Winter 2021 CS291A Syllabus

Checkout the class presentation schedule for additional readings: Class Presentation Schedule

• 1/4 Introduction, logistics, and deep learning.
• 1/6 Tips for a successful class project
• 1/11 Neural network basics, & backpropagation
• 1/13 Word embeddings (Project proposal due 23:59PT 1/13 submission link, HW1 out)
  • Enriching Word Vectors with Subword Information
  • Contextual String Embeddings for Sequence Labeling
• 1/18 NO CLASS (University Holiday: Martin Luther King Jr. Day)
• 1/20 RNNs
  • Recurrent neural network based language model
  • Conditional Random Fields as Recurrent Neural Networks
• 1/25 LSTMs/GRUs
  • Deep contextualized word representations
  • Memory Networks
• 1/27 Sequence-to-sequence models
  • fairseq: A Fast, Extensible Toolkit for Sequence Modeling
  • Sequence Level Training with Recurrent Neural Networks
• 2/1 Convolutional Neural Networks (HW1 due and HW2 out)
  • Densely Connected Convolutional Networks
  • Improved protein structure prediction using potentials from deep learning
• 2/3 Attention mechanisms
  • A Neural Attention Model for Sentence Summarization
  • Neural Machine Translation by Jointly Learning to Align and Translate
• 2/8 Transformer and BERT (Mid-term report due submission link)
  • XLNet: Generalized Autoregressive Pretraining for Language Understanding
  • RoBERTa: A Robustly Optimized BERT Pretraining Approach
  • ALBERT: A Lite BERT for Self-supervised Learning of Language Representations
• 2/10 Mid-term project updates upload your slide here by 2/9 noon
• 2/15 NO CLASS (University Holiday: Presidents' Day)
• 2/17 Language and vision
  • An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
  • Neural Topological SLAM for Visual Navigation
• 2/22 Deep Reinforcement Learning 1
  • Imagination-Augmented Agents for Deep Reinforcement Learning
  • Causal Discovery with Reinforcement Learning
• 2/24 Deep Reinforcement Learning 2 (HW2 due: 02/26 Friday 11:59pm)
  • Curiosity-driven Exploration by Self-supervised Prediction
  • Grandmaster level in StarCraft II using multi-agent reinforcement learning
  • Mastering Atari, Go, chess and shogi by planning with a learned model
• 3/1 Generative Adversarial Networks
  • Towards Principled Methods for Training Generative Adversarial Networks
  • Wasserstein GAN
  • Boundary Equilibrium GAN
• 3/3 Project: final presentation (1) submission link by 3/2 noon.
• 3/8 Project: final presentation (2) submission link by 3/5 noon.
• 3/10 Project: final presentation (3) submission link by 3/9 noon.
• 3/19 23:59PM PT Project Final Report Due submission link.