A simple library for t-SNE animation and a zoom-in feature to apply t-SNE in that region

Reconstructing the topology of a metric graph

Showcasing Manifold Learning with ISOMAP, and compare the model to other transformations, such as PCA and MDS.

Manifold mapping with ISOMAP (MATLAB).

This project aims to compare the performance obtained using a linear Support Vector Machine model whose data was first processed through a Shortest Path kernel with the same SVM, this time with data also processed by two alternative Manifold Learning techniques: Isomap and Spectral Embedding.

The goal here is to use a graph kernel and a manifold learning technique in conjunction with Support Vector Machines to enhance the SVM classification.

Multi-omics image alignment and analysis by information manifolds (MIAAIM)

Dimensionality reduction and data embedding via PCA, MDS, and Isomap.

Pipeline Consisting of LSTM + Variational and Transformer Based Autoencoders + PCA/UMAP (Parameterized and Non-Parameterized) For Generating Low-Dim Manifold Representation of V1 Neural Activity

Filling the 3D 'scattering volume' by appropriately-oriented 2D scattering patterns. An analytical model suggests a numerical procedure (using Diffusion Map and the fisrt 9 non-trivial eigenvectors). The Matlab code here 1) synthesizes 2D scattering patterns; 2) Forms the Distance Matrix of mimages; and 3) retrieves the (relative) orientations u…

Implementation of the method for Density Estimation on an Unknown Submanifold from the paper https://arxiv.org/pdf/1910.08477.pdf

Materiales del Curso Aprendizaje Geométrico Profundo, Posgrado Matemáticas UNAM 2023-1

Coursera Applied Machine Learning in Python

This package is for Baxter and WHILL cooperative movement to perform robotic clothing assistance.

High-dimensional image preparation module for MIAAIM

Implements the principal flow algorithm by Professor Yao Zhi Gang using a greedy approach. Includes results of algorithm on various image datasets. Written to fulfill the requirements of my honors thesis in Statistics at the National University of Singapore.

Embedding with the Heat-geodesic dissimilarity

A model-based, unsupervised manifold learning method that factors complex cellular trajectories into interpretable bifurcating Gaussian processes of transcription.

Two ensemble models made from ensembles of LightGBM and CNN for a multiclass classification problem.