PhD: Detection of hidden information with AI

In recent years, the criminal use of information hiding techniques (i.e. steganography) in digital media such as images, video, audio and text files has increased rapidly. An important reason for this is that many steganographic tools are publicly available. As an example, the platform GitHub2 is currently reporting more than 4,800 steganography related results in its repository. Moreover, this multitude of tools is often made available as program source code packages. Consequently, perpetrators can easily and selectively pick, adapt and combine information hiding tools for their criminal activities. We are observing an increase in the use of steganographic methods in a wide variety of types of crime including child pornography, industrial espionage, criminal attacks on enterprises, credit card fraud, system intrusion, and backdoor injection and delivery methods.

The increasing complexity of security challenges combined with the accumulation of significant amounts of digital data calls for better and more widespread use of artificial intelligence (AI) capabilities for law enforcement agencies, also in relation to information hiding techniques. AI can provide benefits to forensic institutes on a number of levels given the right understanding, tools, data and protection while increasing the validation and making it explainable for court. You will work in the Forensic Image Analysis and Biometrics team of the Digital and Biometric Traces division. The team utilises data analysis to respond to forensic investigation queries about the use of evidence in legal proceedings, for example. You will work closely with the University of Amsterdam’s Informatics Institute (team Multix), where you will obtain your PhD. In addition, you will contribute to EU projects on Stego-analysis and the processing of large amounts of multimedia data.

Applicable research questions

• How can AI and other methods help detect the criminal use of information hiding techniques in images, video and audio?
• How can we make the methods used explainable in court and do it forensically correct with validation?
  
  

Key Responsibilities

• You complete an article based PhD-thesis within 4 years.
• You present your research in leading scientific journals and at international conferences.
• You develop software tools based on the research you conduct.
• You work in the Forensic Image Analysis and Biometrics team to validate the developed methods against data obtained in practice.
• You work at the Informatics Institute in Amsterdam one day a week (on average) to gain wider input and ideas for your research.
• You contribute to European project meetings.
  
  

Functie-eisen

• You have an MSc in computer science, artificial intelligence, mathematics, statistics or equivalent.
• You have experience in data science (i.e. machine learning, natural language processing, social network analysis, and visualisation).
• You are able to programme and have experience of programming languages such as Python.
• You enjoy working for a social cause and have an affinity with the security and justice sector.
• You communicate and write in excellent English and you are willing to learn Dutch.
• You are enthusiastic, curious and creative.
• You can work independently and as part of a team.

Competencies

• analysis
• accuracy of work
• collaboration
• planning and organising
• judgement