Course Objective & Scope
Going beyond calling hosted generative APIs, this course covers the mathematical and architectural families of deep generative modelling for engineers who need domain-specific synthesis, controllable multimodal generation or synthetic data. It compares Variational Autoencoders (VAEs and ELBO), GANs, Denoising Diffusion Probabilistic Models (DDPM, DDIM), Latent Diffusion, Flow Matching and discrete-token autoregressive models.
Classifier-free guidance, structural conditioning with ControlNet and LoRA, synthetic tabular/time-series generation and evaluation of fidelity, diversity and training-data memorisation are practised hands-on in PyTorch and Hugging Face Diffusers.
Who Should Attend?
- Deep learning engineers, data scientists and ML researchers
- Teams working on image, audio, design synthesis or synthetic tabular/time-series data
- Engineers adapting generative models to enterprise domains with LoRA and ControlNet
- Prerequisite: PyTorch and deep learning fundamentals
Key Takeaways
- Navigate the generative trilemma across VAEs, GANs, Diffusion, Flow Matching and autoregressive models
- Train and fine-tune conditional latent diffusion and Flow Matching pipelines with PyTorch and Diffusers
- Enforce structural constraints and domain style using ControlNet and LoRA adapters
- Generate realistic synthetic tabular and time-series data for rare-event augmentation and privacy
- Evaluate generative outputs on fidelity, coverage, diversity and memorisation/privacy leakage
Curriculum
Generative Modelling Taxonomy: VAEs and GANs
- Latent variable models, ELBO, VQ-VAE and discrete codebooks
- Adversarial losses, mode collapse and the generative learning trilemma
Diffusion and Flow Matching Architectures
- Forward noising, reverse denoising (DDPM/DDIM) and score matching
- U-Net and Diffusion Transformer (DiT) backbones; fast sampling with Flow Matching
Conditional Generation, LoRA and ControlNet
- Cross-attention conditioning and Classifier-Free Guidance (CFG)
- LoRA fine-tuning and structural conditioning with ControlNet in Diffusers
Synthetic Tabular and Time-Series Generation
- Synthetic data for rare-class augmentation and safe test environments
- CTGAN, TVAE and diffusion-based tabular/temporal generators
Evaluation, Memorisation and Safety
- FID, CLIP-Score, distributional precision/recall and Train-on-Synthetic-Test-on-Real (TSTR)
- Checking for training-data memorisation and provenance watermarking
Other Courses in This Category
-
MDPs, Q-Learning, DQN, Policy Gradients (PPO, SAC), contextual bandits and Offline RL using Gymnasium and Stable-Baselines3 for dynamic pricing, resource allocation and sequential control problems.
- Live Course
- 5 modules
-
Deep fraud models over transaction sequences, session clickstreams and payment networks: Transformer/RNN sequence encoders, autoencoder anomaly scores, GNNs for fraud rings and hybrid GBDT+DL architectures.
- Live Course
- 5 modules
-
Deep defect and fault detection across inspection images, acoustic/vibration signals and system logs using PatchCore/PaDiM, Anomalib, YOLO/segmentation, 1D-CNNs/autoencoders and normal-only training.
- Live Course
- 5 modules
-
Multivariate, long-horizon and probabilistic time-series forecasting with N-BEATS, N-HiTS, Temporal Fusion Transformer (TFT), PatchTST, DeepAR and TCN using NeuralForecast and PyTorch Forecasting.
- Live Course
- 5 modules
-
Large-scale deep recommendation architectures: Two-Tower retrieval with ANN (FAISS/Qdrant), sequential session recommenders (SASRec/BERT4Rec), DeepFM/DCN-v2/DLRM ranking and multi-task MMoE learning.
- Live Course
- 5 modules
-
Text preprocessing, Turkish morphology, regex, tokenisation, TF-IDF, word/sentence embeddings and building first text classification, NER and sentiment pipelines with spaCy, scikit-learn and Hugging Face.
- Live Course
- 5 modules
Get in touch
