Diffusion models have become the state-of-the-art technology for generative A.I. so we will investigate their capabilities by understanding their theoretical background and implementing their architecture.

We are going to explore a fragment of the Yelp social network - a website where both restaurants and users can post food-related activity - with the goal of finding groups of users with similar restaurant preferences.

This work addresses central mathematical issues in modern optimization efforts of wireless communications systems via Intelligent Reflecting Surfaces and offers bounds and algorithms designed to optimize these systems and work around the induced mathematical difficulties.

We want to study in-detail all functionalities of the skip-gram word2vec model and implement the architecture on a corpus of restaurant reviews, with the goal of creating an embedding space for words and testing the reliability of the generated space to complete analogies (‘a’ is to ‘b’ as ‘c’ is to ‘d’).

Although EEG/MEG signals have a notoriously low SNR, it is possible to remove much of the different sources of noise corrupting our target signals via our multi-stage processing pipeline, such that our neural network is able to optimally classify and execute further control based on these signals.

Often there are useful variables not directly visible from a data set which can be exploited, and sometimes there are variables which offer negligible gains. After analyzing our data, we want to derive the classic linear regression model and evaluate its robustness against other classic A.I. methods.

Where does the logistic regression model along with all of its components come from? Is it a robust model for decision making? We will investigate these questions, implement the model, and evaluate its capability to detect cancerous cells given physical tissue measurements.

In this article, we want to investigate whether an accurate “middle-ground” technique can be exist between unsupervised and supervised classification.