Research map / Learning theory and algorithms / Prediction and decision losses
Learning by Abstraction: The Neural State Machine
Drew Hudson, Christopher D. Manning
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 Neural State Machine with a semantic scene graph and sequential graph traversal instead of the strongest listed single-model baseline | GQA · GQA | Accuracy: 60.34 -> 63.17 | Improved |
| Use Neural State Machine instead of the strongest listed baseline | VQA-CPv2 · VQA-CPv2 | Accuracy: 40.42 -> 45.80 | Improved |
| Use Neural State Machine instead of MAC | GQA generalization: Content · GQA generalization: Content | Accuracy: 31.12 -> 40.24 | Improved |
| Use Neural State Machine instead of MAC | GQA generalization: Structure · GQA generalization: Structure | Accuracy: 47.27 -> 55.72 | Improved |
| Use Neural State Machine instead of the strongest listed baseline | GQA · GQA | Plausibility: 85.48 -> 84.28 | Hurt |
Closest papers by meaning
- Deciphering Raw Data in Neuro-Symbolic Learning with Provable Guarantees (AAAI 2024 · similarity 0.82)
- Learning with Language-Guided State Abstractions (ICLR 2024 · similarity 0.82)
- Data-Efficient Learning with Neural Programs (NeurIPS 2024 · similarity 0.82)
- A learnability analysis on neuro-symbolic learning (NeurIPS 2025 · similarity 0.81)
- Transformers Converge to Invariant Algorithmic Cores (NeurIPS 2026 · similarity 0.81)
- Concise and Logically Consistent Conformal Sets for Neuro-Symbolic Concept-Based Models (NeurIPS 2026 · similarity 0.81)
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.
