nripstein / Learning-Tensorflow

Notes from my process of learning tensorflow from Daniel Bourke's 64-hour tensorflow course

Tensorflor Tutorial

These are my Jupyter Notebook files as I learn tensorflow using Daniel Bourke's Tensorflow Youtube tutorial. It is a two part series over 14 hours in duration. After the free content on youtube, I began taking notes on the full 64 hour course.

This repository also includes some indipendent practice with tensorflow (see "notebook practice"), including a neural network I designed to predict house prices using the Boston housing dataset.

Why do I want to learn tensorflow?

       As a psychology, neuroscience and behaviour student, I am interested in understanding the underlying mechanisms of human cognition and behavior. Machine learning and artificial intelligence offer powerful tools for analyzing complex data and making predictions, which can be applied to a wide range of research questions in psychology and neuroscience. TensorFlow, in particular, is a popular and widely-used machine learning library that offers many tools for working with large datasets, building predictive models, and analyzing patterns in data. By learning TensorFlow, I hope to gain the skills necessary to work with large datasets and build models that can help us better understand the brain and behavior. I also hope to see how I can apply these skills to real-world problems, such as developing better diagnostic tools for mental health disorders and building more effective interventions for people with cognitive and behavioral challenges. Ultimately, I believe that learning TensorFlow will give me a powerful set of tools for conducting research and developing innovative solutions to complex problems in psychology and neuroscience.

What is Tensorflow?

       TensorFlow is an open-source library for numerical computation that allows developers to build and train machine learning models. It was developed by Google and is widely used for a variety of applications, from computer vision to natural language processing. TensorFlow makes it relativley easy to create complex models using high-level APIs, and can run on a variety of platforms, from CPUs to GPUs to mobile devices. It also includes tools for distributed training, which allows for large models to be trained more quickly. TensorFlow is one of the most popular machine learning libraries available today and is used by many companies and researchers to build cutting-edge machine learning applications.

What can be done with Tensorflow?

1. Image recognition: TensorFlow can be used to build computer vision applications such as image recognition, object detection, and image segmentation.
2. Natural language processing: TensorFlow has many tools for working with text data, including pre-processing, text classification, and sequence-to-sequence models for translation and summarization.
3. Speech recognition: TensorFlow includes tools for building speech recognition models, including pre-processing, feature extraction, and deep learning models.
4. Recommender systems: TensorFlow can be used to build personalized recommendation systems based on user data and preferences.
5. Anomaly detection: TensorFlow can be used to build models that can identify anomalies in time series data, such as fraud detection in financial transactions.
6. Generative models: TensorFlow can be used to build generative models such as GANs, which can create new images, music, and other types of data.
7. Reinforcement learning: TensorFlow includes tools for building reinforcement learning models, which can learn how to make decisions based on feedback from their environment.