Boostcamp-AI-Casual-Talk

개요

• 네이버 AI랩의 하정우 소장님의 Weekly Arxiv Casual Talk에 자극받아 시작하게 된 잡담 세션입니다.
• 매주 1시간 부담없이 가볍게 AI 관련 이야기를 합니다.
• 꼭 논문 관련일 필요도 없고 시각적 자료도 준비할 필요없습니다.
• 얕고 넓게가 목적인 잡담입니다! 논문 리뷰와 다르게 디테일, 수식적 부분에 대한 설명은 부족 할 수도 있습니다.

운영방법

• 매주 수요일 밤 10시~ 11시
• 정규 모더레이터: 이호진 김인정 조영하 정인식 박성호
• 소개 하고 싶은 내용이 있다면 Issue에 링크, 자료 등을 올려주시고 DM이나 줌에서 반응 해주시면 발언권을 드리겠습니다!

다시보기

• ep5: 5회차:20220427

  • StyleGAN-Human: A Data-Centric Odyssey of Human Generation/Photorealistic Monocular 3D Reconstruction of HumansWearing Clothing
  • Performer: A Novel PPG to ECG Reconstruction Transformer For a Digital Biomarker of Cardiovascular Disease Detection
  • Prefrontal cortex as a meta-reinforcement learning system
  • 크래프톤 딥러닝 가능성 보여준 퍼즐게임 '푼다' / 세계 최초, 강화학습 AI 기술 기반 새로운 MMORPG 콘텐츠를 제시하다

• ep4: 4회차:20220420 Youtube 링크

  • Solving ImageNet: a Unified Scheme for Training any Backbone to Top Results/Neighborhood Attention Transformer
  • Social Diversity and Social Preferences in Mixed-Motive Reinforcement Learning
  • Mip-NeRF: A Multiscale Representation for Anti-Aliasing Neural Radiance Fields
  • Transformers4Rec/RQ Transformer/ Grad Cache

• ep3: 3회차:20220413 Youtube 링크

  • PaLM / Training Compute-Optimal Large Language Models/MultiMAE: Multi-modal Multi-task Masked Autoencoders
  • Recovering individual emotional states from sparse ratings using collaborative filtering
  • Face scanner Clearview AI aims to branch out beyond police/ DALL.E2 /Salesforce CodeGen: Program Using Your Voice

• ep2: 2회차:20220330 Youtube 링크

  • GTC SORDI/Recommendation as Language Processing (RLP):P5 /Make-A-Scene: Scene-Based Text-to-Image Generation with Human Priors
  • GTC 2022 review: Chelsea Finn / Using deep reinforcement learning to reveal how the brain encodes abstract state-space representations in high-dimensional environments
  • Ukraine uses facial recognition to identify dead Russian soldiers, minister says/Revisiting Neighborhood-based Link Prediction for Collaborative Filtering
  • CLIP Guidance Diffusion model/ novel-illustration-disco-diffusion

• ep1: 1회차:20220323 Youtube 링크

  • 이루다/Stanford AI Index Report 2022/FORCE: A Framework of Rule-Based Conversational Recommender System
  • Aesthetic preference for art can be predicted from a mixture of low- and high-level visual features
  • OpenAI, Davinci model/P-tuning
  • GopherCite: Teaching language models to support answers with verified quotes/Using Deep Learning to Annotate the Protein Universe/ProteInfer: deep networks for protein functional inference
  • Web 3.0