Informationen zur Anzeige:
Master Thesis - Detecting and Eliminating Adversarial Attacks using XAI - REF73153D
Berlin
Aktualität: 09.01.2025
Anzeigeninhalt:
09.01.2025, Continental AG
Berlin
Master Thesis - Detecting and Eliminating Adversarial Attacks using XAI - REF73153D
Über uns:
Continental develops pioneering technologies and services for sustainable and connected mobility of people and their goods. Founded in 1871, the technology company offers safe, efficient, intelligent and affordable solutions for vehicles, machines, traffic and transportation. In 2023, Continental generated sales of 41.4 billion and currently employs around 200,000 people in 56 countries and markets.
Aufgaben:
- Adversarial attacks challenge the reliability of machine learning models, particularly in computer vision tasks. These attacks exploit vulnerabilities in neural networks to mislead their outputs. This project focuses on detecting such attacks by analyzing neural network internal representations and latent spaces , employing techniques from Explainable AI (XAI) and data science.
- You will contribute to innovative research aimed at understanding and mitigating adversarial behavior in computer vision deep learning models, emphasizing robust and real-time detection methodologies.
- Explore how adversarial attacks manifest in neural networks' latent spaces and internal activations
- Investigate techniques to detect adversarial behavior using insights from XAI and advanced machine learning methods
- Develop, test and refine approaches to improve the resilience of models against adversarial scenarios
Qualifikationen:
- You study in the field of Artifical Intelligence, Computer Science, Computer Engineering or comparable courses of study, so that you have basic aspects of our area of responsibility
- Proficiency Programming experience in Python and deep learning frameworks like PyTorch are absolutely necessary in the context of this position
- You should also have an understanding of deep neural networks, machine learning, and computer vision, should be familiar with adversarial attacks and defenses in machine learning and have basic knowledge of Explainable AI concepts and their applications
- We expect as well the ability to analyze complex data and generate insightful visualizations
- Very good English language skills in word and writing are absolutely necessary for us
- You are independent, goal-oriented and structured and know how to work effectively without losing track - always pay attention to high-quality results
- Analytical thinking is also beneficial and can help you solve the tasks effectively
- A very good communication and teamwork skills are particularly important for us - after all, we always have to work together on solutions
Wir bieten:
- You will receive qualified support from our experienced specialists to improve your professional knowledge and practical skills
- Your commitment to us is fairly remunerated : We know that your work is valuable and want to ensure that you are adequately remunerated
- We offer you flexible working hours with hybrid working models, where you have the opportunity to occasionally work from home, because we know how important the interaction of private and professional life is.
- Our location is c entrally located in Berlin and we look forward to welcoming you to our Co-Working Space AI Campus
- We provide a powerful GPU infrastructure , because only with excellent technology can we achieve great
- Participate in journal clubs and networking events and share technical topics with your colleagues on campus
Unser Kontakt:
Jana Janotta
Weitere Informationen:
Applications from severely handicapped people are welcome.
Standorte
