/Weekly-Report

Primary LanguageHTML

My Learning Progress as a PhD Student @RPI

I started my PhD journey from September 2024 at RPI under the advise of Prof. Qiang Ji. During the first few months, my mainly topics will be Uncertainty in Deep Learning, say, the Probabilistic Machine Learning or Bayesian Machine Learning.

Weekly Report

I present weekly report and discuss questions with Prof. Ji, and I will upload my material (mainly presentation slides) here.

2024 Fall

  1. Week 1: A Small Step into Bayeisan Deep Learning and Uncertainty
  2. Week 2: Read Hanjing's Work about UQ, UA
  3. Week 3: Hands-on | Deep Ensemble, MC Drop-out
  4. Week 4: Hands-on | UA via Vanilla Gradient and FullGrad
  5. Week 5: UA hands-on & Readings on Uncertianty❌Generative Models
  6. Week 6: [Results about Uncertainy-guided Denoising & DAG-GFlowNet: understand paper and run the code & More Readings on Uncertainty]

Paper Reading

I will record some of my reading progress (the material, the paper) here.

  • [Basics in BNN] A Primer on Bayesian Neural Networks: Review and Debates [arxiv 2023]
  • [Basics in BNN] An Introduction to Bayesian Neural Networks @[ProbAI 2022 summer school] [slides] [notes]
  • [Basics in BNN] A Survey on Bayesian Deep Learning [ACM computing surveys 2020]
  • [Uncertainty Attribution] (work from our group's senior studetns) Gradient-based uncertainty attribution for explainable bayesian deep learning [CVPR 2023]
  • [Uncertainty Quantification | Pre-trained Model] (work from our group's senior studetns) Epistemic Uncertainty Quantification For Pre-Trained Neural Networks [CVPR 2024]
  • [UQ | Ensemble] (work from our group's senior studetns) Diversity-enhanced probabilistic ensemble for uncertainty estimation [UAI 2023]
  • [UQ | Latent/Semantic] (work from our group's senior studetns) Semantic Attribution for Explainable Uncertainty Quantification [Epi UAI 2023
  • [Uncertainty and Transformer] (work from our group's senior studetns) Uncertainty-guided probabilistic transformer for complex action recognition [CVPR 2022]
  • [Uncertainty in Causal Graph] (work from our group's senior studetns) Quantifying Uncertainty in Causal Graphs: A Key to Improved Domain Generalization Predictions [not published yet]
  • [Uncertainty in Diffusion] BayesDiff: Estimating Pixel-wise Uncertainty in Diffusion via Bayesian Inference [ICLR 2024]
  • [Uncertainty in Diffusion] Hyper-Diffusion: Estimating Epistemic and Aleatoric Uncertainty with a Single Model [Arxiv 2024]
  • [Estimate Probabilistic Graphs] Bayesian Structure Learning with Generative Flow Networks [UAI 2022]