![[2505.13289] RECON: Robust symmetry discovery via Explicit Canonical Orientation Normalization](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2505.13289] RECON: Robust symmetry discovery via Explicit Canonical Orientation Normalization
Summary
The paper introduces RECON, a method for robust symmetry discovery through Explicit Canonical Orientation Normalization, enhancing data representation in machine learning.
Why It Matters
RECON addresses the challenge of instance-specific symmetries in real-world data, enabling improved model performance without retraining. This is significant for applications in computer vision and machine learning, where accurate data representation is crucial for effective learning and prediction.
Key Takeaways
- RECON provides a method for unsupervised discovery of instance-specific pose distributions.
- It enables detection of out-of-distribution poses, enhancing model robustness.
- The test-time canonicalization layer can be integrated with existing models without retraining.
- Validation shows RECON matches or outperforms other canonicalization methods in classification tasks.
- The approach is applicable to both images and molecular ensembles.
Computer Science > Machine Learning arXiv:2505.13289 (cs) [Submitted on 19 May 2025 (v1), last revised 24 Feb 2026 (this version, v3)] Title:RECON: Robust symmetry discovery via Explicit Canonical Orientation Normalization Authors:Alonso Urbano, David W. Romero, Max Zimmer, Sebastian Pokutta View a PDF of the paper titled RECON: Robust symmetry discovery via Explicit Canonical Orientation Normalization, by Alonso Urbano and 3 other authors View PDF Abstract:Real world data often exhibits unknown, instance-specific symmetries that rarely exactly match a transformation group $G$ fixed a priori. Class-pose decompositions aim to create disentangled representations by factoring inputs into invariant features and a pose $g\in G$ defined relative to a training-dependent, \emph{arbitrary} canonical representation. We introduce RECON, a class-pose agnostic \emph{canonical orientation normalization} that corrects arbitrary canonicals via a simple right translation, yielding \emph{natural}, data-aligned canonicalizations. This enables (i) unsupervised discovery of instance-specific pose distributions, (ii) detection of out-of-distribution poses and (iii) a plug-and-play \emph{test-time canonicalization layer}. This layer can be attached on top of any pre-trained model to infuse group invariance, improving its performance without retraining. We validate on images and molecular ensembles, demonstrating accurate symmetry discovery, and matching or outperforming other canonicalizations i...
