Deep Learning for visual recognition
Understand the details of deep learning for visual recognitionDel siden Print
Understanding Deep Learning
The deep learning revolution has given us self-driving cars, Google translate, Siri and much more.
Within the area of deep learning, visual recognition (or computer vision) is the most established field and additionally the best way to learn more about deep learning.
Throughout this two-day course, you will improve your skills by learning how to implement, train and debug your own neutral networks. You will also gain a comprehensive understanding of neutral network architectures and insight into cutting-edge research within deep learning.
The hands-on programming exercises on the course involves setting up computer vision problems, such as:
- Image classification
- Applying learning algorithms
- Practical engineering tricks for training as well as fine-tuning the networks
We will discuss how to apply deep learning for other data types, such as text, speech, and tabular data.
Who can participate?
The course is aimed at software developers or engineers, who wants to expand their toolbox with deep learning.
Participants will be reading and writing code in Python throughout the course. It is not necessary to be familiar with Python, but basic programming skills are required.
Participants does not need any prior knowledge regarding machine learning and only basic mathematics are required for the course.
Benefits for you:
- Identifying and describing visual recognition tasks that can be solved using deep learning
- Describing and comparing different neural networks architectures
- Explaining and relating techniques for training neural networks
- Applying deep learning to standard visual recognition tasks and assessing the results
- Defining and scoping your own deep learning projects
Benefits for your company:
- A competitive advantage by having employees with deep learning competences
- Improved ability to collect and exploit data in the future
- A higher chance of retaining employees with an interest in AI
The course covers deep learning theory, but primarily focuses on the practical use of the techniques. You will have hands-on training under expert instruction and the opportunity to ask any questions you might have under the entire duration of the course. Using the knowledge and implementing it in a practical context will allow you to continue to use deep learning in your daily work.
- Machine learning fundamentals
- Logistic and linear regression
- Simple neural networks
- Training neural networks (this will include topics such as optimization, transfer learning, backpropagation, and regularization techniques to avoid overfitting)
- Convolutional neural networks
- Advanced neural network architectures (this will include topics such as Recurrent Neural Networks, ResNet, and Generative Adversarial Networks)
Before the course
You should either have or be willing to create a Google (i.e., gmail) account, preferably prior to attending the course.
To benefit as much as possible from the course, make sure to brush up on the following areas in advance:
- basic linear algebra (such as inner products and matrix/vector multiplication)
- basic calculus (such as differentiation and especially the chain rule)
- simple probability theory (such as probability distributions)
During the course
You are expected to bring your own laptop.
There are no hardware requirements (like GPU), and there is no need for you to install any special software. All you will need is an internet browser (ideally Google Chrome).
After the course
After the course there will be no hand-ins, but you are more than welcome to consult with the instructor in case you need any help setting up your own projects.
The instructor on this course is Henrik Pedersen (PhD), Head of Visual Computing Lab at the Alexandra Institute. Henrik is lecturer at the Department of Computer Science at Aarhus University (AU), where he teaches the full-semester course, “Deep Learning for Visual Recognition”. Throughout his career, Henrik has been in various academic roles, both covering research as well as teaching in computer vision and deep learning. With years of experience, he is a skilled educator and has played a key role in creating the vibrant deep learning community at both the Alexandra Institute and AU’s computer science department.
Mannaz works in close collaboration with IDA, The Danish Society of Engineers.
When you sign up for this course, Mannaz handles your registration, while IDA manages the course execution.
Hvis du ønsker at komme på venteliste så udfyld felterne herunder. Vi kontakter dig hvis der kommer en ledig plads på kurset.