[B급 프로그래머] 3월 2주 소식(빅데이터/인공지능, 하드웨어/블록체인, 읽을거리 부문)

(오늘의 짤방: Timeline history from Eliza to ChatGPT via @s3thr1n)

1. 빅데이터/인공지능
   • mindsdb - Build ML powered applications fast.
   • Microsoft Designer - AI를 이용한 디자인 도구 (designer.microsoft.com)
   • Using ChatGPT for DevOps
   • Learn To Reproduce Papers: Beginner’s Guide
   • Jupyter Notebook as a Product
   • MosaicBERT: Pretraining BERT from Scratch for $20
   • OpenXLA 공개 - 머신러닝을 가속하고 단순화 시켜주는 프로젝트 (opensource.googleblog.com)
   • An interesting difference I’ve noticed between ChatGPT and davinci-003:
   • 11 Powerful output formats ChatGPT can produce ( with examples ) 🧵:
   • PaLM-E: An embodied multimodal language model
   • Creating a Resume Automation Website With ChatGPT in 10 Minutes
   • Show GN: ChatGPT 기술을 사용한 AI 면접 코칭 서비스 (wanted.co.kr)
   • Kindle GPT - AI search & chat on your Kindle highlights.
   • Data2vec 2.0: Highly efficient self-supervised learning for vision, speech and text
   • Inference of Facebook's LLaMA model in pure C/C++
   • Evidence of a predictive coding hierarchy in the human brain listening to speech
   • Prismer: A Vision-Language Model with Multi-Modal Experts
   • *If* GPT-4 is multimodal, we can predict with reasonable confidence what GPT-4 *might* be capable of, given Microsoft’s prior work Kosmos-1:
   • ‘생성 AI, 신뢰성 확보 쉽지 않네’··· 빙AI, 덕어시스트의 ‘출처 제공 기능’ 검토해보니...
   • 마이크로소프트, 애저 오픈AI 서비스에 챗GPT 프리뷰 추가
   • CHATGPT IS NOW FINDING BUGS IN DATABASES
   • Visual ChatGPT connects ChatGPT and a series of Visual Foundation Models to enable sending and receiving images during chatting.
   • “이미 실용성 가득하더라”··· 생성 AI에 대한 IT리더들의 접근법
   • 덕덕고, 생성AI 기반 응답 서비스 '덕어시스트' 발표··· "환각 현상 최소화"
   • 세일즈포스, 아인슈타인GPT ‘비공개 베타 테스트’로 공개...생성 AI 도입에 ‘신중’
   • Online gradient descent written in SQL
   • 구글 Universal Speech Model(USM) : 100개 이상의 언어를 지원하는 최첨단 Speech AI (ai.googleblog.com)
   • Prompt Engineering Overview
   • ChatGPT is now available in Azure OpenAI Service
   • Fine-tuning 20B LLMs with RLHF on a 24GB consumer GPU
   • GPT-4 is coming next week – and it will be multimodal, says Microsoft Germany
   • ML Papers Explained - Explanations to key concepts in ML
   • How Unstructured and LlamaIndex can help bring the power of LLM’s to your own data
   • Machine learning with Flowchart
   • Minimal LLaMA
   • ChatGPT's API is So Good and Cheap, It Makes Most Text Generating AI Obsolete
   • A Comprehensive Survey of AI-Generated Content (AIGC): A History of Generative AI from GAN to ChatGPT
   • 18 Data Science Project Ideas for Beginners
   • Deep Learning with Python: Neural Networks (complete tutorial)
   • The training notes of GLM-130B
   • I was asked this question a lot; why BLOOM-175B performance is not as good as expect?
   • How to avoid machine learning pitfalls: a guide for academic researchers
   • Microsoft, 비즈니스 앱 제품군에 AI 기반 Copilot 제공 (techcrunch.com)
   • BIGKINDS-LAB - 한국언론진흥재단과 (주)포스웨이브에서 제작한 KPF-BERT fine-tunning 모델을 제공합니다.
   • 개발자 관점에서 한글 번역기 비교해봄 (생산성 향상 TIP?)
