Click on the above “Beginner Learning Vision” to choose to add Star or Pin.
Significant content delivered at the first time
Source: Coggle Data Science
The field of machine learning is vast. To avoid getting lost in learning, the following list can help guide your studies. It outlines some learning details of deep learning.
Stage 1: Beginner Level
At the beginner level, you should master the following skills:
- Able to handle small datasets
- Understand key concepts of classical machine learning techniques
- Understand classical networks DNN, CNN, and RNN
Data processing at the beginner level uses small datasets that can fit into main memory. Only a few lines of code are needed to apply such operations. At this stage, data types include Audio, Image, Time-series, and Text.
Classical machine learning is a good choice to study basic machine learning techniques before delving deeper into deep learning, including regression, clustering, SVM, and tree models.
Network: Master common network layers and corresponding neural networks; GAN, AE, VAE, DNN, CNN, RNN, etc. At the beginner stage, prioritize mastering DNN, CNN, and RNN.
Theory: Without neural networks, there is no deep learning. Without (mathematical) theory, there are no neural networks. Start learning by understanding mathematical symbols, beginning with matrices, linear algebra, and probability theory.
Stage 2: Intermediate Level
There is no real boundary between the intermediate and beginner levels. At the intermediate level, you can handle larger datasets and use advanced networks to process custom models:
- Handle larger datasets
- Customize models to complete tasks
- Improve network model accuracy
Data processing can handle datasets of several GB, requiring custom data augmentation methods and data processing functions.
Completing tasks independently means developing code based on specific tasks rather than following the MNIST tutorial.
Custom networks: How to process data when handling custom projects? How to define your network layers?
Model training: Master the concept of transfer learning and learn to use pre-trained weights to accomplish new tasks. Also, master the method of freezing certain network layers.
Deep learning theory: Master the forward and backward propagation of deep learning models, especially the chain rule for derivatives. Understand the roles of activation functions and objective functions, and be able to select appropriate activation functions and objective functions.
Stage 3: Proficient Level
Compared to the intermediate level, you need to master more data processing methods and methods to accelerate model training:
- Processing and storing large-scale data
- Tuning network models
- Unsupervised learning and reinforcement learning
Data processing requires mastering the handling of datasets of hundreds of GB and learning Linux operations. At this stage, you may encounter multimodal tasks.
Unsupervised projects: Start attempting to build unsupervised network models, such as autoencoders and GAN models, and master the principles of these models.
Model training: Master tuning methods and common logging and visualization tools, such as TensorBoard. Master learning rate adjustment methods, such as cosine annealing. Master multi-machine and mixed-precision training.
Stage 4: Expert Level
Master the development of cutting-edge academic models, know your interests, and be able to propose new models:
- Learn to use JAX or DALI for data processing
- Familiarize yourself with graph neural networks and transformer models
This article has been condensed from the original. Original link: https://towardsdatascience.com/a-guide-to-the-field-of-deep-learning-9bb9b21dae2
Disclaimer: Some content is sourced from the internet for readers' learning and exchange purposes. Article copyright belongs to the original author. If there are any issues, please contact for deletion.
Download 1: OpenCV-Contrib Extension Module Chinese Version Tutorial
Reply "Extension Module Chinese Tutorial" in the backend of the "Beginner Learning Vision" public account to download the first OpenCV extension module tutorial in Chinese, covering installation of extension modules, SFM algorithms, stereo vision, object tracking, biological vision, super-resolution processing, and more than twenty chapters of content.
Download 2: Python Visual Practical Projects 52 Lectures
Reply "Python Visual Practical Projects" in the backend of the "Beginner Learning Vision" public account to download 31 visual practical projects, including image segmentation, mask detection, lane line detection, vehicle counting, eyeliner addition, license plate recognition, character recognition, emotion detection, text content extraction, facial recognition, etc., to help quickly learn computer vision.
Download 3: OpenCV Practical Projects 20 Lectures
Reply "OpenCV Practical Projects 20 Lectures" in the backend of the "Beginner Learning Vision" public account to download 20 practical projects based on OpenCV, achieving advanced OpenCV learning.
Group Chat
Welcome to join the public account reader group to exchange with peers. Currently, there are WeChat groups for SLAM, 3D vision, sensors, autonomous driving, computational photography, detection, segmentation, recognition, medical imaging, GAN, algorithm competitions, etc. (These will gradually be subdivided). Please scan the WeChat ID below to join the group, and note: "Nickname + School/Company + Research Direction", for example: "Zhang San + Shanghai Jiao Tong University + Visual SLAM". Please follow the format; otherwise, you will not be approved. After successful addition, you will be invited to the relevant WeChat group based on your research direction. Please do not send advertisements in the group; otherwise, you will be removed. Thank you for your understanding~