Meetup
Introduction
Wir organisieren Events mit Industriepartnern, Workshops und Coding-Sessions in unregelmässigen Zeitabständen an verschiedenen ZHAW Campus-Locations.
Summary & Key Outcomes of the International Conference on Machine Learning (ICML) 2024
When: 10.10.2024, 18:00-20:00
Where: ZHAW Campus Winterthur Technikumstrasse(PDF 2,1 MB) TS O1.13
Organiser: Pavel Sulimov
Title: Summary & Key Outcomes of the International Conference on Machine Learning (ICML) 2024
Description: Regardless of your specific area of interest within ML - be it natural language processing (NLP), computer vision (CV), reinforcement learning (RL), diffusion models, applied or theoretical frameworks - you are sure to find valuable insights and concise summaries of novel developments during our recap session. We will focus on
- a tutorial on the physics of large language models (LLMs) which emerged as one of the most talked-about sessions,
- reinforcement learning from human feedback
- superalignment,
- supervised fine-tuning and reward/RL for LLMs
- and scalable recommender systems
Generative AI on Images - Painting like Monet with VAE, GAN or Diffusion
When: 18.7.2024, 18:00-20:00
Where: ZHAW Campus Winterthur Technikumstrasse(PDF 2,1 MB) TN O1.46
Organiser: Pavel Sulimov
Title: Generative AI on Images - Painting like Monet with VAE, GAN or Diffusion
Description: Recent breakthroughs in image generation, have showcased the incredible potential of these technologies. State-of-the-art models in this field, such as DALL-E 2, leverage advanced diffusion mechanisms. But do you always need to use the latest and most complex methods? Depending on your specific task, simpler models might be more effective.
During this meetup, we'll delve into the key milestones of generative models in the continuous generation domain, including Variational Autoencoders (VAEs), Generative Adversarial Networks (GANs), and Diffusion Models. We'll explore the nuances of their training processes and discuss their applicability to and effectiveness in different problem types compared to simpler approaches. We will also discuss how incorporating attention mechanisms can improve such generative models.
In addition to structured presentations, you'll have the chance to participate in engaging quizzes and try your hand at a Kaggle competition. Whether you're a seasoned AI practitioner or just starting out, this meetup will provide valuable insights and practical knowledge.
Prerequisites:
- basic understanding of neural networks
- basic understanding of deep learning Python frameworks (e.g. PyTorch, JAX, Keras etc.)
Registration: Please reserve your spot here
Quantum Machine Learning: The next big thing after Large Deep Learning models?
When: 23.5.2024, 17:45-20:00
Where: ZHAW Campus Winterthur Technikumstrasse(PDF 2,1 MB) TN O1.46
Organiser: Pavel Sulimov
Title: Quantum Machine Learning: The next big thing after Large Deep Learning models?
Description: The conceptual input will provide answers to what types of machine learning problem could be solved by quantum algorithms faster and/or with better accuracy and how a "quantum version" of a "classical" algorithm can be developed. In the workshop, we will dive deeper into the basics of quantum computation, how quantum models differ architecturally from classical ones, when it makes sense to switch to quantum models, and how ZHAW students and employees can construct their first ansatz and run experiments on IBM quantum computers with 100+ quantum bits.
Registration: Please reserve your spot here