UML creation scripting (similar to LaTex)

• https://plantuml.com/
• plantuml in VS' marketplace : https://marketplace.visualstudio.com/items?itemName=jebbs.plantuml
• try the following code in this editor https://editor.plantuml.com

@startuml

Alice -> Bob : hello

@enduml

Multilingual AI detector (not free)

https://app.originality.ai/home/scan 

It supports Thai language.

Call number codes

 

Superlinear, Linear, Sublinear

Superlinear คือ slop เพิ่มขึ้นเรื่อยๆ (โตเร็วกว่า linear)

Sublinear คือ slope ลดลงเรือยๆ (โตช้ากว่า linear)

Example:

Settrade streaming pro

 

Speech recognition model by OpenAI

Whisper is a general-purpose speech recognition model. It is trained on a large dataset of diverse audio and is also a multitasking model that can perform multilingual speech recognition, speech translation, and language identification.

https://github.com/openai/whisper

Major types of recommendation systems

Collaborative Filtering (finding similar users/items based on past behavior), Content-Based Filtering (recommending items with similar features to what a user liked), and Hybrid Systems (combining both for better results). Other advanced types include Knowledge-Based, Deep Learning, and Demographic systems, which leverage different data points like item attributes, complex patterns, or user demographics to provide personalized suggestions. 

Collaborative Filtering (CF):

• Concept: "People who liked X also liked Y." It finds patterns in user-item interactions (ratings, purchases).
• Sub-types: User-based (If User A buys Item A, and a neighboring User B (who is found to be highly similar to User A based on their shared past interactions like rating or purchases) also buy Item B but User A has not yet bought it, then the system should recommend Item B to User A.) and Item-based (If a user likes item A, and many users who liked item A also liked item B, then the system should recommend item B to that user.).
• Techniques: Matrix Factorization (like SVD), Nearest Neighbors.

Content-Based Filtering:

• Concept: Recommends items similar to those a user has liked before, based on item features (e.g., movie genre, director, actors). If a user liked item A, they will also like item B if item B shares many of the same features as item A.
• How it works: Builds a user profile from features of liked items, then matches it to other items based on similarity. The features can also include those of the user like age and gender.

Hybrid Recommendation Systems:

• Concept: Merges CF and Content-Based methods, or other techniques, to overcome individual limitations (like the cold-start problem). 
  

--Gemini

LightFM is a Python implementation of a number of popular recommendation algorithms for both implicit and explicit feedback.

LightFM can solve cold-start problem: For a new user (cold-start user), LightFM uses their provided features (e.g., "age: 25, gender: female") to do content-based filtering to recommend items that share similar features with the user's profile. As the cold-start user interacts with items, the model gradually updates their information, shifting from pure content-based to a more personalized collaborative prediction.

https://making.lyst.com/lightfm/docs/home.html#examples

 

Google AI Studio vs Colab (answered by Gemini)

 

Daysie AI-Edge PaaS by Nectec

https://daysie.io/

https://www.conferencethaiseries.com/wp-content/uploads/2025/10/%E0%B8%9B%E0%B8%B2%E0%B8%90%E0%B8%81%E0%B8%96%E0%B8%B2-6-The-Rise-of-Edge-Computing-AI-Beyond-the-Cloud-%E0%B9%82%E0%B8%94%E0%B8%A2-%E0%B8%94%E0%B8%A3.%E0%B8%9E%E0%B8%99%E0%B8%B4%E0%B8%95%E0%B8%B2-%E0%B8%9E%E0%B8%87%E0%B8%A9%E0%B9%8C%E0%B9%84%E0%B8%9E%E0%B8%9A%E0%B8%B9%E0%B8%A5%E0%B8%A2%E0%B9%8C.pdf

infrastructure as code

 AWS Cloudformation

Terraform (https://developer.hashicorp.com/terraform/tutorials/aws-get-started/infrastructure-as-code)

Google Optimization Tool

OR-Tools is an open source software suite for optimization, tuned for tackling the world's toughest problems in vehicle routing, flows, integer and linear programming, and constraint programming.

https://developers.google.com/optimization

Online learning vs Offline learning

Online learning (or online machine learning) is a method in machine learning where the model continuously learns from a sequential stream of data, updating its parameters incrementally with each new data instance or small batch of data.

It is a dynamic process that allows a model to adapt to new patterns and changes in the data distribution in real-time.

Online learning is the opposite of the more traditional Batch Learning (or Offline Learning) approach.

Low code RAG development tool

Workflow automation tool

https://n8n.io/ 

AI builder tool

https://www.langflow.org/

Evaluate RAG

https://towardsdatascience.com/evaluating-your-rag-solution/

Notebooklm is RAG as we can add files to let users ask anything about the files.

PR-AUC

You can use this plot to make an educated decision when it comes to the classic precision/recall dilemma. Obviously, the higher the recall, the lower the precision. Knowing at which recall your precision starts to fall fast can help you choose the threshold and deliver a better model.

https://neptune.ai/blog/f1-score-accuracy-roc-auc-pr-auc

The precision recall curve is a handy plot to showcase the relationship and tradeoff between precision recall values as we adjust the decision threshold of the classifier. What is the decision threshold? The decision threshold, also called the classification threshold, is a cutoff point used in binary classification to convert the probability score output by a machine learning model into a final class prediction (positive or negative). Most binary classification models (like logistic regression) output a probability between 0 and 1 that an instance belongs to the positive class. The decision threshold determines which probability values map to which class: If the predicted probability is greater than or equal to the threshold, the instance is classified as the positive class. If the predicted probability is less than the threshold, the instance is classified as the negative class. How it Works By default, the threshold is often set to 0.5. A probability of \ge 0.5 \rightarrow Positive Class A probability of < 0.5 \rightarrow Negative Class However, this default isn't always optimal. The threshold is a hyperparameter that can be tuned to balance the trade-off between precision and recall, which is what the precision-recall curve helps to visualize. Threshold and Precision/Recall Trade-off Adjusting the decision threshold directly impacts the number of false positives (FP) and false negatives (FN), which in turn changes the precision and recall values.

A higher AUC-PR value signifies better performance, with a maximum value of 1 indicating perfect precision and recall trade-off. https://www.superannotate.com/blog/mean-average-precision-and-its-uses-in-object-detection

Agentic AI vs AI Agent

The primary difference is that AI Agents are individual tools that execute pre-defined tasks with limited autonomy, while Agentic AI is a broader concept representing the use of autonomous systems that can independently set goals, make real-time decisions, adapt, and collaborate to solve complex, dynamic problems. Think of AI agents as specific tools or employees, and agentic AI as the system or project manager coordinating them to achieve a larger, more complex goal.  

Stochastic Gradient Descent

• Gradient Descent (Batch): You take a step in the steepest downhill direction. To find the steepest direction, you have to survey the slope of the entire landscape (the entire dataset) before taking each single step. This is accurate but very slow if the landscape is vast (a huge dataset).Stochastic 
• Gradient Descent (SGD): Instead of surveying the entire landscape, you just pick one random spot on the landscape and measure the slope there. You then take a small step in that single spot's steepest downhill direction. You repeat this process many times, picking a new random spot for each step.