AI track and course roadmap

• AI Skill Development Roadmap - Project Phase
  • Specialized Tracks (Focus Areas)
  • Team Projects (팀 프로젝트)
  • Individual Missions (개인 미션)
    • AI 응용 (AI Application)
    • 고급 자연어 처리 (Advanced NLP)
    • 이미지 생성 (Image Generation)
    • 텍스트 생성 (Text Generation)
  • Key Features of This Roadmap
  • Implementation Strategy
• Detailed Skill Map Analysis
  • Skill Map Breakdown
    • Phase 1: Core NLP Skills (Left Column)
    • Phase 2: Advanced NLP/ML Skills (Middle Column)
    • Phase 3: Applied Projects (Right Column)
  • Skill Map Summary

AI Skill Development Roadmap - Project Phase

This detailed roadmap provides a comprehensive guide for AI skill development, focusing specifically on the Project phase and Advanced Individual Missions. The roadmap identifies several specialized tracks and details team projects that align with those tracks.

Specialized Tracks (Focus Areas)

• 콘텐츠 생성 및 자동화 집중형 (Content Generation and Automation Focus)
• 데이터 분석 및 시각화 집중형 (Data Analysis and Visualization Focus)
• 생성형 AI 적용 SDLC 집중형 (Generative AI Application in SDLC Focus)
• 자연어 처리 및 응용 집중형 (Natural Language Processing and Application Focus)
• 생성형 AI 응용 서비스 개발 집중형 (Generative AI Application Service Development Focus)

Team Projects (팀 프로젝트)

The projects are grouped by number, with each containing a Team-Only (팀단독형) mission and several Team-Linked (팀연계형) missions.

Individual Missions (개인 미션)

This section details advanced individual learning modules covering various AI applications and techniques.

AI 응용 (AI Application)

고급 자연어 처리 (Advanced NLP)

이미지 생성 (Image Generation)

텍스트 생성 (Text Generation)

Key Features of This Roadmap

1. Progressive Difficulty: Projects are structured with increasing complexity from basic text generation to advanced conversational AI systems.

2. Practical Focus: Each mission includes hands-on implementation with specific technologies and frameworks.

3. Team Collaboration: Team projects emphasize collaborative development and real-world application scenarios.

4. Individual Growth: Personal missions allow for specialized skill development in specific AI domains.

5. Technology Stack: Covers modern AI technologies including:

   • Deep Learning: PyTorch, Keras, TensorFlow
   • NLP: BERT, GPT-2, BART, NLTK
   • Computer Vision: OpenCV, GANs, Autoencoders
   • Data Science: SHAP, t-SNE, Bayesian Optimization
   • Development Tools: Figma, APIs, Graph Neural Networks

Implementation Strategy

1. Assessment: Evaluate current skill level and choose appropriate starting point
2. Track Selection: Select specialized track based on career goals and interests
3. Progressive Learning: Complete individual missions to build foundational skills
4. Team Collaboration: Participate in team projects for practical experience
5. Continuous Improvement: Apply optimization and analysis techniques to ongoing projects

Note: Estimated time units and difficulty ratings are based on the original roadmap specifications. Time units likely represent abstract learning points or hours of dedicated study and practice.

Detailed Skill Map Analysis

This image appears to be a skill map or a learning path for an AI/ML developer, likely focusing on Natural Language Processing (NLP) and Computer Vision. The diagram is structured into sections based on themes, with individual “cards” representing skills, projects, or learning modules.

The following provides a text-based, numbered breakdown of the diagram. The structure flows from left to right in phases or themes. The labels are primarily in Korean, with core concepts translated below.

Skill Map Breakdown

The diagram is organized into major thematic sections (like “Input Preparation,” “Image,” “Beyond Basic,” and “GPT-3”) and project columns (labeled “Project 1,” “Project 2,” etc.).

Phase 1: Core NLP Skills (Left Column)

This section focuses on foundational and initial skill building in NLP and related areas.

1. 입문 미션 1 (Intro Mission 1):
   • Content: Applying a dictionary using NLTK to train and improve the performance of a basic natural language processing model.
   • Est. Time/Cost: 150 (Units not specified, likely hours or points).
   • Rating/Difficulty: 3 stars.
2. 입문 미션 2 (Intro Mission 2):
   • Content: Developing a basic NLP system that applies techniques like Word2Vec and TF-IDF to understand and infer the meaning of a given text.
   • Est. Time/Cost: 200.
   • Rating/Difficulty: 3 stars.
3. 이미지 미션 (Image Mission) / 입문 미션 3 (Intro Mission 3):
   • Content: Using Keras to implement a deep learning model for image classification (e.g., using ResNet or VGG).
   • Est. Time/Cost: 300.
   • Rating/Difficulty: 4 stars.
4. 입문 미션 4 (Intro Mission 4):
   • Content: Developing and implementing a simple computer vision program using OpenCV to experience basic image processing techniques.
   • Est. Time/Cost: 250.
   • Rating/Difficulty: 3 stars.
5. 고급 지식 배우기 (Learn Advanced Knowledge) / Mission:
   • Content: Applying PyTorch to develop a basic deep learning model for NLP and experience text preprocessing and learning results.
   • Est. Time/Cost: 300.
   • Rating/Difficulty: 3 stars.
6. Mission (Continuing Advanced Knowledge):
   • Content: Developing a basic QA (Question-Answering) system using the BERT model.
   • Est. Time/Cost: 350.
   • Rating/Difficulty: 4 stars.

