Meet Samantha. She is your intimate assistant from 2025. She can help you organize your mail, set up your meetings, and order your grocery. She will draw and write poetry. She is your best friend. She is the artificial intelligence in the movie "She", which is what we can imagine how Siri will transform our lives into something.
Now, large and small high-tech companies are competing to make this a reality. You will hear these jargon when you look at the news: AI, machine learning, deep learning, neural networks, natural language processing.
Maybe it's all a bit confusing. Therefore, the following is the basic knowledge of these concepts and how they are related.
What is artificial intelligence (AI)?
Simply put, AI is trying to make computers smarter and even smarter than humans. This is to let the computer have human-like behavior, thinking process and reasoning ability.
There are two types of artificial intelligence:
Narrow AI (weak AI)
This AI only focuses on a narrow task. Now we are weak AI everywhere. It has beaten us humans in chess, the TV contest "The Dangerous Edge", and the recent Go game.
Digital assistants like Siri and Cortana can provide us with weather information and self-driving cars on the road. However, they have great limitations. Self-driving cars don't play chess. Siri also cannot read and delete unimportant messages. Weak AI has a narrow scope: it can't go beyond what was first set for it.
Cortana AI is a weak example <br> General AI (strong AI)
Below we enter the kingdom of science fiction. Samantha is undoubtedly defined as a strong AI. She can learn new things and modify her own code base. She can defeat humans in both chess and driving.
Analysis of AI <br> Now that we know that general AI is our ultimate goal, how can we achieve it? Here are 5 areas that need to be proficient:
Perception: Like us humans, a computer also needs five senses to interact with the world. But it is not limited to these five aspects. It can feel like someone doesn't have it. Perspective eye? Sonar detection? All possible.
Natural Language Processing (NLP): Beyond the world of perception, AI needs to understand language and writing. They need to parse sentences and understand the nuances, accents, and meanings between them. The same sentence can have different meanings depending on the context, so the difficulty of this task is well known.
Knowledge expression: Since it can perceive things - objects, people, concepts, words and mathematical symbols - it needs a way to represent the world in their brains.
Reasoning: Once it collects data through its senses and links it to the concept, it can use that data to logically solve the problem. For example, if a chess software detects that a piece of chess moves around the board, it can calculate the strategy to deal with.
Planning and navigation: To be truly human, AI must not only think like humans. You should also live among us. Therefore, one of the big problems for researchers is to help artificial intelligence move and plan the best path in the three-dimensional world. Autonomous vehicles must do this because a mistake can be fatal.
![image description]()
The Singapore Prime Minister took a ride in a self-driving car. Image source: Kenji Soon, MCI.
You can see how these aspects work together, such as machine vision, which uses imaging and image analysis to solve problems. For example, Facebook parses the photos you upload to social networks to suggest who you should tag, and this is fairly accurate.
Autonomous vehicles may be the most complex machine vision processing task available today. It needs to recognize road signs, obey the lanes, and pay attention to vehicles, objects and people. It works well in poor weather conditions, day or night, on dilapidated roads or on new roads.
Implementation Tools <br> These concepts are not new. They were presented at Dartmouth Conferences as early as 1956, a groundbreaking event in the field of artificial intelligence.
It will take decades for technology to keep up with our imagination. We seem to be on the cusp of the AI ​​revolution. With more venture capital investment, more large-scale technology companies are investing in AI research and development. In the coming, we are increasingly using AI in our daily lives.
Important factors that contribute to the rise of AI include Moore's Law, which allows us to inject more computing power into smaller, more efficient chips. Once computing power reaches a certain level, AI will become both practical and cost-effective.
![image description]()
Google made a major breakthrough in the recognition of cats. Image source: zbeads
Big data is another trend that has led to the rise of AI: After Google entered massive amounts of data for neural networks in 2012, it made breakthroughs, including 10 million YouTube video stills.
