일 | 월 | 화 | 수 | 목 | 금 | 토 |
---|---|---|---|---|---|---|
1 | 2 | 3 | 4 | 5 | ||
6 | 7 | 8 | 9 | 10 | 11 | 12 |
13 | 14 | 15 | 16 | 17 | 18 | 19 |
20 | 21 | 22 | 23 | 24 | 25 | 26 |
27 | 28 | 29 | 30 |
- 회귀
- LG Aimers 4th
- LG Aimers
- 오블완
- OpenAI
- 딥러닝
- 티스토리챌린지
- Classification
- 분류
- AI
- supervised learning
- 해커톤
- GPT-4
- 머신러닝
- PCA
- 지도학습
- LLM
- regression
- gpt
- ChatGPT
- LG
- Machine Learning
- deep learning
- Today
- Total
목록4학년 1학기 전공 (11)
SYDev

경희대학교 허선영 교수님의 컴파일러 수업을 기반으로 정리한 글입니다. 1. Front-end Compilation- Lexical Analysis문장을 토큰 단위로 분류하는 과정정규 표현식을 사용하여 문장을 의미 단위로 구분- Syntax AnalysisLexical Analysis 과정에서 생성된 토큰을 기반으로 parse tree를 생성- Semantic Analysis프로그램의 선언과 진술이 의미론적으로 정확한지 확인하는 작업syntax analysis와 달리 type checking 진행 2. Syntax AnalysisAbstract Syntax Tree(AST)를 빌드 -> token stream이 valid한지 체크if the token stream is not valid -> syntax ..

경희대학교 허선영 교수님의 컴파일러 수업을 기반으로 정리한 글입니다. Goals of this AssignmentBuild a lexical analyzer for a new language, using a lexer generator for OCaml(ocamllex)know how to write a regular expression to represent a given set of stringsUnderstand how to write and debug ocamllex specification ocamllexocamllex: lexer generator for OCamlocamllex specification consists of three parts as shown in the example in..

PA-LSTMhttps://www.sciencedirect.com/science/article/pii/S0263224121009969PEMS(Portable Emission Measurement Systme): 이동식 배출가스 측정 장비로, 운행 중 발생되는 배출가스 측정을 위하 사용하는 배기가스 인증규제 시험절차의 제도적 보완장치장기 측정 시에 drift 및 비용 문제로 활용에 제한이 있음vehicle emissions는 일반적인 시계열 데이터 -> time-series prediction model 적용 가능이런 PEMS의 한계를 보완하고자, 배출가스의 사전 예측 및 제어 전략 반영이 가능한 예측 모델로 Parallel Attention-based LSTM 제안모델 선정 과정traditional pr..

LSTM(Long Short-Term Memory): RNN(Recurrent Neural Network)의 한 종류로, RNN의 장기 의존성 문제(long-term dependencies)를 해결하기 위해서 나온 모델직전 데이터 뿐만 아니라, 좀 더 거시적으로 과거 데이터를 고려하여 미래 데이터를 예측-> 6개의 파라미터와 4개의 게이트로 구성 Cell State작은 linear interaction만을 적용시키면서 전체 체인을 계속 구동정보가 전혀 바뀌지 않고 그대로만 흐르게 하는 부분Gate라 불리는 구조에 의해서 정보 추가 or 제거Gate는 Training을 통해서 어떤 정보를 유지하고 버릴지 학습 Forget Gate1이면 보존, 0이면 폐기previous Cell State에서 어떤 정보를 ..

경희대학교 최진우 교수님의 기계학습 수업을 기반으로 정리한 글입니다.

경희대학교 최진우 교수님의 기계학습 수업을 기반으로 정리한 글입니다. 1. ML Definitions 명시적으로 프로그램하지 않아도, 기계가 스스로 무언가 배울 수 있는 능력을 가지도록 연구하는 분야Machine Learning: 기계가 Experience E에 따라서, 특정 Task T를 Performance P만큼의 성능을 수행할 수 있도록 학습시키는 것ex) 손글씨 인식Task T: 손글씨 사진이 어떤 숫자인지 구별하는 것Performance measure P: 100개 중에서 몇 개를 정확히 인식하는지Training experience E: 데이터 2. ML TypesSupervised Learning: 특징 벡터(input) X와 목푯값(output) Y가 모두 주어진 상황회귀 문제(Regres..

Time Series?Time Series: In mathematics, a time series is a series of data points indexed (or listed or graphed) in time orderthus it is a sequence of discrete-time data간격이 동일하지 않으면 Time Series Data X Characteristics of Time Series DataA time series ccan be stationary or non-stationaryStationary Time Series: has statistical properties that are constant over timestatistics like the mean, variance..

경희대학교 허선영 교수님의 컴파일러 수업을 기반으로 정리한 글입니다. OCaml(Objective Caml)Functional Programming Language: 함수의 구성 & 추가로 프로그램이 설계되는 Functional Programming 방식을 사용하는 언어Functional programming language에서 function은 인자로서 전달되고, 자료구조에 저장되고, function calls의 결과로 반환된다. OCaml은 정적 타입 언어 -> 변수와 함수의 타입이 Compile time에 결정됨한 번 변수를 바인딩하면 변경 불가 x = x + 1 형태로 변경 불가값을 바꾸고 싶으면 새롭게 리바인딩해야함 let x = x + 1Syntax- Data Types - Operators-..