Internship | Advancing federated learning for defense applications

Internship | Advancing Federated Learning for Defense Applications

What will be your role?

Introduction:

In conventional training of machine learning models, data is stored in one location. More data often means better performance, thus if multiple parties want to contribute to the training, they have to share their data. Nevertheless, when data is sensitive, data sharing is often not possible nor allowed. This is the case in defense applications, where AI models have to be trained on sensitive military data, which cannot be shared across different countries or organizations.

Federated Learning (FL) is a technique whereby an AI model can be collaboratively trained with data from multiple parties without having to upload this data to a central location. This technique has proven to be very effective and more secure than conventional training. Nevertheless, training in a FL setup has two downsides: it is slow due to the communication overhead and it is difficult to perform well on realistic data distributions.

Research on federated learning often focuses on situations where all participating parties in the FL setup have similar private datasets (“independent and identically distributed” a.k.a. “IID”), for example where the number of samples and classes in each client are uniform, and the images in each individual dataset are similar.

In real-world scenarios, this is often not the case. Here, data can be very unbalanced (“non-IID”). Research has shown that non-IID data can result in a performance drop, raise security risks, and increase the training time even more compared to training on IID data.

The aim of the thesis is to make FL work in a realistic scenario. For this, three objectives are defined:

• Investigate how to create datasets in each FL client that realistically mimic a possible real-world scenario.
• Explore strategies to account for the negative effects of a realistic non-IID data distribution.
• Investigate how training on realistic non-IID data can be more efficient, for example by applying less communication steps.

What will be your role?

Your role will be to explore the types and the effects of non-IID data on FL, and investigate the state-of-the-art strategies to account for effects of different types of non-IID data. You will start with a literature study on the types of non-IID data, and the effects that the different types of non-IID data have on the training of the model. It will especially be interesting to gain knowledge on the effects of a combination of non-IID data distribution types, since the more types that are combined, the more “realistic” the dataset becomes.

Your research could make a direct impact in the development of automatic threat detection systems using federated learning, where the goal is to identify and mitigate potential security risks. The proposed solution can enhance the accuracy and security of defense applications while also being more robust to real-world data distributions. This allows multiple defense organizations to collaboratively develop models without sharing sensitive data, thereby streamlining processes and protecting classified information.

You will perform this assignment within TNO’s Intelligent Imaging department. The Intelligent Imaging department is a passionate, creative, and dedicated team of professionals (60 people) specializing in developing groundbreaking applications in the field of computer vision. Our team members have diverse backgrounds, ranging from the medical field to artificial intelligence. Intelligent Imaging is a young and growing department that has built up a lot of expertise over the past years in AI and deep learning.

What we expect from you

We are looking for a master’s student who wants to join our cutting-edge research team to explore the boundaries of training FL models for computer vision tasks. This position is perfect for students who are passionate about AI, computer vision, and advanced machine learning techniques. You should be interested in taking a deep dive into the world of non-IID data in FL and helping us uncover how to handle training FL setups on datasets that mimic realistic real-world data distributions.

Additional requirements include being in the final stages of your master's degree in artificial intelligence, computer science, physics, mathematics, electrical engineering, or a similar field. You should have some experience in computer vision, artificial intelligence, deep learning and Python programming. Experience with federated learning or handling of non-IID data is not required, but basic understanding of training an AI model is.

What you'll get in return

You want an internship opportunity on the precursor of your career; an internship gives you an opportunity to take a good look at your prospective future employer. TNO goes a step further. It’s not just looking that interests us; you and your knowledge are essential to our innovation. That’s why we attach a great deal of value to your personal and professional development. You will, of course, be properly supervised during your work placement and be given the scope for you to get the best out of yourself. Furthermore, we provide:

• A highly professional, innovative internship environment, within a team of top experts.
• A suitable internship allowance (615 euro for wo-, hbo- and mbo-students, for a full-time internship).
• Possibility of eight hours of free leave per internship month (for a full-time internship).
• A free membership of Jong TNO, where you can meet other TNO professionals and join several activities, such as sports activities, (work-related) courses or the yearly ski-trip.
• Use of a laptop.
• An allowance for travel expenses in case you don’t receive an OV-card.

TNO as an employer

At TNO, we innovate for a healthier, safer and more sustainable life. And for a strong economy. Since 1932, we have been making knowledge and technology available for the common good. We find each other in wonder and ingenuity. We are driven to push boundaries. There is all the space and support for your talent and ambition. You work with people who will challenge you: who inspire you and want to learn from you. Our state-of-the-art facilities are there to realize your vision. What you do at TNO matters: impact makes the difference. Because with every innovation you contribute to tomorrow’s world.

At TNO we encourage an inclusive work environment, where you can be yourself. Whatever your story and whatever unique qualities you bring to the table. It is by combining our unique strengths and perspectives that we are able to develop innovations that make a real difference in society.