VJViktor Stenby Johansson
Viktor Stenby Johansson
*This is an algorithmically generated author record based on publicly available information.
Viktor Stenby Johansson has a total of 7 co-authors. Here are the top co-authors by number of shared publications:
- Akshit Achara
King's Institute for Artificial Intelligence King's College London
1 shared publication - Donato Crisostomi
Sapienza University of Rome
1 shared publication - Emanuele Rodolà
Sapienza University of Rome Italy
1 shared publication - Mateo Mahaut1 shared publication
- Melih Barsbey1 shared publication
- Pritish Chakraborty
Department of Computer Science and Engineering IIT Bombay
1 shared publication
Latest Publications
- Open Access
Multi-Way Representation Alignment
The Platonic Representation Hypothesis suggests that independently trained neural networks converge to increasingly similar latent spaces. However, current strategies for mapping these representations are inherently pairwise, scaling quadratically with the number of models and failing to yield a consistent global reference. In this paper, we study the alignment of $M \ge 3$ models. We first adapt Generalized Procrustes Analysis (GPA) to construct a shared orthogonal universe that preserves the internal geometry essential for tasks like model stitching. We then show that strict isometric alignment is suboptimal for retrieval, where agreement-maximizing methods like Canonical Correlation Analysis (CCA) typically prevail. To bridge this gap, we finally propose Geometry-Corrected Procrustes Alignment (GCPA), which establishes a robust GPA-based universe followed by a post-hoc correction for directional mismatch. Extensive experiments demonstrate that GCPA consistently improves any-to-any retrieval while retaining a practical shared reference space.Open AccessMulti-Way Representation Alignment
The Platonic Representation Hypothesis suggests that independently trained neural networks converge to increasingly similar latent spaces. However, current strategies for mapping these representations are inherently pairwise, scaling quadratically with the number of models and failing to yield a consistent global reference. In this paper, we study the alignment of $M \ge 3$ models. We first adapt Generalized Procrustes Analysis (GPA) to construct a shared orthogonal universe that preserves the internal geometry essential for tasks like model stitching. We then show that strict isometric alignment is suboptimal for retrieval, where agreement-maximizing methods like Canonical Correlation Analysis (CCA) typically prevail. To bridge this gap, we finally propose Geometry-Corrected Procrustes Alignment (GCPA), which establishes a robust GPA-based universe followed by a post-hoc correction for directional mismatch. Extensive experiments demonstrate that GCPA consistently improves any-to-any retrieval while retaining a practical shared reference space.