Phase 2: Advanced NLP/ML Skills (Middle Column)

This column introduces more advanced model types and concepts, bridging core skills with practical projects.

1. Mission (Related to Intro Mission 1):
   • Content: Developing an advanced text classification model by applying techniques like LSTM and Attention mechanisms.
   • Est. Time/Cost: 300.
   • Rating/Difficulty: 4 stars.
2. Mission (Related to Intro Mission 2):
   • Content: Developing a Transformer model-based information retrieval system and evaluating its performance.
   • Est. Time/Cost: 350.
   • Rating/Difficulty: 5 stars.
3. Mission (Related to Intro Mission 3/4):
   • Content: Developing an advanced computer vision model by implementing object detection/segmentation using models like Mask R-CNN or YOLO.
   • Est. Time/Cost: 350.
   • Rating/Difficulty: 4 stars.
4. Mission (Related to Advanced Knowledge):
   • Content: Developing a conversational dialogue system through fine-tuning of pre-trained models.
   • Est. Time/Cost: 300.
   • Rating/Difficulty: 4 stars.
5. Mission (Related to Advanced Knowledge):
   • Content: Developing a text summarization and generation system by utilizing the BART or T5 model structure.
   • Est. Time/Cost: 400.
   • Rating/Difficulty: 5 stars.
6. Mission (GPT-3 API/Beyond Basic):
   • Content: Developing a web service or mobile app by applying GPT-3 API using a language like C# or Python.
   • Est. Time/Cost: 350.
   • Rating/Difficulty: 4 stars.
7. Mission (GPT-3 API/Beyond Basic):
   • Content: Developing a model that generates and visualizes various images based on a user’s prompt (e.g., using DALL-E or Stable Diffusion).
   • Est. Time/Cost: 400.
   • Rating/Difficulty: 5 stars.

Phase 3: Applied Projects (Right Column)

This column represents practical application through development projects, likely to be completed after mastering the skills in the left and middle columns.

1. 팀 프로젝트 1 (Team Project 1) / Project 1 (Related to Intro Mission 1/Middle Mission 7):
   • Content: Developing a natural language classification model for a real-world problem and improving its prediction accuracy.
   • Est. Time/Cost: 500.
   • Rating/Difficulty: 4 stars.
2. Project 1 (Continuing from 14):
   • Content: Developing a real-time data processing system by applying the developed language model (from 14) to a data pipeline.
   • Est. Time/Cost: 450.
   • Rating/Difficulty: 4 stars.
3. Project 1 (Continuing from 15):
   • Content: Developing a recommendation system by combining the language model with traditional machine learning techniques (like Collaborative Filtering).
   • Est. Time/Cost: 500.
   • Rating/Difficulty: 4 stars.
4. 팀 프로젝트 2 (Team Project 2) / Project 2 (Related to Intro Mission 3/4/Middle Mission 9):
   • Content: Developing a computer vision-based real-time object detection and analysis system for smart cities or smart factories.
   • Est. Time/Cost: 400.
   • Rating/Difficulty: 4 stars.
5. Project 2 (Continuing from 17):
   • Content: Developing a web-based service using Figma and the computer vision model to easily verify results.
   • Est. Time/Cost: 500.
   • Rating/Difficulty: 4 stars.
6. 팀 프로젝트 3 (Team Project 3) / Project 3 (Related to Middle Mission 10/11):
   • Content: Developing a web application that improves efficiency by applying the BART/T5 model to various internal documents.
   • Est. Time/Cost: 450.
   • Rating/Difficulty: 4 stars.
7. 팀 프로젝트 4 (Team Project 4) / Project 4 (Related to Middle Mission 12/13):
   • Content: Developing a new service by utilizing the GPT model to automate internal document summarization and analysis.
   • Est. Time/Cost: 500.
   • Rating/Difficulty: 4 stars.
8. Project 4 (Continuing from 20):
   • Content: Developing a service that sequentially summarizes and analyzes various documents to extract key information using the GPT model.
   • Est. Time/Cost: 500.
   • Rating/Difficulty: 4 stars.

Skill Map Summary

This map provides a clear progression from foundational NLP/CV techniques to advanced models (BERT, Transformer, GPT) and finally to real-world, team-based development projects. The structure emphasizes:

• Foundational Learning: Starting with basic NLP and computer vision concepts
• Progressive Skill Building: Moving from simple implementations to complex model architectures
• Practical Application: Culminating in comprehensive team projects that integrate multiple skills
• Technology Integration: Covering the full spectrum from traditional ML to modern transformer architectures

The following wiki, pages and posts are tagged with likelion

{# nothing on index to avoid visible raw text #}