instead of non-additive (ablation) | decoder for latent identification · BlockLatents independent z | RMSE: .02 -> .03 | Hurt |
| use additive decoder instead of non-additive (ablation) | decoder for latent identification · BlockLatents dependent z | RMSE: .02 -> .01 | No clear effect |
| violate path-connected latent support assumption (ablation) | additive decoder · Disconnected-support variant of ScalarLatents | LMS_Spear | Hurt |
Closest papers by meaning
- Multi-View Causal Representation Learning with Partial Observability (ICLR 2024 · similarity 0.84)
- Disentangling Factors of Variation by Mixing Them (CVPR 2018 · similarity 0.84)
- Local Disentanglement in Variational Auto-Encoders Using Jacobian $L_1$ Regularization (NeurIPS 2021 · similarity 0.84)
- Overcoming the Disentanglement vs Reconstruction Trade-off via Jacobian Supervision (ICLR 2019 · similarity 0.84)
- ADDP: Learning General Representations for Image Recognition and Generation with Alternating Denoising Diffusion Process (ICLR 2024 · similarity 0.84)
- BiGR: Harnessing Binary Latent Codes for Image Generation and Improved Visual Representation Capabilities (ICLR 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
