- This list of submodules of all the projects done / in progress for AI Algorithm branch in school
42
Project name | Description | Score |
---|---|---|
Python Module | Introductory Python project covering topics like Python syntax, creating packages, utilizing data science libraries, and concluding with K-means clustering. | 100/100 |
ML Module | Series of projects involving linear regression and logistic regression implementation from scratch, based on the knowledge from the Python Module. The project involves training, predicting, evaluating, and an understanding of model evaluation metrics. | 100/100 |
ft_linear_regression | In-depth implementation of a linear regression model. | 125/100 |
matrix | Project focused on learning linear algebra by implementing Vector and Matrix classes from scratch in Python. Requires understanding of linear algebra laws and spatial concepts of matrices. | 100/100 |
ready-set-boole | Introduction to essentials of Boolean Algebra and Set Theory in both mathematics and computer science, with exercises involving programming techniques, bitwise operations, logical evaluation, truth tables, various rewrite rules, with no external math libraries. | 100/100 |
dslr | In-depth implementation of a logistic regression model. | 125/100 |
multilayer-perceptron | Implementing a multilayer perceptron neural network from scratch, focusing on core concepts like feedforward, backpropagation, and gradient descent, as well as preprocessing data into training and validation sets, and building a modular program to support customizable hidden layers and parameters. | in progress |
lem_in | Elementary algorithm project requiring a solid understanding of the 'graph' data structure to design an algorithm that efficiently tackles optimal pathfinding for network flow and resource management, implementing techniques such as breadth-first search, depth-first search, or Dijkstra's algorithm, Edmonds-Karp algorithm, and disjoint path finding. | 125/100 |
expert-system | A project that implements a backward-chaining inference engine based on an input file containing rules, initial facts, and queries. The engine must determine if each query is true, false, or undetermined. This project supports a variety of logical operators, and rule grouping, and offers additional features such as user interaction and reasoning visualization. | 125/100 |
krpsim | This project requires consideration of various aspects in algorithm design, operational research, artificial intelligence, and industrial applications. The primary objective is to implement an algorithm that finds the most efficient way to produce a specific product or achieve it within the shortest possible time. It involves analyzing interconnected processes within a complex graph chain. To accomplish this goal, the team has utilized reinforcement learning techniques, specifically implementing a Q-agent. | 115/100 |
- 아래 프로젝트 (서브모듈 링크) 들은 '42' 교내에서 진행중인 AI/알고리즘 심화과정의 프로젝트 진행상황과 설명입니다.
프로젝트명 | 설명 | 점수 |
---|---|---|
Python Module | 파이썬 문법, 패키지 생성, 데이터 과학 관련 라이브러리 사용 및 K-means 클러스터링을 다루는 파이썬 입문 과정. | 100/100 |
ML Module | Python Module을 바탕으로 선형 회귀와 로지스틱 회귀를 처음부터 구현하는 일련의 프로젝트. 학습, 예측, 평가 과정을 거치며, 모델 평가 지표에 대한 이해가 필요함. | 100/100 |
ft_linear_regression | 선형 회귀 모델을 깊이 있게 구현하는 프로젝트. | 125/100 |
matrix | 파이썬을 사용하여 벡터와 행렬 클래스를 처음부터 구현하며 선형 대수학 및 행렬의 공간 개념을 학습하는 프로젝트. | 100/100 |
ready-set-boole | 불 대수와 집합 이론의 핵심 개념을 수학과 컴퓨터 과학에 적용하는 프로젝트. 프로그래밍 기술, 비트 연산, 논리 평가, 진리 표, 다양한 재작성 규칙 등을 다루며, 제공 라이브러리를 사용하지 않고 구현함. | 100/100 |
dslr | 로지스틱 회귀 모델을 깊이 있게 구현하는 프로젝트. | 125/100 |
multilayer-perceptron | 다층 퍼셉트론 신경망을 처음부터 구현하는 프로젝트로, 전방향(feed forwarding), 역전파(back propagation), 경사 하강법(gradient descent) 등의 핵심 개념에 집중하며, 학습 및 검증 데이터 세트로 데이터 전처리를 진행하고, 맞춤형 은닉층(hidden layer)과 매개 변수를 지원하는 모듈화된 프로그램 구조를 구축 하여야 함. | in progress |
lem_in | 그래프 데이터 구조에 이해가 필요한 기초 알고리즘 프로젝트로, 최적의 경로 찾기와 자원 관리를 효율적으로 다루는 최대 유량 알고리즘을 응용 설계하며, 너비 우선 탐색, 깊이 우선 탐색, 및 다익스트라 알고리즘, 에드몬드-카프, 서로소 경로 찾기(disjoint path finding) 등의 지식이 필요함. | 125/100 |
expert-system | 입력 파일을 기반으로 규칙, 초기 사실, 질의를 처리하여 각 질의가 참인지, 거짓인지, 또는 결정되지 않은 상태인지를 판별하는 역추론 엔진(Backward chaining inference engine)을 구현하는 프로젝트. 다양한 논리 연산자와 묶음 처리를 지원하며, 사용자와 상호작용하거나 추론 과정을 시각화하는 추가 기능 포함. | 125/100 |
krpsim | 알고리즘 설계, 운영 연구, 인공 지능, 그리고 산업 응용 분야의 다양한 측면을 고려하여야 하는 프로젝트로, 주요 목표는 특정 제품을 가장 효율적으로 생산하거나 가능한 한 짧은 시간 내에 이를 달성하는 방법을 찾는 알고리즘을 구현 하여야 함. 복잡한 그래프 체인에서 상호 연결된 프로세스를 조사/분석 하여야 하며, 팀은 이 목표를 달성하기 위해 강화 학습 기술을 활용하였으며, 구체적으로는 Q-에이전트를 구현하였음. | 115/100 |