Research map / Probabilistic and scientific modeling / Variational and generative sampling
VAEL: Bridging Variational Autoencoders and Probabilistic Logic Programming
Eleonora Misino, Giuseppe Marra, Emanuele Sansone
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 |
|---|---|---|---|
| replace neural label head with ProbLog logic program prior (VAEL vs CCVAE) | conditional VAE · 2digit MNIST | predictive accuracy: 0.5284 -> 0.8477 | Improved |
| replace neural label head with ProbLog logic program prior (VAEL vs CCVAE) | conditional VAE · Mario | predictive accuracy: 1.0 -> 0.977 | No clear effect |
| replace neural label head with ProbLog logic program prior (VAEL vs CCVAE) | conditional VAE · 2digit MNIST | generative accuracy: 0.5143 -> 0.7922 | Improved |
| replace neural label head with ProbLog logic program prior (VAEL vs CCVAE) | conditional VAE · Mario | generative accuracy: 0.0 -> 0.8135 | Improved |
| replace neural label head with ProbLog logic program prior (VAEL vs CCVAE) | conditional VAE · 2digit MNIST | reconstruction loss: 1549 -> 1542 | No clear effect |
| replace neural label head with ProbLog logic program prior (VAEL vs CCVAE) | conditional VAE · Mario | reconstruction loss: 43461 -> 42734 | Improved |
| train VAEL with 10 images per pair vs CCVAE with 100 per pair | VAEL · 2digit MNIST data efficiency | discriminative accuracy: 0.121 -> 0.445 | Improved |
| train VAEL with 10 images per pair vs CCVAE with 100 per pair | VAEL · 2digit MNIST data efficiency | generative accuracy: 0.284 -> 0.415 | Improved |
| swap ProbLog program at test time to new tasks without retraining | VAEL · 2digit MNIST and Mario novel tasks | qualitative generation consistency | Improved |
| more training images per pair for CCVAE on simplified 0-2 digit dataset | CCVAE · simplified 2digit MNIST | accuracy: 2500 images per pair | Improved |
Closest papers by meaning
- DeepProbLog: Neural Probabilistic Logic Programming (NeurIPS 2018 · similarity 0.88)
- Variational Autoencoders with Jointly Optimized Latent Dependency Structure (ICLR 2019 · similarity 0.87)
- InfoVAE: Balancing Learning and Inference in Variational Autoencoders (AAAI 2019 · similarity 0.87)
- DeepStochLog: Neural Stochastic Logic Programming (AAAI 2022 · similarity 0.86)
- Learning to Disentangle Latent Reasoning Rules with Language VAEs: A Systematic Study (AAAI 2026 · similarity 0.86)
- LogicMP: A Neuro-symbolic Approach for Encoding First-order Logic Constraints (ICLR 2024 · similarity 0.85)
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.
