A differentiable version of SPTK

Comprehensive topic modeling system using Non-negative Matrix Factorization (NMF)

A modular toolkit for NMF-based sound source localization

Python Implementation of DRNMF-SP

Official implementation of PRISM : Pruning for Rank-adaptive Interpretable Segmentation Model, ICCVW 2025. PRISM integrates NMF in a neural structure to automatically extracts interpretable components from multispectral images, enhancing text binarization of historical documents.

Python Implementation of the Instance-wise Distributionally Robust Nonnegative Matrix Factorization (iDRNMF)

Research Fellowship work developed at University of Bologna in Spring 2025. Modular structure for analysing various kind of synergy extraction methods (PCA, Autoencoders, Classical NMF, Sparse NMF) on three different hand-pose datasets.

Customer Retention Project

Pipeline complet de modélisation de sujets (LDA, NMF, BERTopic) avec application Streamlit interactive.

Bachelor Thesis work developed in 2025 at University of Bologna. See Readme for more information on the project.

Convert multivariate time series into evolutionary games; fit replicator dynamics and test for ESS (ts2eg).

Repository containing the codes used for the development of a source separation system for cardiorespiratory sounds using Non-negative Matrix Factorisation (NMF). The aim is to obtain a pure respiratory sound and a pure cardiac sound using this algorithm.

🎛️ [Audio] Source separation using Non-negative Matrix Factorization (NMF)

Pysoundlocalization aims to localize stationary and moving sound sources in polyphonic, noisy environments using sensors placed around the sound source(s). The library offers ready-for-use implementations to support sound source localization based on time difference of arrival (TDoA) and multilateration.

analyzing and detecting hot topics in blog data using various natural language processing and topic modeling techniques

libfmp - Python package for teaching and learning Fundamentals of Music Processing (FMP)

A flexible source separation library in Python

[ACM Computing Surveys'23] Implementations or refactor of some temporal link prediction/dynamic link prediction methods and summary of related open resources for survey paper "Temporal Link Prediction: A Unified Framework, Taxonomy, and Review" which has been accepted by ACM Computing Surveys.

A combine research process of radio astronomy, including line identification, spectrum fitting, dimension reduction, and clustering analysis.

A fast and versatile implementation of spike-triggered non-negative matrix factorization (STNMF) based on accelerated/fast HALS algorithms