Job Description

Adversarial samples became popular in the area of deep learning, where they have been defined as input samples subtly modified to cause a machine learning misclassification.
Generative Adversarial Networks (GAN) are a deep learning approach for generating artificial examples that plausibly could be drawn from a certain data category. GANs architecture is composed of two networks: a generator and a discriminator. The generator is learning a generative model close to the data model, in order to generate artificial data samples, while the discriminator compares the generated sample with the training data and computes the probability that it belongs to the training data. The objective of the discriminator is to force the generator to learn a good data representation during the adversarial process.

In the case of EO multi-spectral images, the adversarial samples may occur “naturally”. Sensor artifacts like LSB stripes or saturation are at the origin of such samples. In addition, EO multispectral data contains physical information. Thus, the adversarial information shall represent it in a consistent meaningful model. As a simple example is the transparency of a cloud to infrared radiation. This project aims to provide solutions for deep learning for EO multi-spectral images in the presence of naturally occurring adversarial samples and also considering their physical nature and models.

The successful candidate will be employed for 3 years by University Politehnica of Bucharest and will be part of CEOSpaceTech, a very dynamic research laboratory oriented towards space applications. She or he receives a financial package, which is twice the average salary in the country and additional mobility and family allowance, granted according to the rules for Early Stage Researchers (ESRs) in an EU Marie Sklodowska-Curie Actions Innovative Training Networks (ITN). A career development plan will be prepared for her/him in accordance with the supervisor. The plan will include a choice of more than 20 streamed or registered courses, one stage in Germany, and various outreach activities. For more information please visit the Marie Sklodowska-Curie Actions Innovative Training Networks website.

YOUR TASKS

• Study the concept of adversarial samples for EO multi-spectral images.
• Build of a database of specific adversarial samples and use GANs to generate them.
• Study of the effects of adversarial samples for the case of DNN applied to EO multi-spectral images and design of specific DNN paradigms to alleviate the sequels of adversarial samples.

PROFILE

• A Master of Science in Computer Science is required. It could comprise the full range of mathematical, physical, engineering and technology disciplines related to sensor data acquisition and programming.
• Proficient English level.
• Image analysis, neural networks, programming languages, basic knowledge on optics could be an advantage.

PLANNED SECONDMENTS

• DLR, Munich, Germany, Prof. M. Datcu, 10 months, theoretical aspects of DNN for EO multi-spectral images and related topics regarding the impact of adversarial samples.

ADDITIONAL INFORMATION

References

1. Tsagkatakis, Grigorios, et al. “Survey of Deep-Learning Approaches for Remote Sensing Observation Enhancement.” Sensors18 (2019): 3929.
2. Qiu, Shilin, et al. “Review of artificial intelligence adversarial attack and defense technologies.” Applied Sciences5 (2019): 909.
3. Czaja, Wojciech, et al. “Adversarial examples in remote sensing.” Proceedings of the 26th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems. ACM, 2018.
4. Making Machine Learning Robust Against Adversarial Inputs.

Follow @eurasip