Euclidean Embeddings

Euclidean embeddings represent data points as vectors in a Euclidean space, aiming to capture relationships between data instances through their proximity. Current research focuses on improving the effectiveness of these embeddings in various applications, including few-shot learning, graph classification, and structured prediction, often comparing them to alternative approaches like hyperbolic embeddings. A key challenge lies in optimizing embedding methods for high-dimensional data and noisy inputs, with recent work exploring techniques like low-rank approximations, subsampling, and the incorporation of magnitude information to enhance performance. These advancements have implications for diverse fields, improving the accuracy and efficiency of machine learning models across numerous applications.

Papers

October 8, 2024

Weighted Embeddings for Low-Dimensional Graph Representation
Thomas Bläsius, Jean-Pierre von der Heydt, Maximilian Katzmann, Nikolai Maas
Jina Embeddings Hyperbolic Embeddings Weight Freezing Euclidean Embeddings

August 9, 2024

Hyperbolic Learning with Multimodal Large Language Models
Paolo Mandica, Luca Franco, Konstantinos Kallidromitis, Suzanne Petryk, Fabio Galasso
Vision Language Model Multimodal Large Language Model Hyperbolic Space Hyperbolic Embeddings Hyperbolic Model Euclidean Embeddings

September 18, 2023

Hyperbolic vs Euclidean Embeddings in Few-Shot Learning: Two Sides of the Same Coin
Gabriel Moreira, Manuel Marques, João Paulo Costeira, Alexander Hauptmann
LeArning Abstract Representation Learning Shot Classification Hyperbolic Space Spatial Encoder Hyperbolic Embeddings Hyperbolic Neural Network Euclidean Embeddings

November 24, 2022

Lifting Weak Supervision To Structured Prediction
Harit Vishwakarma, Nicholas Roberts, Frederic Sala
Pseudo Label Label Noise Weak Supervision Structured Prediction Euclidean Embeddings

October 20, 2022

Low-Rank Representations Towards Classification Problem of Complex Networks
Murat Çelik, Ali Baran Taşdemir, Lale Özkahya
Real World Meaningful Representation Complex Network Classification Problem LOw Rank Euclidean Embeddings Network Classification

September 1, 2022

Johnson-Lindenstrauss embeddings for noisy vectors -- taking advantage of the noise
Zhen Shao
Industrial Disturbing Noise Advantage Feedback Sparse Embeddings Euclidean Embeddings Dense Embeddings High Dimensional Noisy Johnson Lindenstrauss

March 14, 2022

Permutation Invariant Representations with Applications to Graph Deep Learning
Radu Balan, Naveed Haghani, Maneesh Singh
Financial Application Jina Embeddings Machine Permutation Euclidean Embeddings Permutation Invariant Representation

February 28, 2022

Magnitude-aware Probabilistic Speaker Embeddings
Nikita Kuzmin, Igor Fedorov, Alexey Sholokhov
Speech Recognition Speaker Verification Speaker Embeddings Hyperspherical Learning Euclidean Embeddings

January 22, 2022

Collaborative Representation for SPD Matrices with Application to Image-Set Classification
Li Chu, Rui Wang, Xiao-Jun Wu
Application Proficiency Classification Code Feature Space Invariant Representation Reproducing Kernel Hilbert Space Gradient Information Symmetric Positive Definite Laplacian Kernel Euclidean Embeddings

January 5, 2022

Asymptotics of $\ell_2$ Regularized Network Embeddings
Andrew Davison
Stochastic Gradient Descent Jina Embeddings Link Prediction Asymptotic Behavior Node2vec Application Euclidean Embeddings