As a result, the neural network learned to recognize cats without anyone teaching it, achieving 75% accuracy. It is impossible to achieve without this 10 million video database.
When the machine learns <br> Now, let's clarify a few concepts that are often confusing. Machine learning is an AI technology that focuses on learning insights from data and using them to predict the world.
![image description]()
Decision tree algorithm. Image source: Wikipedia
Machine learning has been implemented using algorithms. A task is done through a series of algorithms. Examples of such algorithms include decision tree learning and association rule learning.
Machine learning algorithms make the whole world shine, but it's just an artificial neural network, a technique inspired by how our brain's nerves work.
It even entered popular culture: in the comedy series "Silicon Valley", the startup Pied Piper runs its compression service on a neural network.
Video: https://youtu.be/E2YcOV5C2x4
Here is a simple explanation: the neural network is composed of several layers of neurons. The input is passed to the first layer. A single neuron receives input, gives each input value a weight value, and produces an output based on this weight value.
The output is passed from the first layer to the second layer for processing, and so on. The final output is generated as such.
Then the miracle happened. The person running the network defines what the "correct" final output should be. Each time the data is passed over the network, the end result is compared to the "correct" value, each time it adjusts the weight value until it creates the correct final output. This network is actually self-training.
For example, this artificial brain can learn how to recognize a chair from a photo. Over time, it learns what the characteristics of the chair are to increase the probability that they will recognize the chair.
Yann LeCun, the head of Facebook's AI, uses analogy to explain neural networks:
The pattern recognition system is like a black box with a camera on one end, a green light and a red light at the top, and a large number of knobs on the front. The learning algorithm is to try to adjust the knob, when the dog is in front of the lens, the red light is on; when there is a car in front of the camera, the green light is on.
When you show a dog to the machine. Just turn the red light on, don't do anything. If it is dark, adjust the knob to make the light brighter. If the green indicator lights up, adjust the knob to make it darker. Then, when the car is displayed, adjust the knob to make the red light darker and the green light brighter.
If you show a lot of cars and dogs, and you adjust the knobs a little bit each time, the machine will get the correct answer every time.
![image description]()
Image source: a16z
Now let's talk about deep learning, which is a simple method of training multi-layer artificial neural networks. It has proven to be particularly effective in identifying patterns from data. Whenever the media talks about neural networks, it is likely to refer to deep learning.
A great explanation for machine learning and deep learning: How Does Your Phone Know This Is A Dog?
The role of deep learning in promoting AI is obvious. It is now used in many industries outside the software industry.
Facebook M, an artificial intelligence-driven virtual assistant, is using deep learning to help users with a variety of tasks – including research, booking tickets, and buying coffee.
Google is using a deep learning system called RankBrain to filter search results to compare with more traditional search results. According to Bloomberg:
The system helps Google handle 15% of queries that were not encountered by previous systems. It is good at handling ambiguous query requests, such as "What do consumers call at the top of the food chain?"
The system is now the third-largest signal of Google search results, behind anti-links and content.
Can the neural network recognize cats? This is deep learning.
From Siri to Samantha <br> Deep learning may be a key puzzle for building smarter, more human-like AI.
Google scans the cat's brain to run 16,000 computer processors. Defeat the AlphaGo program of Go World Champion Li Shishi, running on 48 processors. In the future, neural networks can run on inexpensive mobile phones.
Deep learning can improve all aspects of AI, from natural language processing to machine vision, and can be seen as a better brain that can improve computer learning.
It can improve the ability of virtual assistants such as Siri or Google to handle unfamiliar requests. It can process videos and generate short films of summary content.
Maybe one day it will win an Oscar, who knows?
Now, large and small high-tech companies are competing to make this a reality. You will hear these jargon when you look at the news: AI, machine learning, deep learning, neural networks, natural language processing.
Maybe it's all a bit confusing. Therefore, the following is the basic knowledge of these concepts and how they are related.
What is artificial intelligence (AI)?
