arXiv research

This work learns low-rank hyperbolic embeddings for tasks with hierarchical structures.

New algorithm tackles dynamic query routing to multiple embedding models.

As opposed to manual feature engineering which is tedious and difficult to scale, network representation learning has attracted a surge of research interests as it automates the process of feature learning on graphs. The learned low-dimensional node vector representation is generalizable and eases the knowledge discove…

Predict missing movie ratings or graph embeddings with low rank matrices.

Proposes a method to learn from multiple views with low-rank embeddings.

Estimates low-rank distributional matrices from incomplete samples.

This paper addresses the problem of low-rank distance matrix completion. This problem amounts to recover the missing entries of a distance matrix when the dimension of the data embedding space is possibly unknown but small compared to the number of considered data points. The focus is on high-dimensional problems. We r…

Although the convolutional neural networks (CNNs) have become popular for various image processing and computer vision task recently, it remains a challenging problem to reduce the storage cost of the parameters for resource-limited platforms. In the previous studies, tensor decomposition (TD) has achieved promising co…

Paper analyzes word embedding composition using tensor decomposition.

Proposes a faster Isomap algorithm by reducing eigenvalue decomposition complexity.

Framework for incomplete multi-view learning improves efficiency and clustering accuracy.

Attention improves edge prediction in e-commerce graphs.

New approach reduces model size for Transformer architectures.

The objective in extreme multi-label learning is to train a classifier that can automatically tag a novel data point with the most relevant subset of labels from an extremely large label set. Embedding based approaches make training and prediction tractable by assuming that the training label matrix is low-rank and hen…

Most existing word embedding methods can be categorized into Neural Embedding Models and Matrix Factorization (MF)-based methods. However some models are opaque to probabilistic interpretation, and MF-based methods, typically solved using Singular Value Decomposition (SVD), may incur loss of corpus information. In addi…

The problem of low rank matrix completion is considered in this paper. To exploit the underlying low-rank structure of the data matrix, we propose a hierarchical Gaussian prior model, where columns of the low-rank matrix are assumed to follow a Gaussian distribution with zero mean and a common precision matrix, and a W…

Proposes BHT-ARIMA for forecasting multiple short time series.

New embeddings capture local structure in complex networks.

We propose and evaluate new techniques for compressing and speeding up dense matrix multiplications as found in the fully connected and recurrent layers of neural networks for embedded large vocabulary continuous speech recognition (LVCSR). For compression, we introduce and study a trace norm regularization technique f…

Model compression is essential for serving large deep neural nets on devices with limited resources or applications that require real-time responses. As a case study, a state-of-the-art neural language model usually consists of one or more recurrent layers sandwiched between an embedding layer used for representing inp…

The paper tackles transfer learning for growing matrix representations, improving estimation accuracy.

New research shows graph embeddings fail to capture key network properties.

New faster, space-saving methods for subspace embeddings in tensors.

Proposes MvLPE for better multi-view representation learning.

CMF is a technique for simultaneously learning low-rank representations based on a collection of matrices with shared entities. A typical example is the joint modeling of user-item, item-property, and user-feature matrices in a recommender system. The key idea in CMF is that the embeddings are shared across the matrice…

ALF reduces network parameters and operations by 70% and 61%, respectively, on embedded hardware.

Spectral embedding improves with edge weight transformations.

HERA improves PLL by integrating heterogeneous loss and sparse-low-rank regularization.

Paper proposes a new word embedding method optimizing word similarity.

Low-rank matrix recovery has found many applications in science and engineering such as machine learning, signal processing, collaborative filtering, system identification, and Euclidean embedding. But the low-rank matrix recovery problem is an NP hard problem and thus challenging. A commonly used heuristic approach is…

LoRA enhances model adaptability without increasing parameters.

Deep learning models have become state of the art for natural language processing (NLP) tasks, however deploying these models in production system poses significant memory constraints. Existing compression methods are either lossy or introduce significant latency. We propose a compression method that leverages low rank…

Paper introduces G-LowTESTR for efficient tensor bandits.

Proposes a low-rank bilinear pooling model for link prediction in knowledge graphs.

Anchor PCA improves robustness in multi-domain PCA.

Qwen3-8B outperforms classical models in financial text classification.

StatFEM uses low-rank approximations to scale Bayesian statFEM for high-dimensional problems.

Improved LoRA+ adapts large models more efficiently.

The study categorizes knowledge graph relations and explains their embedding representations.

A new method models user-specific parameters as a low-rank plus sparse component for efficient personalization.

Matrix completion works well for smooth non-linear structures, even without low-rank assumptions.

The paper connects neural collapse and low-rank bias in networks with L2 regularization.

Algorithm compresses large matrices by approximating them as low rank and low precision factors.

New method accelerates CNNs for mobile devices by approximating tensors and quantizing weights.

Develops novel techniques for collaborative filtering and multi-label classification.

The matrix completion problem consists of finding or approximating a low-rank matrix based on a few samples of this matrix. We propose a new algorithm for matrix completion that minimizes the least-square distance on the sampling set over the Riemannian manifold of fixed-rank matrices. The algorithm is an adaptation of…

This paper improves spectral embedding for multipartite networks, revealing latent subspaces and providing consistent node representations.

Unified model for signed networks separates balance and anomaly effects.