← Back to all news
arXiv AI By Yuanzhe Peng, Jie Xu

Collaborative In-Context Learning for Indoor Radio Map Estimation

Read the original on arXiv AI →

The paper introduces Poly‑Ripple, a collaborative in‑context learning framework for indoor radio map estimation that leverages large autoregressive vision models without updating them. Clients provide private demonstrations to predict radio maps for shared scene queries, using ripple‑ordered autoregressive modeling to reflect wireless propagation from the transmitter. Importance‑aware aggregation combines heterogeneous client knowledge within context length limits, and experiments on a public benchmark show Poly‑Ripple outperforms existing ICL baselines.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv AI.

Related stories

arXiv Machine Learning
Jun 11

OmniLoc: A Geometry-Aware Foundation Model for Anchor-Free UE Localization Across Diverse Indoor Environments

arXiv:2606. 11490v1 Announce Type: new Abstract: Indoor localization from wireless measurements remains challenging in large-scale deployments due to substantial variation in building geometry, the set of detectable access points (APs), and the heterogeneity of received signals.

By Lei Chu, Yuning Zhang, Omer Gokalp Serbetci, Anushka Katiyar, Bassel Abou Ali Modad, Andreas F. Molisch
llmsragnlpbenchmarks
More like this →
arXiv Machine Learning
Oct 16

Scene Reconstruction Using Ubiquitous Wireless Signals via Diffusion Posterior Inference Over physics-Grounded Representations

The paper presents a method for reconstructing building footprints using only wireless channel measurements. It introduces a two‑stage hierarchical framework: first, a physics‑grounded wireless foundation model converts each channel measurement into a scene representation; second, a diffusion model jointly integrates multiple representations to infer and sample the posterior distribution over building occupancy. Experiments show that this approach can accurately recover building footprints and that both the physics‑grounded representations and joint posterior inference are essential for performance.

By Vahid Yazdnian, Yasaman Ghasempour
diffusion
More like this →
arXiv Machine Learning
Oct 22

A Transfer Learning-Driven Methodology for Efficient and Sustainable Indoor Localization Using VLC

The paper presents a transfer learning–driven approach for indoor localization using visible light communication (VLC) that aims to reduce energy consumption and improve scalability. By applying transfer learning, the method addresses environmental variability such as lighting changes and obstacles, lowering computational overhead compared to conventional techniques. Experiments on a BOSCH factory dataset show a 47% increase in localization accuracy and up to 67.5% model efficiency under high‑noise conditions, outperforming non‑TL models.

By Masood Jan, Wafa Njima, Xun Zhang, Alexander Artemenko
fine-tuning
More like this →
arXiv AI
Jun 15

Fusion of Pervasive RF Data with Spatial Images via Vision Transformers for Enhanced Mapping in Smart Cities

arXiv:2508. 03736v2 Announce Type: replace-cross Abstract: In this paper, we present a deep learning-based approach that integrates the DINOv2 architecture to improve building mapping by combining (possibly erroneous) maps from open-source platforms with pervasive radio frequency (RF) data collected from multiple wireless user equipments and base stations.

By Rafayel Mkrtchyan, Armen Manukyan, Hrant Khachatrian, Theofanis P. Raptis
llmscomputer-vision
More like this →