v2vk(Vevtor2Vector Kit), annoy server (or faiss server in future).

Project on extending a Neural Partitioner for the M149 - Database Systems course, NKUA, Spring 2023.

Introduction to Approximate Nearest Neighbor Search (ANN) for Information Retrieval

Implementation and analysis of various algorithms, libraries and systems, distributed and not, for Approximate Nearest Neighbors searches

Image and text search_engine using elasticsearch and and Approximate nearest neighbors.

The engine for Sembed written in C++. Sembed is a vector database made for small projects and dosent require a ton of memory to work

A Python implementation of Approximate Nearest Neighbors algorithm with triangle inequality from scratch using numpy

Approximate k-Nearest Neighbours in high-dimensional space via Locality Sensitive Hashing

Embedding based semantic search app for poetry [App and EDA notebooks]

Autoencoder dimensionality reduction, EMD-Manhattan metrics comparison and classifier based clustering on MNIST dataset.

Simple Image Searching on CIFAR10 dataset using Conv Autoencoder

This project combines React, Express, OpenAI's CLIP model, and Spotify's ANNOY for searching an uploaded directory with text and image queries.

Testing the effectiveness of Approximate k-NN search with LSH and Hypercube on MNIST. Also, implemented k-medians++ for the same dataset.

Simple and Efficient DiskANN implementation

A recommendation system leveraging similarity measures to generate top-N item recommendations for various applications..

Search Engine that can recognize the species of a bird from an image uploaded by the user, proposing also similar pictures. Our own version of the PP-Index has been implemented to enable fast similarity search (developed for the Multimedia Information Retrieval and Computer Vision course)

End-to-end content-based fashion product recommendation system.