Saved in:
| Main Author: | |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2510.24356 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917048002543616 |
|---|---|
| author | Sanyal, Suman |
| author_facet | Sanyal, Suman |
| contents | We introduce Perception Learning (PeL), a paradigm that optimizes an agent's sensory interface $f_ϕ:\mathcal{X}\to\mathcal{Z}$ using task-agnostic signals, decoupled from downstream decision learning $g_θ:\mathcal{Z}\to\mathcal{Y}$. PeL directly targets label-free perceptual properties, such as stability to nuisances, informativeness without collapse, and controlled geometry, assessed via objective representation-invariant metrics. We formalize the separation of perception and decision, define perceptual properties independent of objectives or reparameterizations, and prove that PeL updates preserving sufficient invariants are orthogonal to Bayes task-risk gradients. Additionally, we provide a suite of task-agnostic evaluation metrics to certify perceptual quality. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_24356 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Perception Learning: A Formal Separation of Sensory Representation Learning from Decision Learning Sanyal, Suman Machine Learning Artificial Intelligence We introduce Perception Learning (PeL), a paradigm that optimizes an agent's sensory interface $f_ϕ:\mathcal{X}\to\mathcal{Z}$ using task-agnostic signals, decoupled from downstream decision learning $g_θ:\mathcal{Z}\to\mathcal{Y}$. PeL directly targets label-free perceptual properties, such as stability to nuisances, informativeness without collapse, and controlled geometry, assessed via objective representation-invariant metrics. We formalize the separation of perception and decision, define perceptual properties independent of objectives or reparameterizations, and prove that PeL updates preserving sufficient invariants are orthogonal to Bayes task-risk gradients. Additionally, we provide a suite of task-agnostic evaluation metrics to certify perceptual quality. |
| title | Perception Learning: A Formal Separation of Sensory Representation Learning from Decision Learning |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2510.24356 |