Simply put, AI is trying to make computers smarter and even smarter than humans. This is to let the computer have human-like behavior, thinking process and reasoning ability.
There are two types of artificial intelligence:
Narrow AI (weak AI)
This AI only focuses on a narrow task. Now we are weak AI everywhere. It has beaten us humans in chess, the TV contest "The Dangerous Edge", and the recent Go game.
Digital assistants like Siri and Cortana can provide us with weather information and self-driving cars on the road. However, they have great limitations. Self-driving cars don't play chess. Siri also cannot read and delete unimportant messages. Weak AI has a narrow scope: it can't go beyond what was first set for it.
Cortana AI is a weak example <br> General AI (strong AI)
Below we enter the kingdom of science fiction. Samantha is undoubtedly defined as a strong AI. She can learn new things and modify her own code base. She can defeat humans in both chess and driving.
Analysis of AI <br> Now that we know that general AI is our ultimate goal, how can we achieve it? Here are 5 areas that need to be proficient:
Perception: Like us humans, a computer also needs five senses to interact with the world. But it is not limited to these five aspects. It can feel like someone doesn't have it. Perspective eye? Sonar detection? All possible.
Natural Language Processing (NLP): Beyond the world of perception, AI needs to understand language and writing. They need to parse sentences and understand the nuances, accents, and meanings between them. The same sentence can have different meanings depending on the context, so the difficulty of this task is well known.
Knowledge expression: Since it can perceive things - objects, people, concepts, words and mathematical symbols - it needs a way to represent the world in their brains.
Reasoning: Once it collects data through its senses and links it to the concept, it can use that data to logically solve the problem. For example, if a chess software detects that a piece of chess moves around the board, it can calculate the strategy to deal with.
Planning and navigation: To be truly human, AI must not only think like humans. You should also live among us. Therefore, one of the big problems for researchers is to help artificial intelligence move and plan the best path in the three-dimensional world. Autonomous vehicles must do this because a mistake can be fatal.
The Singapore Prime Minister took a ride in a self-driving car. Image source: Kenji Soon, MCI.
You can see how these aspects work together, such as machine vision, which uses imaging and image analysis to solve problems. For example, Facebook parses the photos you upload to social networks to suggest who you should tag, and this is fairly accurate.
Autonomous vehicles may be the most complex machine vision processing task available today. It needs to recognize road signs, obey the lanes, and pay attention to vehicles, objects and people. It works well in poor weather conditions, day or night, on dilapidated roads or on new roads.
Implementation Tools <br> These concepts are not new. They were presented at Dartmouth Conferences as early as 1956, a groundbreaking event in the field of artificial intelligence.
It will take decades for technology to keep up with our imagination. We seem to be on the cusp of the AI ​​revolution. With more venture capital investment, more large-scale technology companies are investing in AI research and development. In the coming, we are increasingly using AI in our daily lives.
Important factors that contribute to the rise of AI include Moore's Law, which allows us to inject more computing power into smaller, more efficient chips. Once computing power reaches a certain level, AI will become both practical and cost-effective.
Google made a major breakthrough in the recognition of cats. Image source: zbeads
Big data is another trend that has led to the rise of AI: After Google entered massive amounts of data for neural networks in 2012, it made breakthroughs, including 10 million YouTube video stills.
As a result, the neural network learned to recognize cats without anyone teaching it, achieving 75% accuracy. It is impossible to achieve without this 10 million video database.
When the machine learns <br> Now, let's clarify a few concepts that are often confusing. Machine learning is an AI technology that focuses on learning insights from data and using them to predict the world.
Decision tree algorithm. Image source: Wikipedia
Machine learning has been implemented using algorithms. A task is done through a series of algorithms. Examples of such algorithms include decision tree learning and association rule learning.
Machine learning algorithms make the whole world shine, but it's just an artificial neural network, a technique inspired by how our brain's nerves work.
It even entered popular culture: in the comedy series "Silicon Valley", the startup Pied Piper runs its compression service on a neural network.
