Deep Learning

Neurons, backpropagation, convolution, pooling, and LeNet

• Representation learning
• Forward & backward propagation
• Convolution geometry
• LeNet shape tracking

Training loops, hyperparameters, regularization, and residual learning

• Keras training workflow
• Dense vs. CNN on MNIST
• Callbacks & regularization
• VGG and ResNet

Boxes, detector families, mAP, and instance masks

• IoU, AP, mAP & NMS
• R-CNN to Faster R-CNN
• YOLO and DETR
• Mask R-CNN

GANs, diffusion, verification, and metric learning

• GAN min-max training
• Forward and reverse diffusion
• Identification vs. verification
• Contrastive & triplet learning

Sequence state, BPTT, LSTM, GRU, and image captioning

• Sequence layouts
• BPTT & vanishing gradients
• LSTM and GRU gates
• CNN-RNN captioning

Sparse text features, Word2Vec, similarity, and contextual meaning

• NLP tasks & levels
• Bag-of-words and TF-IDF
• Skip-gram and CBOW
• Static vs. contextual embeddings

Query-key-value attention, masking, Transformer blocks, and BERT

• Alignment & context vectors
• Self and multi-head attention
• Position & causal masking
• Transformer and BERT

QA metrics, RAG, prompting strategies, and learned prompts

• QA taxonomy & SQuAD
• Retrieval-augmented generation
• Few-shot & reasoning prompts
• BiomedCoOp prompt learning