   • Word Embedding of BERT
   • [정보] 구글 검색결과에서 스팸 사이트가 발생하는 원인 및 해결법
   • PyTorch Crash Course
   • There's an AI for That (theresanaiforthat.com)
   • LLaMa CPU fork
   • Microsoft Designer
   • HackerFM - 해커뉴스 내용을 기반으로 AI가 진행하는 팟캐스트 (hackerfm.com)
   • “AI+집단 지성”··· 링크드인, AI 콘텐츠와 전문가 통찰력 합친 ‘협동 기사’ 출시
   • Salesforce follows Microsoft in launching A.I. tools for salespeople with help from OpenAI
   • Linear Algebra — Survival Kit for Machine Learning
   • 2022년에 가장 많이 인용된 AI 논문 100 (zeta-alpha.com)
   • The Grind a Day: thousands of Apple II floppy disks archived
   • Amazon SageMaker로 컬리(Kurly) 상품 후기 분류 모델 개발하기
   • ChatGPT Explained: A Normie's Guide To How It Works
   • This repository contains my lecture notes from graduate school on following topics 👇🏼(machine learning, data science, computer science, statistics, and more. 🌟)
   • Here are 6 Colab hacks that are really worth knowing: 🧵👇
   • Category Theory ∩ Machine Learning
   • Graphs with Python: Overview and Best Libraries
   • PaLM-E: An Embodied Multimodal Language Model
   • "돈 보내줘"…아들 목소리 AI에 2천만원 털렸다
   • OpenICL provides an easy interface for in-context learning, with many state-of-the-art retrieval and inference methods built in to facilitate systematic comparison of LMs and fast research prototyping.
   • GPT-3.5 + ChatGPT: An illustrated overview
   • A Full End-to-End Deployment of a Machine Learning Algorithm into a Live Production Environment
   • Object detection is a fundamental problem in computer vision. Although Vision Transformers (ViTs) achieve state-of-the-art performance today, the history of object detection proceeded in three distinct generations of innovation… 🧵 [1/7]
   • The 10 Best Data Visualizations of 2021
   • GitHub Enhanced Copilot with New AI Model and Security-Oriented Capabilities
   • LLaMA: INT8 edition
   • 빙 AI 검색 사용해 보기(2) – 샌드위치 모델은?
   • How to format inputs to ChatGPT models
   • ChatGPT invented its own puzzle game.
   • Whisper.cpp - OpenAI Whisper의 고성능 C/C++ 구현체 (github.com/ggerganov)
   • Why AI Won't Cause Unemployment
   • Neural Networks Hate Text
   • Benchmarking the new Pandas PyArrow Backend
   • Grokking: Generalization Beyond Overfitting on Small Algorithmic Datasets
   • This repository provides very basic flask, Streamlit, and docker examples for the llama_index (FKA gpt_index) package.
   • AI Alignment Course - AGI Safety Fundamentals
   • Using CNN for financial time series prediction
   • ChatGPT API Zero Shot Sentiment Analysis on User-Defined Emotions
   • 🤗 Diffusers is the go-to library for state-of-the-art pretrained diffusion models for generating images, audio, and even 3D structures of molecules.
   • MLSys Seminars - Stanford MLSys Seminars
   • chatgpt-clone - Build Yourown ChatGPT with OpenAI API & Gradio
   • Logistic Regression from First Principles in Python
   • The Machine Learning repo on GitHub is where I put code for my twitter tutorials. You will find ⬇️
   • ChatGPT API 소식
   • High-resolution image reconstruction with latent diffusion models from human brain activity
   • TextSynth provides access to large language or text-to-image models such as GPT-J, GPT-Neo, M2M100, CodeGen, Stable Diffusion thru a REST API and a playground.
   • 31 Datasets For Your Next Data Science Project
   • whisper.cpp - High-performance inference of OpenAI's Whisper automatic speech recognition (ASR) model:
   • Machine learning algorithms - A collection of minimal and clean implementations of machine learning algorithms.
   • 상상을 현실로 만드는, 이미지 생성 모델을 위한 엔지니어링
   • ChatGPT가 촉발한 초거대 AI시대 우리의 대응 전략｜하정우 NAVER AI Lab 연구소장
   • 오픈AI, 챗GPT-위스퍼용 API 출시 “10배 저렴해진다”
   • brilliant - Join 10+ million people excelling in math and science
   • ChatGPT로 하루 만에 이커머스 앱 제안서 작성하기
   • "사실을 원해? 재미를 원해?" 정밀성과 창의성 모드 추가한 빙 챗 AI
   • 목소리 복제하는 AI 음성 비서, 금융권 본인 인증도 뚫는다
   • Paul Graham GPT
   • “이미 도래한 미래” 생성형 AI 시대의 SW 개발자를 위한 4가지 조언
   • a - CLI tool to generate code from GPT3 that pretty-prints the output by language, and copies to your clipboard
   • [독서광] 그림으로 배우는 StatQuest 머신러닝 강의(본인이 작성한 글)
   • Show GN: AiShell - ChatGPT 기반의 자연어 셸 툴 킷 (github.com/code-yeongyu)
   • 뤼튼, 전문적인 글쓰기 위한 AI 문장 생성 ‘뤼튼 도큐먼트’ 선보여
   • Microsoft Kosmos-1: Multimodal LLM(MLLM) (arxiv.org)
   • ChatLLaMA - LLaMa 기반 ChatGPT의 오픈소스 구현체 (github.com/nebuly-ai)
   • New – Amazon Lightsail for Research with All-in-One Research Environments
   • ChatML: ChatGPT API expects a structured format, called Chat Markup Language
   • Improving Students’ Learning With Effective Learning Techniques: Promising Directions From Cognitive and Educational Psychology
   • With the release of the ChatGPT API, developers can now do chat and traditional completion workflows for 90% cheaper than text-davinci-003!
