non-negative matrix factorization
PulseAugur coverage of non-negative matrix factorization — every cluster mentioning non-negative matrix factorization across labs, papers, and developer communities, ranked by signal.
3 day(s) with sentiment data
-
Dissertation details AI for structured knowledge retrieval and reasoning
This dissertation introduces a novel architecture for transforming unstructured domain-specific text into structured knowledge for retrieval and reasoning. It presents Binary Bleed, an adapted binary search method for N…
-
New SCALPEL method enables selective AI model unlearning
Researchers have developed SCALPEL, a novel contrastive sparse autoencoder designed for selective machine unlearning. This method aims to remove specific information from AI models, such as personal data under GDPR, whi…
-
Research paper links NMF, LCA, EMA, and PLSA models
A new research paper published on arXiv details the equivalences and identifiability of several compositional models used across machine learning, social science, and geology. The paper, titled "Nonnegative matrix facto…
-
New Diff Mining Framework Reveals Language Model Finetuning Objectives
Researchers have introduced Diff Mining, a novel framework designed to identify the specific objectives and behaviors learned by language models during the finetuning process. This method compares the logits of a finetu…
-
New MaxVol NMF method offers improved sparse decomposition
Researchers have introduced Maximum-Volume Nonnegative Matrix Factorization (MaxVol NMF) as an alternative to the existing Minimum-Volume NMF (MinVol NMF). While MinVol NMF aims to minimize the volume of its factor matr…
-
New NMF method enhances analysis of highly mixed grain-size data
Researchers have developed a new method called maximum-distance nonnegative matrix factorization (NMF) to improve the analysis of highly mixed grain-size distribution data. This technique generalizes existing NMF-based …
-
New framework analyzes team communication dynamics in VR using LLMs
Researchers have developed a computational framework to analyze team communication dynamics within collaborative virtual reality environments. This system uses late chunking and penalized Gaussian-kernel change-point de…
-
New framework tackles EV charging data issues with music-inspired approach
Researchers have developed the Note-Chord-Voice framework, a novel pipeline inspired by music theory to address challenges in electric vehicle (EV) charging data. This framework separates data cleaning, structural patte…
-
Google Maps POIs used to estimate income in Sao Paulo
Researchers have developed a method to estimate household income at a sub-municipal level in São Paulo, Brazil, by analyzing crowd-sourced data from Google Maps Points of Interest (POIs). This approach uses POI categori…
-
AutoML pipeline automates trend prediction from text data
This paper introduces AutoCluster, AutoTopicModeling, and AutoTrendAnalysis, a comprehensive AutoML pipeline designed to predict emerging trends from textual data with temporal attributes. The system automates the selec…
-
New MemNMF method enhances anomalous sound detection using LPC spectra
Researchers have developed MemNMF, a novel method for anomalous sound detection that operates on Linear Predictive Coding (LPC) spectra. This approach utilizes a memory module initialized from a non-negative matrix fact…
-
New R package 'nnmf' offers performance comparison for non-negative matrix factorization
A new R package named nnmf has been developed for non-negative matrix factorization (NMF), a technique used for dimensionality reduction across various fields like bioinformatics, text mining, and image analysis. This s…
-
New Newton Algorithm Enhances Nonnegative Matrix Factorization with KL Divergence · 2 sources tracked
Researchers have developed a novel Newton-type algorithm for Nonnegative Matrix Factorization (NMF) that utilizes the Kullback-Leibler (KL) divergence. This new method offers an efficient approach for analyzing count da…
-
New framework offers faithful, named explanations for AI classifiers
Researchers have introduced Language-Anchored Decomposition (LAD), a novel post-hoc framework designed to provide faithful and human-interpretable explanations for deep neural network classifiers without altering the or…
-
New convex model advances Nonnegative Matrix Factorization research
Researchers have developed a new convex model for Smooth Separable Nonnegative Matrix Factorization (SSNMF), a technique used for dimensionality reduction in nonnegative data. This model aims to address the challenges o…
-
New Graph-based Model Enhances Visual Explanation Interpretability
Researchers have developed a Graph-based Concept Bottleneck Model (G-CBM) that enhances interpretability in visual explanations. This new framework performs unsupervised concept discovery using Non-negative Matrix Facto…
-
AI models identify radioisotopes using computer vision techniques
Researchers have developed a novel machine learning approach for identifying radioisotopes in urban environments, converting gamma-ray data into spectrograms for analysis by computer vision architectures. This method en…
-
New EventNMF model analyzes continuous-time event data directly
Researchers have developed EventNMF, a novel continuous-time non-negative matrix factorization model designed to analyze event data directly. Unlike previous methods that require binning or smoothing, EventNMF operates …