Ole Reissmann’s Post

AI That Plays Flappy Bird Using NEAT! I'm thrilled to share a project I developed: an AI that learns to beat Flappy Bird using a genetic algorithm. Technologies Used: - Pygame: For game display. - NEAT-Python: A genetic algorithm library to evolve strategies for playing the game. How It Works: The AI learns through trial and error, trying to keep the bird airborne while avoiding pipes. It relies on three key pieces of data: 1. Bird's height   2. Upper pipe's height   3. Lower pipe's height  Using this data, the AI decides when to jump while gravity pulls the bird down. Evolutionary Approach: With NEAT, the AI improves by simulating multiple birds each generation, retaining the best performers, and refining their behaviors until it masters the game. Check out the project on GitHub: https://lnkd.in/g7KBpkwT Watch a demo of the project: https://lnkd.in/gExmrTkv #AI #MachineLearning #NEAT #Python #FlappyBird #GeneticAlgorithm #GameDev

Anthropic has recently released the Claude Sonnet 3.5 API, allowing AI models to operate computer applications like a humans. It sounds cool but use it cautiously, as this technology is still in the experimental phase. Let’s take a look at what makes this new API so remarkable. Claude Sonnet 3.5 can now call three new tools through its API: Computer, Text Editor, and Bash. This means it can: - Set computer screen resolution - Access the keyboard, mouse, and applications - Write Python scripts in a text editor - Run command-line programs in the Bash terminal - Store outputs in spreadsheets #AI #Claude #Sonnet #Anthropic #agent #AGI #AIGC

A smart and simple way to explain the attention mechanism .

The UK AI Safety Institute has launched Inspect - a framework for large Language model evaluations..... The press release (https://lnkd.in/eFew4A5Q) states "New UK-built AI safety testing platform to strengthen and accelerate global safety evaluations. Inspect will make it easier for a wide range of groups to develop AI evaluations, boosting collaboration with researchers and developers." "Inspect provides many built-in components, including facilities for prompt engineering, tool usage, multi-turn dialog, and model graded evaluations. Extensions to Inspect (e.g. to support new elicitation and scoring techniques) can be provided by other Python packages" Access the Inspect platform --> https://lnkd.in/eXFaRDMC Manchester Met University Research - Department of Computing & Mathematics

🎉 Exciting News! 🎉 I am thrilled to share that I received an A in Georgia Tech's OMSCS course "Machine Learning for Trading"! This experience has significantly enhanced my skills in machine learning, finance, and Python. Grateful for the opportunity to learn and grow in such an innovative program. #MachineLearning #AI #Trading #GeorgiaTech #OMSCS #LifelongLearning --- This post and silly image brought to you by ChatGPT 4o. 🤣

Understanding the mathematical expression in machine learning has being my target due to my background in mathematics.I've been diving into the fascinating world of gradient descent, a cornerstone algorithm in machine learning. This powerful technique iteratively adjusts model parameters to minimize a cost function, leading to improved model performance. In my latest pursuit for truly understanding the rhythm of the descent itself, I've visualized the gradient descent process step-by-step, unraveling the mathematical intuition behind it. I've also experimented with different learning rates and explored how they impact convergence. While the core concept is relatively straightforward, implementing it efficiently can be challenging. I've been working on creating a simplified Python function that encapsulates the key steps, making it easier to understand and apply in various machine learning tasks. #machinelearning #mathematics #appliedMathematics #datascience

I just found a great document that explains the self-attention mechanism in a really simple and clear way. If you want to understand self-attention better, this is a must-read.

Self-attention as a directed graph! Self-attention is at the heart of transformers, the architecture that led to the LLM revolution that we see today. In this post, I'll clearly explain self-attention & how it can be thought of as a directed graph. Keep Swiping through the images below!👇 For those interested in Ai Engineering, I would also encourage you to check Lightning AI's ⚡️ LLM learning Lab. A curated collection of blogs, tutorials, and how-to videos on: - Training - Fine-tuning - And deploying LLMs 🚀 Check this out👇  https://lnkd.in/d-X-gcCe _____________ That's a wrap, hope you enjoyed reading! If you're interested in: • LLMs 🧠 • MLOps 🛠 • Python 🐍 • And Machine Learning ⚙️ Find me → https://lnkd.in/em_V4unu ✔️ Everyday, I share tutorials on above topics! _____________ Cheers!! 🙂 #artificialintelligence #llms #machinelearning

Synthetic data can beat its teacher! The AI-MO team released their winning dataset with an additional fine-tuned Qwen2 model that approaches or surpasses OpenAI GPT-4o and Anthropic Claude 3.5 in match competitions. 👀 There was a sentiment that fine-tuned models from synthetic datasets cannot beat their teachers. Well, they can! NuminaMath 72B TIR matches GPT-4o and Anthropic Claude 3.5 on AMC 2023 and AIME 2024 with TIR. Tool Integrated Reasoning (TIR) first generates step-by-step explanations to solve the problem, followed by Python code, runs the code, and fixes any errors to solve math problems. Dataset & Collection: https://lnkd.in/ejKtryRq Dataset Tech report: https://lnkd.in/eH85Vgi4 Blog: https://lnkd.in/eh98BdPZ

The core of #MachineLearning (#ML) is #optimization.   The core of #optimization is #GradientDescent.   The core of #GradientDescent is:   - Writing the #lossfunction as #matrices   - Taking #derivatives of those matrices with respect to #parameters.   If you really want to understand #MachineLearning, #LinearAlgebra and #VectorCalculus are the most important tools.   That #LinearAlgebra lecture where you were taught to express #summations in terms of #matrices will help you master the #closedform ordinary least squares solution.   That #VectorCalculus lecture where you learnt how to take #derivatives of one #Matrix with respect to another #Matrix will help you master #Backpropagation and #NeuralNetworks.   #DataScience #AI #DeepLearning #Statistics #Mathematics #Tech #Education #Learning #Algorithms #ComputationalMathematics #Research #Innovation #ArtificialIntelligence #DataAnalysis #PredictiveModeling #BigData #DataMining #ScientificComputing #MathematicalOptimization #ModelTraining #FeatureEngineering #Programming #Python #RProgramming #MachineLearningAlgorithms #NeuralNetworkArchitecture #ContinuousLearning #STEM #QuantitativeAnalysis #DataVisualization #TechTrends