Research Paper Presentations


Each group of students will give a presentation on a research paper that is broadly related to transfer learning. Overall, the presentation will be worth 30% of the total grade. Each group will need to sign up for the presentation in a google sheet (the link will be shared on Piazza).

Each talk should be around 30 mins in length, with 20~25 mins for the technical presentations and 5~10 mins for questions and discussion. Please send your slides to the instructor at least one week before your presentation for feedback.

Research Papers

The following list of papers only serve as a reference for you to choose from. You are free to choose any paper that interests you, as long as it is relevant to the course topic (broadly).
Domain Adaptation / Generalization
Domain adaptation: Learning bounds and algorithms
Invariant Risk Minimization
Correcting Sample Selection Bias by Unlabeled Data
Impossibility Theorems for Domain Adaptation
Robust Supervised Learning
Joint Distribution Optimal Transportation for Domain Adaptation
The Risks of Invariant Risk Minimization
Domain Generalization via Invariant Feature Representation
Generalizing from Several Related Classification Tasks to a New Unlabeled Sample
In Search of Lost Domain Generalization
Adjusting the Outputs of a Classifier to New a Priori Probabilities: A Simple Procedure
A Unified View of Label Shift Estimation
Distributionally Robust Neural Networks for Group Shifts: On the Importance of Regularization for Worst-Case Generalization
Multitask Learning
A Convex Formulation for Learning Task Relationships in Multi-Task Learning
Multi-Task Feature Learning
On the Theory of Transfer Learning: The Importance of Task Diversity
Learning Task-Agnostic Embedding of Multiple Black-Box Experts for Multi-Task Model Fusion
Which Tasks Should Be Learned Together in Multi-task Learning?
Multi-Task Learning as Multi-Objective Optimization
Pareto Multi-Task Learning
Bayesian Multitask Learning with Latent Hierarchies
Multi-Task Learning Using Uncertainty to Weigh Losses for Scene Geometry and Semantics
Provable Meta-Learning of Linear Representations
Recasting Gradient-Based Meta-Learning as Hierarchical Bayes
On First-Order Meta-Learning Algorithms
Rapid Learning or Feature Reuse? Towards Understanding the Effectiveness of MAML
Meta-Learning with Implicit Gradients

Grading Scheme

Each presentation will be graded based on the following criterion:
  • Clarity of the presentation;
  • Literature review, background and motivation of the work;
  • A high-level summary of the contributions of the work;
  • Detailed introduction to a key technical result of the paper;
  • Discussion on potential future research directions and applications;
  • Answers to the potential questions from the audience.