Neural Concept Verifier: Scaling Prover-Verifier Games via Concept Encodings
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866915773452124160 |
|---|---|
| author | Turan, Berkant Asadulla, Suhrab Steinmann, David Kersting, Kristian Stammer, Wolfgang Pokutta, Sebastian |
| author_facet | Turan, Berkant Asadulla, Suhrab Steinmann, David Kersting, Kristian Stammer, Wolfgang Pokutta, Sebastian |
| contents | While Prover-Verifier Games (PVGs) offer a promising path toward verifiability in nonlinear classification models, they have not yet been applied to complex inputs such as high-dimensional images. Conversely, expressive concept encodings effectively allow to translate such data into interpretable concepts but are often utilised in the context of low-capacity linear predictors. In this work, we push towards real-world verifiability by combining the strengths of both approaches. We introduce Neural Concept Verifier (NCV), a unified framework combining PVGs for formal verifiability with concept encodings to handle complex, high-dimensional inputs in an interpretable way. NCV achieves this by utilizing recent minimally supervised concept discovery models to extract structured concept encodings from raw inputs. A prover then selects a subset of these encodings, which a verifier, implemented as a nonlinear predictor, uses exclusively for decision-making. Our evaluations show that NCV outperforms classic concept-based models and pixel-based PVG classifier baselines on high-dimensional, logically complex datasets and helps mitigate shortcut behavior. Overall, we demonstrate NCV as a promising step toward concept-level, verifiable AI. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2507_07532 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Neural Concept Verifier: Scaling Prover-Verifier Games via Concept Encodings Turan, Berkant Asadulla, Suhrab Steinmann, David Kersting, Kristian Stammer, Wolfgang Pokutta, Sebastian Machine Learning Artificial Intelligence 68T01, 68T07 I.2.6 While Prover-Verifier Games (PVGs) offer a promising path toward verifiability in nonlinear classification models, they have not yet been applied to complex inputs such as high-dimensional images. Conversely, expressive concept encodings effectively allow to translate such data into interpretable concepts but are often utilised in the context of low-capacity linear predictors. In this work, we push towards real-world verifiability by combining the strengths of both approaches. We introduce Neural Concept Verifier (NCV), a unified framework combining PVGs for formal verifiability with concept encodings to handle complex, high-dimensional inputs in an interpretable way. NCV achieves this by utilizing recent minimally supervised concept discovery models to extract structured concept encodings from raw inputs. A prover then selects a subset of these encodings, which a verifier, implemented as a nonlinear predictor, uses exclusively for decision-making. Our evaluations show that NCV outperforms classic concept-based models and pixel-based PVG classifier baselines on high-dimensional, logically complex datasets and helps mitigate shortcut behavior. Overall, we demonstrate NCV as a promising step toward concept-level, verifiable AI. |
| title | Neural Concept Verifier: Scaling Prover-Verifier Games via Concept Encodings |
| topic | Machine Learning Artificial Intelligence 68T01, 68T07 I.2.6 |
| url | https://arxiv.org/abs/2507.07532 |