Introduction to Modern AI
Artificial Intelligence:
What is Artificial Intelligence (AI) ?
The Artificial intelligence is made up of 2 words : "artificial" and "intelligence"
Artificial : Man made things.
Intelligence: Ability to think and process.
Therefore, artificial intelligence is a man made ability to think on a particular task and give the output similar to a human.
Therefore, artificial intelligence is a man made ability to think on a particular task and give the output similar to a human.
Machine Learning:
What is machine learning (ML)?
Machine learning is a subfield of AI that allows a computer (program) learn from the data and improve tasks over time without being dynamically programmed at every step.
There are multiple types of AI tools that are commonly used in our daily life, such as Chatgpt, Gemini, Claude etc.
An AI tool can be used in various ways according to the purpose it serves.
An AI tool can be used in various ways according to the purpose it serves.
Computer Vision
It is a function of AI that helps to detect the object from an image.
How does it works :
How does it works :
1. Object Detection: This process is done my using to functions i.e :
a) Object localization: locate an object with a bounding box (a rectangle).
b) Image classification: identify the category (a text label) that an image belongs to.
2. Similarity Search: This is the process where when an image is searched through the AI lens, it shows all the results similar to the image as the AI has no any additional information given to search on the image.
Therefore, the output can be not related to the image sometime.
3. Image Segmentation: In this process the image is segregated by pixels for each object on the image, this helps the AI to detect the objects easily and it can provide better results.
4. Optical Character Recognition (OCR): Used to detect and extract text from the image. The use case of the OCR can be seen when we search any text by the help of Google lens, the text on the image becomes copiable which is helpful to us.
Machine Translation
Machine translation is a method used to understand the text (word) by the help of the condition the is mentioned in the sentence.
eg: the word "park" can be used in 2 different ways
1. I parked my car on the parking area.
2. I go to park in the evening.
Here, the word the "park" has 2 different meanings, so the machine detects the word according to the meaning it has in the sentence.
Machine Translation through context: A machine (eg: Google Translator) translates the text from one language to the other by the context it receives.
Chatbot
A chatbot is an AI tool where the user can chat with a bot (AI) which replies to the user in a human like way, the chatbot gives the answer according to the context it receives from the user.
eg: ChatGPT, Claude, Gork AI etc.
What can you do with chatbots ?
- Expand: Answer questions, generate ideas
- Synthesize: summarize, label
- Transform: translate, reword or rewrite
- Evaluate: fact check, grade
- Chat: dialogue with a chatbot
- Take actions: set a calendar event, search the internet
- Input images and text: Ask a question about a picture
The chatbot predicts one word at a time and record that word with the sentence and predicts again.
That chatbot stores the information in a memory, each chat that is done with the bot is being stored in its memory. This helps the bot to give relevant answer to the user related to the context.
If a user combines multiple topics on a single chat, the AI might not give relevant answers to the questions, this happens because the AI is programmed to implement one task at a time, so when the context changes on a single task the bot might not answer properly.
If a user combines multiple topics on a single chat, the AI might not give relevant answers to the questions, this happens because the AI is programmed to implement one task at a time, so when the context changes on a single task the bot might not answer properly.
eg: Open a new chat in a chatbot and ask a random question.
"what smartphone has the best features ?"
The AI will provide you the results, after the results ask the bot something different than the context
"How far is Madrid from Barcelona?"
The chatbot will answer the question but relevant to the initial question asked.
The chatbot will answer the question but relevant to the initial question asked.
Therefore, whenever the user needs to start a new conversation or change the context of the topic, it is better to start a new chat with the chatbot.
Use Multiple Chatbots for different tasks.
An AI tool cannot perform all the task by itself, each AI has its own features and limitations. There are multiple chatbots that can be used for different tasks.
Gemini: https://gemini.google.com/app
Claude: https://claude.ai/chats
Copilot: https://copilot.microsoft.com/
Mistral: https://chat.mistral.ai/chat
Testing each AI tool is beneficial to us, as we will know which AI tools are best for the particular task.
How to write prompts ?
Prompts are the initial information or queries that we give to the AI. This information tells the AI what and how to perform the task.
Therefore, the prompts which give plays a very important role to perform the task.
Comments
Post a Comment