arXiv:2606. 01734v1 Announce Type: cross Abstract: This paper proposes ``FlatVPR,'' a novel geometric rectification paradigm that effectively bridges the trade-off between map lightweightness and localization accuracy in visual place recognition (VPR) by enforcing a feature manifold structure where any descriptor between two adjacent anchors $\mathbf{z}_A$ and $\mathbf{z}_B$ can be accurately reconstructed via linear interpolation $\hat{\mathbf{z}}_{pseudo} = (1-t)\mathbf{z}_A + t\mathbf{z}_B$, where $t \in [0,1]$ denotes the relative position.
By Rai Hisada, Kanji Tanaka
This paper proposes ``FlatVPR,'' a novel geometric rectification paradigm that effectively bridges the trade-off between map lightweightness and localization accuracy in visual place recognition (VPR) by enforcing a feature manifold structure where any descriptor between two adjacent anchors $\mathbf{z}_A$ and $\mathbf{z}_B$ can be accurately reconstructed via linear interpolation $\hat{\mathbf{z}}_{pseudo} = (1-t)\mathbf{z}_A + t\mathbf{z}_B$, where $t \in [0,1]$ denotes the relative position. While state-of-the-art foundation models such as DINOv2-ViT-S/14 provide robust semantic features, their latent manifolds exhibit prominent curvature, projecting uniform linear motion in physical space onto highly non-linear trajectories in the feature space, which hinders reliable reconstruction under sparse anchor conditions.
RecKAN introduces a learnable recursive polynomial basis for Kolmogorov–Arnold Networks, replacing fixed bases like B-splines or Chebyshev polynomials. The basis is defined by a second‑order polynomial recurrence whose five coefficients are jointly learned with the network, enabling it to encompass classical families such as Chebyshev, Fibonacci, Pell, and Jacobsthal. Experiments across image, text, biomedical time‑series classification, and forecasting tasks show RecKAN outperforming parameter‑matched KAN baselines and achieving state‑of‑the‑art results on several benchmarks.
By Amirhosein Azarpour
arXiv:2607. 07034v1 Announce Type: cross Abstract: We introduce Intrinsic Green's Learning (IGL), a framework that models a target function on a manifold as the solution to a linear PDE whose source term is learned from data.
By Alexandre Quemy
arXiv:2609.31458v1 Announce Type: new
Abstract: Transformers have become a central architecture for in-context learning (ICL), particularly through their state-of-the-art performance in large languag...
By Jaehee Seo, Jisu Kim
The paper introduces a Nested Inductive Bias framework that uses a two‑stage diffeomorphic composition to pull back non‑Euclidean target geometries onto symmetric positive definite (SPD) manifolds. This approach allows the construction of curvature‑aligned Riemannian classifiers that respect both matrix constraints and the intrinsic relational geometry of data. Empirical results on kinematic, signal processing, and synthetic benchmarks show that class separability degrades when metric curvature does not match the data distribution, and the authors also propose the Rational Conformal Metric (RCM) for robust vectorized architectures.
By Tushar Das
Methods operating on Vision Transformer (ViT) feature spaces typically rely on Euclidean distance or cosine similarity. This assumes that every direction is equally meaningful, but there is no reason...
arXiv:2607. 25299v1 Announce Type: cross Abstract: Optimization over the Stiefel manifold plays a significant role in various machine learning tasks.
By Yuan Zhang, Jiang Hu, Zhijian Lai, Lin Lin, Zaiwen Wen
arXiv:2609.27988v1 Announce Type: cross
Abstract: Methods operating on Vision Transformer (ViT) feature spaces typically rely on Euclidean distance or cosine similarity. This assumes that every direc...
By Andrew Bond, Ege Erdem \"Ozl\"u, Tuna \c{C}imen, Ilkin Umut Melanlioglu, Tolga Birdal, Erkut Erdem, Aykut Erdem
The paper introduces a geometry-aware Bayesian fine‑tuning method that uses Stein variational gradient descent on the Stiefel manifold. By transporting low‑rank adapter matrices along this manifold, the approach preserves orthogonality constraints and yields multiple inference solutions, enabling uncertainty quantification. Experiments demonstrate improved model calibration and higher prediction accuracy compared to Euclidean‑space SVGD and related methods.
By Quang-Duy Tran, Trung Le, Bao Duong, Phuoc Nguyen, Thin Nguyen
LoRA-TSD introduces a new optimizer for low‑rank adaptation (LoRA) that treats each update as a tangent vector on the fixed‑rank matrix manifold and applies a Muon‑style spectral‑norm steepest‑descent step within that tangent space. The method avoids costly full‑matrix operations and offers a retraction that is up to 2.8× cheaper than previous manifold approaches. The authors prove that their surrogate recovers LoRA‑Pro, identify the Riemannian gradient as the natural stationarity measure, and provide the first global convergence guarantees for both LoRA‑Pro and LoRA‑TSD, achieving superior performance across multiple benchmarks with Llama and Qwen models.
By Dmitrii Andriianov, Andrey Veprikov, Aleksandr Beznosikov
Curvature-Aware Radius Shrinkage for Adaptive Nearest Neighbor Classification (CARSANN) is a geometry-driven framework that adapts the spatial support of each neighborhood based on local geometric complexity. It estimates intrinsic dimensionality with TwoNN, builds an intrinsic representation via PCA, and uses a shape-operator-based estimate of local mean curvature to shrink the radius in highly curved regions while keeping a broader support in flatter areas. Experiments on over 70 OpenML datasets show that CARSANN consistently outperforms standard k‑NN and rivals other adaptive nearest‑neighbor methods, achieving a mean balanced accuracy increase from 0.6506 to 0.7528 and statistically significant improvements on most datasets.
By Alexandre L. M. Levada