Video: https://youtu.be/E2YcOV5C2x4
Here is a simple explanation: the neural network is composed of several layers of neurons. The input is passed to the first layer. A single neuron receives input, gives each input value a weight value, and produces an output based on this weight value.
The output is passed from the first layer to the second layer for processing, and so on. The final output is generated as such.
Then the miracle happened. The person running the network defines what the "correct" final output should be. Each time the data is passed over the network, the end result is compared to the "correct" value, each time it adjusts the weight value until it creates the correct final output. This network is actually self-training.
For example, this artificial brain can learn how to recognize a chair from a photo. Over time, it learns what the characteristics of the chair are to increase the probability that they will recognize the chair.
Yann LeCun, the head of Facebook's AI, uses analogy to explain neural networks:
The pattern recognition system is like a black box with a camera on one end, a green light and a red light at the top, and a large number of knobs on the front. The learning algorithm is to try to adjust the knob, when the dog is in front of the lens, the red light is on; when there is a car in front of the camera, the green light is on.
When you show a dog to the machine. Just turn the red light on, don't do anything. If it is dark, adjust the knob to make the light brighter. If the green indicator lights up, adjust the knob to make it darker. Then, when the car is displayed, adjust the knob to make the red light darker and the green light brighter.
If you show a lot of cars and dogs, and you adjust the knobs a little bit each time, the machine will get the correct answer every time.
Image source: a16z
Now let's talk about deep learning, which is a simple method of training multi-layer artificial neural networks. It has proven to be particularly effective in identifying patterns from data. Whenever the media talks about neural networks, it is likely to refer to deep learning.
A great explanation for machine learning and deep learning: How Does Your Phone Know This Is A Dog?
The role of deep learning in promoting AI is obvious. It is now used in many industries outside the software industry.
Facebook M, an artificial intelligence-driven virtual assistant, is using deep learning to help users with a variety of tasks – including research, booking tickets, and buying coffee.
Google is using a deep learning system called RankBrain to filter search results to compare with more traditional search results. According to Bloomberg:
The system helps Google handle 15% of queries that were not encountered by previous systems. It is good at handling ambiguous query requests, such as "What do consumers call at the top of the food chain?"
The system is now the third-largest signal of Google search results, behind anti-links and content.
Can the neural network recognize cats? This is deep learning.
From Siri to Samantha <br> Deep learning may be a key puzzle for building smarter, more human-like AI.
Google scans the cat's brain to run 16,000 computer processors. Defeat the AlphaGo program of Go World Champion Li Shishi, running on 48 processors. In the future, neural networks can run on inexpensive mobile phones.
Deep learning can improve all aspects of AI, from natural language processing to machine vision, and can be seen as a better brain that can improve computer learning.
It can improve the ability of virtual assistants such as Siri or Google to handle unfamiliar requests. It can process videos and generate short films of summary content.
Maybe one day it will win an Oscar, who knows?
KNBL1-32 Residual Current Circuit Breaker With Over Load Protection
KNBL1-32 TWO FUNCTION : MCB AND RCCB FUNCTIONS
leakage breaker is suitable for the leakage protection of the line of AC 50/60Hz, rated voltage single phase 240V, rated current up to 63A. When there is human electricity shock or if the leakage current of the line exceeds the prescribed value, it will automatically cut off the power within 0.1s to protect human safety and prevent the accident due to the current leakage.
leakage breaker can protect against overload and short-circuit. It can be used to protect the line from being overloaded and short-circuited as wellas infrequent changeover of the line in normal situation. It complies with standard of IEC/EN61009-1 and GB16917.1.
KNBL1-32 Residual Current Circuit Breaker,Residual Current Circuit Breaker with Over Load Protection 1p,Residual Current Circuit Breaker with Over Load Protection 2p
Wenzhou Korlen Electric Appliances Co., Ltd. , https://www.korlen-electric.com