PyData Amsterdam 2023

Deep look into Deepfakes: Mastering Creation, Impact, and Detection
09-15, 15:20–15:50 (Europe/Amsterdam), Qux

Deepfakes, a form of synthetic media where a person's image or video is seamlessly replaced using Generative AI like GANs, have recieved significant attention. This talk aims to provide a comprehensive exploration of deepfakes, covering their creation process, positive and negative effects, development pace, and tools for detection. By the end of the presentation, attendees will be equipped with how to create and detect deepfakes, a deep understanding of the technology and its impact.

Talk Outline:

I. How Deepfakes Work (Approx. 8 minutes)

  • Step-by-step explanation of deepfake creation using an opensource tool
  • Clarifying the technical aspects behind manipulating existing media with AI algorithms

II. Deepfakes with GANs (Approx. 8 minutes)

  • Introduction to Generative Adversarial Networks (GANs) and their role in deepfake generation
  • Different types of GANs and how to craft realistic deepfakes

III. The Good and the Bad (Approx. 8 minutes)

  • Exploring the positive effects of deepfakes
  • Unveiling the negative implications of deepfakes
  • Real-world examples highlighting the ethical concerns
  • Speculating on the future developments of deepfake technology

IV. How to Recognize Deepfakes (Approx. 6 minutes)

  • Insight into the ongoing efforts to combat the misuse of deepfakes
  • Various approaches and AI-driven tools for detecting deepfake media
  • Understanding the limitations in detecting increasingly sophisticated deepfakes

Key Takeaways:

  • In-depth understanding of deepfake creation and the role of GANs
  • Awareness of the positive and negative impacts of deepfakes in different domains
  • Real-world examples illustrating the ethical concerns surrounding deepfakes
  • Insights into the future trends and advancements in deepfake technology
  • Familiarity with a range of AI-based approaches and tools for detecting deepfakes
Prior Knowledge Expected –

No previous knowledge expected
Maryam Miradi

Maryam Miradi is AI and Data Science Lead at Transactie Monitoring Nederland (TMNL). She has a PhD in Artificial Intelligence Deep Learning, specialised in NLP and Computer Vision from Delft University of Technology. The last 15 years, she has developed different AI solutions for Organisations such as Ahold-Delhaize, Belastingdienst, Alliander, Stedin and ABN AMRO