Easyai. tech has found that getting started with AI can be difficult, especially for non-technical people.

Therefore, we integrate the excellent popular science content at home and abroad in the most accessible way, specifically for non-technical personnel, so that everyone can understand the basic concepts in the field of artificial intelligence.

Let’s start with a long illustration of the main topics covered in the PDF. Download the PDF for more details.

PDF download address

What problem does this PDF solve?

Artificial intelligence is regarded by many people as “black technology”, it can do some magical things, such as: playing go is better than human, playing games is better than human, beauty effect is good to explode…

Google, Microsoft, Facebook, Amazon, Tencent, Alibaba, Baidu, Bytedance……

Many executives also say that AI will bring about the next technological revolution, and if you think about the “Internet” revolution, you can get a sense of how big it will be.

But the big question is: how do I use AI in the AGE of AI?

The above question is too big to answer, so we need to focus on the question: When I am faced with a specific problem in my business, AI is also a way to solve the problem, so is AI suitable to solve the problem?

So, this PDF solves a problem:

Is the specific problem I’m facing suitable for AI to solve? Is there anything that needs to be evaluated?

Four dimensions of evaluation

Four evaluation dimensions are detailed in the PDF:

  1. data
  2. Characteristics of the
  3. learning
  4. Black box

data

One of the biggest differences between ARTIFICIAL intelligence and traditional computer programs is that it is based on data.

This is also the underlying logic of ARTIFICIAL intelligence, so data is the most important resource in artificial intelligence. Therefore, we need to evaluate data dimensions from three aspects:

  1. Is the data available?
  2. Is the data comprehensive?
  3. Is there a lot of data?

Download the PDF to view the full text, or follow the links below to view the full text of the data article:

Data to Evaluate before using AI

Characteristics of the

The basic principle of ARTIFICIAL intelligence is to find hidden features in large amounts of data, and then learn to use those features to complete specific tasks.

Based on this principle, AI should deal with more complex problems rather than simple ones. The complexity of a problem can be judged from the following two dimensions:

  1. Number of features
  2. Determinism of feature

Few features + weak determinism: suitable for manual solution

Fewer features + strong determinism: suitable for regular solution

Multiple features + strong determinism: suitable for regular solution

Multiple features + weak determinism: AI solution “can be considered”

To view the full text, download the PDF. You can also view the full text of the feature article by following the link below:

Evaluation before USING AI — Features

learning

As the previous two articles have explained, the boundaries of rule-based capabilities are so small that many practical problems cannot be solved by a rule-based approach. Artificial intelligence can expand the boundaries of computer capabilities.

In addition to expanding the boundaries of capabilities, ai also has a very important feature — continuous learning, constantly raising the upper limit of capabilities.

In order for machines to learn continuously, we need to achieve two things:

  1. Constantly getting feedback, letting the machine know where it’s good and where it’s bad
  2. By adding feedback data to the closed loop, the machine can continue to learn and improve its ability

Download the PDF to view the full text, or follow the links below to view the full text of the study paper:

Evaluation before USING AI — A Learning chapter

Black box

Much of our computer science used to be rule-based, much like a car, where we knew exactly how the car was put together, so we tightened it when a screw was loose and replaced it when a part was old. You can do the right thing.

Deep learning, on the other hand, when we find a problem, we can only optimize it globally (for example, feeding more data), rather than treating it right.

So, there are three principles when evaluating:

  1. The more a solution needs to explain why, the less appropriate it is to use deep learning
  2. The less tolerant you are of errors, the less appropriate it is to use deep learning
  3. The above two criteria are not absolute, but also need to look at the commercial value and cost performance. Autonomous driving and medical care are counter-examples.

Download the PDF to view the full text, or follow the links below to view the full text of the black box:

What to evaluate before using AI — Black Box

All of the above has been compiled into a 41-page PDF called “Evaluation Before AI Introduction”, which can be downloaded by clicking on the button below.

If the picture above shows a problem, please click here to download: easyAI.tech