Google Gemma is a powerful neural network. While Gemma is primarily focused on research and advanced users, its neural network capabilities could be interesting to a wide range of users.

As the second year of Aggregata draws to a close, we take the opportunity to review the changes and provide some insight into future developments.

Transparency is an important step for user trust in a product. Today we want to focus more on our use of machine learning methods for Aggregata.

Automatically answering questions about images is a powerful tool for making a variety of processes faster and more efficient. In this article, we introduce Matcha-Quarta, a neural network trained for this task.

Image segmentation is crucial for object recognition and advanced image processing, and the Kosmos-2 neural network enhances this by grounding text in the visual world, perceiving object descriptions, and associating them with their respective image regions. Today we will explore use cases of such a model.

Actor-critic reinforcement learning is a significant advancement in the field of reinforcement learning. Actor-critic reinforcement learning combines the advantages of both policy-based and value-based reinforcement learning, allowing for more efficient and effective learning in complex environments. In this post, I would like to introduce this algorithm.

Dimension reduction is an increasingly important part of the learning process of machine learning programs as data sets grow larger. Today we will look at a linear transformation method on low dimensional vector spaces to potentially improve the learning process. Principal Component Analysis will be introduced for this purpose.

Image captioning is important because it provides a textual representation of the content and context of an image, improving accessibility and understanding for all users, especially those with visual impairments. In this post, we introduce BLIP for this use case.

Sentiment analysis is an increasingly important part of the evaluation of news from social networks. In the following article, we would like to present a pretrained transformer that is tailored for this task: the Emotion Text Classifier.

The Naive Bayes classifier is a simple and efficient algorithm for classification tasks that assumes independence of the features. This post aims to introduce this algorithm to the reader.

Large amounts of data can be challenging for many reasons. Today, I will present an algorithm that can be used to reduce the size of a data set: Random Projections.

T5 is a powerful language model capable of performing a wide range of text-to-text tasks, including text classification, language translation or text summarization. The aim of this post is to introduce this pretrained transformer to the reader.

Optimizing a (complex) function can be a difficult task. Here I present a library which I use and which I think is a good way to solve such tasks. In addition, I will show corresponding tasks and an implementation of a comparatively simple optimization task as a usage example.

As in many other areas, ethical issues are necessary for the safe and fair use of various products. In this article, we will look at ethical issues related to machine learning.

The purpose of this article is to introduce the reader to the basic topic of reinforcement learning. The introduction concretizes important terms and basic concepts used there.

To identify similar groups in unknown data and reduce complexity, clustering can be used. Here we describe the k-means algorithm that can be used for clustering.

There are a number of ways to solve a regression task. In this article we will describe how such a regression task can be solved using a neural network.

There are several classical algorithms to perform a classification. Here I describe an implementation using neural networks and my experiences with it.

Q-Learning was one of the first practical reinforcement learning algorithms. This post introduces this algorithm.

Decision trees alone are often not sufficient to perform a meaningful classification. Random forests represent an improvement of the decision trees. I am going to present this method in the following post.

Classification can be done in many different ways and with many different algorithms. Today I will introduce the K-Nearest-Neighbour classifier.

Often machine learning models are complex and difficult to understand. This article describes a method that visualizes clear structures and decision criteria: Decision Trees.

Event occurrence probabilities are often a complex problem to model. In this article, a basic method for this is presented: Logistic Regression.

This post is intended to introduce the reader to the basic topic of reinforcement learning. The introduction clarifies important terms and basic concepts that are used in Reinforcement Learning.

Semi supervised learning represents one of the four topics of machine learning. This post is intended to give an introduction to the topic.

Unsupervised Learning represents one of the four topics of Machine Learning. This post aims to provide an introduction to the topic.

Supervised Learning represents one of the four topics of Machine Learning. This post is intended to provide an introduction to the topic.

Neural networks are a hotly debated topic. But what are they and how exactly do they work? Some of these questions are answered here.

Machine Learning is taking an increasingly important part in our lives. This post will explain why is likely a relevant concept in the future.

The Deep Q network sometimes suffers from various problems. The problems are presented here and a solution for these problems is presented.

Deep Q-Networks sometimes need information from different time steps to converge quickly. The Deep Recurrent Q-Network represents one possibility for this.

In the Big Data business, (co-)rellations are often useful to make decisions. Here is one easy method to find such information: Linear Regression.

Artificial neural networks have achieved success in many fields. Here I present the first algorithm for training neural networks for reinforcement learning: the Deep Q-Network.

Studies contained an Emphasis on Modeling of Aerospace Systems

Topic of the bachelors thesis was "Implementing the finite element method for quantum graphs".