   • Introducing ChatGPT and Whisper APIs
   • Build your own AI: how to teach GPT to search the web and execute code
   • Training language models to follow instructions with human feedback
   • GitHub Enhanced Copilot with New AI Model and Security-Oriented Capabilities
   • 라벨 스무딩(Label smoothing), When Does Label Smoothing Help?
   • How Loneliness Reshapes the Brain
   • 일부 회사들, '중요한 일'에는 의존해서는 안된다는 경고에도 불구하고 일부 직원을 ChatGPT로 대체중 (fortune.com)
   • 🌎 Internet Explorer 🌏 - Targeted Representation Learning on the Open Web
   • Navigating the Grey Area: Expressions of Overconfidence and Uncertainty in Language Models
   • Why didn't DeepMind build GPT3?
   • Say goodbye to messy CSV files and hello to 𝐅𝐞𝐚𝐭𝐮𝐫𝐞 𝐒𝐭𝐨𝐫𝐞𝐬!! The future of machine learning data management🚀 A Thread 🧵👇
   • PatchTST: rather than operating on every individual time step, instead use patching (like in ViTs) to chunk up the time series.
   • ChatGPT에서 채팅 만으로 웹 애플리케이션을 만드는 프롬프트 활용 사례와 시사점
   • Colleges keep publishing their machine learning courses online.
   • gitgpt - Natural Language Git CLI assistant
   • By using JAX! This simple example speeds up a function from 331 ms to 3.81 ms! Learn about JAX here. 1/8🧵
   • Create Presentation from Jupyter Notebook
   • Diff Models - 코드를 편집하는 새로운 방법 (carper.ai)
   • 머신러닝 단기집중과정 - TensorFlow API 사용 by 구글
   • 137 emergent abilities of large language models
   • Sounding The Secrets Of AudioLM
   • Understanding Deep Learning
   • transcribe-anything - OpenAI Whisper를 이용해서 자막만들기 (github.com/zackees)
   • Interactive Composition Explorer 🧊 - ICE is a Python library and trace visualizer for language model programs.
   • Semantic Uncertainty: Linguistic Invariances for Uncertainty Estimation in Natural Language Generation
   • A Curated List of LLMs and the Model's License
   • 👉 NLP Course by @huggingface
   • SymPy makes math fun again
   • Planting Undetectable Backdoors in Machine Learning Models
2. 하드웨어
   • Disambiguating Arm, Arm ARM, Armv9, ARM9, ARM64, AArch64, A64, A78, ...
   • Catch Me If You CAN - The most realistic Mario Kart 64 simulator ever
   • 하드웨어 마이크 연결 해제
   • So powerful, it's kind of ridiculous
   • Van Gogh, AMD’s Steam Deck APU
   • “애플의 탐욕 vs. 보안 강화” MFi 인증 USB-C 케이블 논란
   • Meet the $10,000 Nvidia chip powering the race for A.I.
3. 블록체인과 메타버스
   • [remark] Privatizing our digital identities
   • 디지털 트윈이 메타버스보다 2배는 중요한 이유
4. 읽을거리
   • 실리콘 밸리 은행 영업 중단?
   • The Collapse of SVB Exposes the Largest Crack in the Economy.
   • The Demise of Silicon Valley Bank
   • Smartphones and social media are destroying children’s mental health
   • 2023년 1월 국제수지(잠정) by 한국은행
   • [기획] 아카데미 9개 부문 후보 ‘서부 전선 이상 없다’, 과거의 영광은 재현될 것인가?
   • A Philosophy Professor’s Final Class
   • A top audio engineer explains NPR’s signature sound
   • 전세 제도에 대하여: A Minority View.(민간 사이에 일어나는 환매조건부채권매매로 설명)
   • 토양의 지력을 굉장히 빨리 고갈시키기로 유명한 구황 작물.jpg
   • Satellites like SpaceX’s Starlink are disrupting Hubble observations
   • The dogs of Chernobyl: Demographic insights into populations inhabiting the nuclear exclusion zone
   • 📚세이노의 가르침 PDF/전자책 무료 배포(2023년판)
   • The most dangerous substance known to man

보너스: 하놔 엄청옛날 포켓중국어여행회화 라는 책 펼쳐봤는데 성조 설명이 웃김 상대방이 당치도 않은 말을 하고 있을 때 네 하면서 대답하는 성조가 2성 ㅠㅠㅋㅋㅌㅋㅌㅋㅌㅋㅋvia @o121011

EOB