Ask or search anything...
Two-Scale Latent Dynamics for Recurrent-Depth Transformers
13 Nov 2025
Researchers at Sapienza University of Rome characterized the latent space dynamics of recurrent-depth transformers, revealing distinct small-scale refinements within looped blocks and larger-scale drifts across blocks. Leveraging this geometric understanding, they introduced an acceleration-based early-exit mechanism that reduced inference latency by up to 38% while preserving output quality, outperforming existing methods.
View blog
From Spectra to Geography: Intelligent Mapping of RRUFF Mineral Data
18 Nov 2024
Accurately determining the geographic origin of mineral samples is pivotal for applications in geology, mineralogy, and material science. Leveraging the comprehensive Raman spectral data from the RRUFF database, this study introduces a novel machine learning framework aimed at geolocating mineral specimens at the country level. We employ a one-dimensional ConvNeXt1D neural network architecture to classify mineral spectra based solely on their spectral signatures. The processed dataset comprises over 32,900 mineral samples, predominantly natural, spanning 101 countries. Through five-fold cross-validation, the ConvNeXt1D model achieved an impressive average classification accuracy of 93%, demonstrating its efficacy in capturing geospatial patterns inherent in Raman spectra.
There are no more papers matching your filters at the moment.