Week1<\/strong> <\/span>(03.08): Introduction<\/p>\n<\/div>\n

Reading Assignment<\/p>\n

1. Transformer: Attention Is All You Need, NIPS 2017
\nhttps:\/\/jalammar.github.io\/illustrated-transformer\/<\/p>\n

2. GPT3: Language Models are Few-Shot Learners, NeurIPS 2020
\nhttps:\/\/jalammar.github.io\/illustrated-gpt2\/<\/p>\n

Week2<\/span><\/strong> (03.15): ChatGPT\u00a0(Overview of transformer\/GPT)<\/p>\n

https:\/\/openai.com\/blog\/chatgpt<\/p>\n

https:\/\/platform.openai.com\/docs\/model-index-for-researchers<\/p>\n

Papers for ChatGPT<\/p>\n

L01-1: Training language models to follow instructions with human feedback, NeurIPS 2022 (InstructGPT)<\/p>\n

L01-2: Learning to summarize from human feedback, NeurIPS 2020<\/p>\n

L01-3: Asynchronous Methods for Deep Reinforcement Learning, ICML 2016<\/p>\n

L01-4: Deep Reinforcement Learning from Human Preferences, NIPS 2017<\/p>\n

L01-5: Fine-Tuning Language Models from Human Preferences, arXiv 2019<\/p>\n

L01-6 (PPO): Proximal Policy Optimization Algorithms, arXiv 2017<\/p>\n

References<\/p>\n

R. Sutton and A. Barto, Reinforcement Learning: An Introduction, MIT Press, Cambridge, 2018<\/p>\n

Trust Region Policy Optimization, ICML 2015<\/p>\n

Week3<\/strong> <\/span>(03.22): ChatGPT (Policy optimization, TRPO\/PPO)<\/p>\n

Week4<\/strong><\/span> (03.29): ChatGPT (DRL from HP)<\/p>\n

Week5<\/strong><\/span> (04.05): ChatGPT (FT LM from HP)<\/p>\n

Week6<\/strong> <\/span>(04.12): ChatGPT (InstructGPT)<\/p>\n

Week7<\/strong><\/span> (04.19): AlphaCode<\/p>\n

L02-1: Competition-level code generation with AlphaCode, Science 2022
\nL02-2: Supplementary for Competition-level code generation with AlphaCode, Science 2022
\nL02-3: TEXT GENERATION BY LEARNING FROM DEMONSTRATIONS, ICLR 2021 (GOLD)<\/p>\n

Week8<\/strong><\/span> (04.26): Midterm (ChatGPT)<\/p>\n

Week9<\/strong><\/span> (05.03): DDGM (L03-1: Tutorial Part 1, L03-2: DDPM)<\/p>\n

Papers for DDGM<\/p>\n

L03-1: Denoising Diffusion-based Generative Modeling (DDGM): Foundations and Applications, CVPR 2022 Tutorial
\nL03-2: Denoising-diffusion-probabilistic-models (DDPM), NeurIPS 2020
\nL03-3: Deep Unsupervised Learning using Nonequilibrium Thermodynamics, ICML 2015<\/p>\n

Week10<\/strong><\/span> (05.10): DDGM (L03-2: DDPM)<\/p>\n

Useful link:
\nWhat are diffusion models? https:\/\/lilianweng.github.io\/posts\/2021-07-11-diffusion-models\/<\/p>\n

References
\nPrince, S.J.D., Computer Vision: Models, Learning and Inference, 2012 (5.6 Product of two normals)<\/p>\n

Week11<\/strong> <\/span>(05.17): DDGM (Score-based: L03-4)<\/p>\n

L03-4: Generative Modeling by Estimating Gradients of the Data Distribution, https:\/\/yang-song.net\/blog\/2021\/score\/
\nL03-4 ref-1: Sliced Score Matching: A Scalable Approach to Density and Score Estimation, UAI 2019 (Section 2.1)
\nL03-4 ref-2: Estimation of Non-Normalized Statistical Models by Score Matching, JMLR 2005 (Theorem 1)<\/p>\n

Week12<\/strong><\/span> (05.24): DDGM (Score-based: L03-5 1\/2)<\/p>\n

L03-5: SCORE-BASED GENERATIVE MODELING THROUGH STOCHASTIC DIFFERENTIAL EQUATIONS, ICLR 2021
\nL03-5 ref-1: A Connection Between ScoreMatching and Denoising Autoencoders, 2011 ((4.3), (4.4))
\nL03-5 ref-2: Generative Modeling by Estimating Gradients of the Data Distribution, NeurIPS 2019 ((1), (2), (5), (6))<\/p>\n

References
\nSimo Sarkka and Arno Solin. Applied stochastic differential equations, volume 10. Cambridge University Press, 2019 (Def. 3.9 (White noise), Def. 4.1 (Brownian motion), Theorem 5.4 (Fokker-Planck-Komogorov equation), Equations (5.50) and (5.51))<\/p>\n

Week13<\/strong> <\/span>(05.31): DDGM (Score-based: L03-5 2\/2)<\/p>\n

Week14<\/strong> <\/span>(06.07): DDGM (Accelerated Sampling)<\/p>\n

L03-6: DENOISING DIFFUSION IMPLICIT MODELS, ICLR 2021 (DDIM)<\/p>\n

Note: Download the lecture note from the course BBS.<\/p>\n

Week15<\/strong><\/span> (06.14): Final (DDGM)<\/p>\n

Week16<\/strong><\/span> (06.21): DDGM (L03-1: Tutorial Part 2 for Conditional generation and cascaded generation for Imagen and DALL\u00b7E 2)<\/p>\n

L03-7-0.9: Diffusion models beat GANs on image synthesis, NeurIPS 2021 (Section 4.1-4.2)
\nL03-7-1:\u00a0 CLASSIFIER-FREE DIFFUSION GUIDANCE,\u00a0 NeurIPS Workshop 2021 (Section 3)
\nL03-7-2: Cascaded Diffusion Models for High Fidelity Image Generation, JMLR 2022 (Section 2-3)<\/p>\n

 <\/p>\n","protected":false},"excerpt":{"rendered":"

Random Signal Analysis ECE629 Spring 2023 Professor Sanghoon Sull School […]<\/p>\n","protected":false},"author":1,"featured_media":0,"parent":56393,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"spay_email":""},"jetpack_sharing_enabled":true,"_links":{"self":[{"href":"https:\/\/dml.korea.ac.kr\/index.php?rest_route=\/wp\/v2\/pages\/56483"}],"collection":[{"href":"https:\/\/dml.korea.ac.kr\/index.php?rest_route=\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/dml.korea.ac.kr\/index.php?rest_route=\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/dml.korea.ac.kr\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/dml.korea.ac.kr\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=56483"}],"version-history":[{"count":48,"href":"https:\/\/dml.korea.ac.kr\/index.php?rest_route=\/wp\/v2\/pages\/56483\/revisions"}],"predecessor-version":[{"id":56616,"href":"https:\/\/dml.korea.ac.kr\/index.php?rest_route=\/wp\/v2\/pages\/56483\/revisions\/56616"}],"up":[{"embeddable":true,"href":"https:\/\/dml.korea.ac.kr\/index.php?rest_route=\/wp\/v2\/pages\/56393"}],"wp:attachment":[{"href":"https:\/\/dml.korea.ac.kr\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=56483"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}