Author: lechuck park

AI Infrastructure Architect & Technical Visualizer "Complex Systems, Simplified. I translate massive AI infrastructure into visual intelligence." I love to learn computer tech and help people by the digital.

Data is the next of the AI

Posted on by lechuck park

Data is the backbone of AI’s evolution.

Summary 🚀

  1. High-quality data is the key to the AI era.
    • Infrastructure has advanced, but accurate and structured data is essential for building effective AI models.
    • Garbage In, Garbage Out (GIGO) principle: Poor data leads to poor AI performance.
  2. Characteristics of good data
    • High-resolution data: Provides precise information.
    • Clear labeling: Enhances learning accuracy.
    • Structured data: Enables efficient AI processing.
  3. Data is AI’s core competitive advantage.
    • Domain-specific datasets define AI performance differences.
    • Data cleaning and quality management are essential.
  4. Key messages
    • “Data is the backbone of AI’s evolution.”
    • “Good data fuels great AI!”

Conclusion

AI’s success now depends on how well data is collected, processed, and managed. Companies and researchers must focus on high-quality data acquisition and refinement to stay ahead. 🚀

With ChatGPT

GPU vs NPU on Deep learning

Posted on by lechuck park

This diagram illustrates the differences between GPU and NPU from a deep learning perspective:

GPU (Graphic Process Unit):

  • Originally developed for 3D game rendering
  • In deep learning, it’s utilized for parallel processing of vast amounts of data through complex calculations during the training process
  • Characterized by “More Computing = Bigger Memory = More Power,” requiring high computing power
  • Processes big data and vectorizes information using the “Everything to Vector” approach
  • Stores learning results in Vector Databases for future use

NPU (Neuron Process Unit):

  • Retrieves information from already trained Vector DBs or foundation models to generate answers to questions
  • This process is called “Inference”
  • While the training phase processes all data in parallel, the inference phase only searches/infers content related to specific questions to formulate answers
  • Performs parallel processing similar to how neurons function

In conclusion, GPUs are responsible for processing enormous amounts of data and storing learning results in vector form, while NPUs specialize in the inference process of generating actual answers to questions based on this stored information. This relationship can be summarized as “training creates and stores vast amounts of data, while inference utilizes this at the point of need.”

With Claude

AI in the data center

Posted on by lechuck park

AI in the Data Center

This diagram titled “AI in the Data Center” illustrates two key transformational elements that occur when AI technology is integrated into data centers:

1. Computing Infrastructure Changes

  • AI workloads powered by GPUs become central to operations
  • Transition from traditional server infrastructure to GPU-centric computing architecture
  • Fundamental changes in data center hardware configuration and network connectivity

2. Management Infrastructure Changes

  • Increased requirements for power (“More Power!!”) and cooling (“More Cooling!!”) to support GPU infrastructure
  • Implementation of data-driven management systems utilizing AI technology
  • AI-based analytics and management for maintaining stability and improving efficiency

These two changes are interconnected, visually demonstrating how AI technology not only revolutionizes the computing capabilities of data centers but also necessitates innovation in management approaches to effectively operate these advanced systems.

with Claude

Operation with LLM

Posted on by lechuck park

This image is a diagram titled “Operation with LLM,” showing a system architecture that integrates Large Language Models (LLMs) with existing operational technologies.

The main purpose of this system is to more efficiently analyze and solve various operational data and situations using LLMs.

Key components and functions:

  1. Top Left: “Monitoring Dashboard” – Provides an environment where LLMs can interpret image data collected from monitoring screens.
  2. Top Center: “Historical Log & Document” – LLMs analyze system log files and organize related processes from user manuals.
  3. Top Right: “Prompt for chatting” – An interface for interacting with LLMs through appropriate prompts.
  4. Bottom Left: “Image LLM (multimodal)” – Represents multimodal LLM functionality for interpreting images from monitoring screens.
  5. Bottom Center: “LLM” – The core language model component that processes text-based logs and documents.
  6. Bottom Right:
    • “Analysis to Text” – LLMs analyze various input sources and convert them to text
    • “QnA on prompt” – Users can ask questions about problem situations, and LLMs provide answers

This system aims to build an integrated operational environment where problems occurring in operational settings can be easily analyzed through LLM prompting and efficiently solved through a question-answer format.

With Claude

The Optimization of Parallel Works

Posted on by lechuck park

The image illustrates “The Optimization of Parallel Works,” highlighting the inherent challenges in optimizing parallel processing tasks.

The diagram cleverly compares two parallel systems:

  • Left side: Multiple CPU processors working in parallel
  • Right side: Multiple humans working in parallel

The central yellow band emphasizes three critical challenges in both systems:

  • Dividing (splitting tasks appropriately)
  • Sharing (coordinating resources and information)
  • Scheduling (timing and sequencing activities)

Each side shows a target/goal at the top, representing the shared objective that both computational and human systems strive to achieve.

The exclamation mark in the center draws attention to these challenges, while the message at the bottom states: “AI Works is not different with Human works!!!!” – emphasizing that the difficulties in coordinating independent processors toward a unified goal are similar whether we’re talking about computer processors or human teams.

The diagram effectively conveys that just as it’s difficult for people to work together toward a single objective, optimizing independent parallel processes in computing faces similar coordination challenges – requiring careful attention to division of labor, resource sharing, and timing to achieve optimal results.

With Claude

Operation with system

Posted on by lechuck park

Key Analysis of Operation Cost Diagram

This diagram illustrates the cost structure of system implementation and operation, highlighting the following key concepts:

  1. High Initial Deployment Cost: At the beginning of a system’s lifecycle, deployment costs are substantial. This represents a one-time investment but requires significant capital.
  2. Perpetual Nature of Operation Costs: Operation costs continue indefinitely as long as the system exists, making them a permanent expense factor.
  3. Components of Operation Cost: Operation costs consist of several key elements:
    • Energy Cost
    • Labor Cost
    • Disability Cost
    • Additional miscellaneous costs (+@)
  4. Role of Automation Systems: As shown on the right side of the diagram, implementing automation systems can significantly reduce operation costs over time.
  5. Timing of Automation Investment: While automation systems also require initial investment during the early phases, they deliver long-term operation cost reduction benefits, ultimately improving the overall cost structure.

This diagram effectively visualizes the relationship between initial costs and long-term operational expenses, as well as the cost optimization strategy through automation.

With Claude