Possible list of topics for presentation

These are examples of topics covered by successful projects of the last few years. These are just as an example, we strongly encourage new topics, state of the art possibly, particularly in variational methods, causal ML, and generative AI. 

• Deep Boltzmann Machines
• Langevin Monte Carlo and other advanced MC methods
• Probabilistic Programming 
• Gaussian Processes (sparsification, Deep Gaussian Processes)
• Randomized projections for kernel machines (Ali Rahimi)
• Bayesian Generalised Linear Models
• speech recognition / tagging with deep probabilistic state space models for time series
• Causal ML
• Bayesian Convolutional NN
• MCMC + HMM for Cryptoanalysis (cryptography related tasks)
• Predicting movies preferences with VAE
• Piano music with latent diffusion
• Variational-based methods for real world clinical data
• Probabilistic random forests
• Inverse Problems via shortcut sampling diffusion
• Dirichlet process mixture models for image segmentation
• HMM for handwritten digit on pen strokes
• Clustering with Dirichelet mixture models (advanced stuff only)
• Advanced topics on Bayesian and kernel PCA
• VAE for music data
• Anomaly detection with VAE
• Combining different modalities in diffusion models
• synthetic data generation (with LLM or other models)
• financial crisis detection with RVAE
• Copula based HMM for  football tactics

Examples of unsuccessful projects include:

• discussing basic methods on a simple case study: e.g. linear regression on financial data 
• PCA and probabilistic PCA (oh no, not again!)
• Basic examples of Dirichelet clustering or LDA. 
• Anything you already studied in a course of statistics :-)
• A project solely based on LLMs...