DOCSEARCHbyAlgolia
Submit Docs

pluinyiasnhg Knowledge Base

website·Indexed 31 days ago

https://pluinyiasnhg.top

A personal technical knowledge base and learning log covering a wide range of topics including Python basic syntax and object-oriented programming, PyTorch for NLP, MySQL installation and SELECT statements, Arch Linux installation, Docker, and JavaScript. The site also includes notes on academic writing, Markdown, and various tool configurations like Emacs and Goldendict.

Tokens
89.1K
Snippets
458
Records
624
Agent score
49%

What's inside pluinyiasnhg

  1. Overview of MySQL Data Types

    MySQL provides a variety of data types to handle different kinds of information. Key categories include:

    • Integer Types: TINYINT, SMALLINT, MEDIUMINT, INT (INTEGER), BIGINT
    • Floating-Point Types: FLOAT, DOUBLE
    • Fixed-Point Type: DECIMAL
    • Bit Type: BIT
    • Date and Time Types: YEAR, TIME, DATE, DATETIME, TIMESTAMP
    • Text String Types: CHAR, VARCHAR, TINYTEXT, TEXT, MEDIUMTEXT, LONGTEXT
    • Enum/Set Types: ENUM, SET
    • Binary String Types: BINARY, VARBINARY, TINYBLOB, BLOB, MEDIUMBLOB, LONGBLOB
    • JSON Type: JSON objects and arrays
    • Spatial Data Types: GEOMETRY, POINT, LINESTRING, POLYGON, and their collection counterparts.

  2. Overview of specialized MySQL storage engines

    Beyond InnoDB and MyISAM, MySQL provides specialized engines for specific use cases:

    • Archive: Used for data archiving.
    • Blackhole: Discards all write operations; read operations return empty.
    • CSV: Stores data in comma-separated values format.
    • Memory: Stores tables in RAM for high speed, but data is volatile (lost on restart).
    • Federated: Allows access to remote tables on other MySQL servers.
    • Merge: Manages a collection of MyISAM tables.
    • NDB: Specialized engine for MySQL Cluster.

  3. Overview of Java utility classes (System, Runtime, Math, BigInteger, BigDecimal, Random)

    Java provides several utility classes for common system and mathematical operations:

    • System: Provides access to standard input/output/error streams (in, out, err) and methods like currentTimeMillis(), gc(), exit(int status), and getProperty(String property).
    • Runtime: A singleton class representing the runtime environment of the Java process and its memory usage.
    • Math: Contains methods for basic numeric operations (e.g., trigonometry, logarithms).
    • BigInteger & BigDecimal: Used for high-precision arithmetic. BigInteger handles arbitrary-length integers; BigDecimal handles arbitrary-precision floating-point numbers.
    • Random: Used to generate random numbers. nextInt(int bound) returns a random integer within a specified range.

  4. Overview of AI Engineering from Scratch project

    The 'AI Engineering from Scratch' project is a comprehensive learning path consisting of 20 phases and 503 lessons. It covers implementation across four programming languages: Python, TypeScript, Rust, and Julia. Each lesson follows a standardized workflow: understanding the problem, mathematical derivation, coding implementation, running tests, and saving the final artifacts.
    # Project Repository
https://github.com/rohitg00/ai-engineering-from-scratch

  6. Understand JVM memory architecture

    The Java Virtual Machine (JVM) divides memory into several key areas to manage data and execution flow.
    RegionPurpose
    JVM StackStores local variables for each executing Java method. Automatically released after method execution.
    Heap MemoryStores all objects, including array objects (created via new).
    Method AreaStores class information, constants, static variables, and JIT compiled code.
    Native Method StackMemory area used during the execution of native methods.
    Program CounterA register in the CPU containing the address of the next instruction to be executed by the thread.

  7. Understand MySQL system databases

    MySQL comes with four default databases used for internal management:

    • mysql: Stores user accounts, privileges, stored procedures, events, and timezone information.
    • information_schema: Contains metadata about all other databases (tables, views, columns, indexes).
    • performance_schema: Monitors various MySQL performance metrics.
    • sys: A set of views that combine information_schema and performance_schema for easier performance monitoring.

  8. Compare Relational (RDBMS) and Non-Relational (NoSQL) Databases

    Relational databases (RDBMS) store data in tables with rows and columns, using SQL for queries and maintaining relationships between entities. Non-Relational databases (NoSQL) are often high-performance alternatives that use key-value pairs or other structures and are suitable for specific use cases like caching, logging, or complex networks.

    Common NoSQL types include:

    • Key-Value: Fast lookup, used for memory caching (e.g., Redis).
    • Document: Stores XML/JSON documents (e.g., MongoDB).
    • Search Engine: Optimized for full-text indexing (e.g., Elasticsearch).
    • Column-family: Reduces I/O, suitable for distributed systems (e.g., HBase).
    • Graph: Manages complex relationships between entities (e.g., Neo4j).

  9. Compare RNN Architectures: Single-layer, Multi-layer, and Bidirectional

    PyTorch RNNs can be implemented in several structural variations:

    • Single-layer: A basic RNN where the same weights (and tanh activation) are shared across all time steps of the sequence.
    • Multi-layer: Stacked RNNs where the output of one layer serves as the input to the next. The x-axis represents time steps and the y-axis represents the layer depth.
    • Bidirectional: Processes the sequence in both forward and backward directions. This allows the model to capture both preceding and succeeding context. The forward and backward layers are typically trained in parallel and their outputs are concatenated (concat) to serve as input for subsequent layers.

  11. Understand the Template Method Design Pattern

    The Template Method pattern uses an abstract class to define the skeleton of an algorithm, providing a common template for multiple subclasses. The parent class implements the fixed steps of the algorithm, while exposing variable parts as abstract methods for subclasses to implement or override. This is used when the overall process is constant, but specific steps differ between implementations.

  12. Structure and write the thesis Introduction (绪论)

    The Introduction should be between 3,000 and 5,000 words and include:

    1. Background, Purpose, and Significance: Reasons for the research, including national funding projects, hot topics, and consensus issues.
    2. Current State of Research: A review of domestic and international progress. Categorize by sub-field if multiple exist. Identify gaps or deficiencies in previous work. This section typically contains the most citations.
    3. Research Content and Organization: Define the problem, break the solution into stages/chapters, and provide a visual diagram (e.g., "Research Framework" or "Main Work Schematic") to show the relationship between chapters.