# Historical trend of deep learning

So far, deep learning has experienced three development waves:

From 1940s to 1960s, the rudiment of deep learning appeared in cybernetics;

From 1980s to 1990s, deep learning was manifested as connectionism;

It was not until 2006 that it really revived in the name of deep learning.

# Get to know some concepts

**Automatically learn various algorithms of features and orange types from the data, and the appearance of this model is your rule base.**

# Where is deep learning in artificial intelligence

# Understanding deep learning

## 1、 The basic unit of neural network — neuron

The artificial neuron simulated by mathematical model deals with the signal sources of all dendrites and the calculation of correlation strength.

The calculation formula is as follows: S = p1w1 + p2w2 + p3w3 + B

## 2、 Structure of neural network

## 3、 Concept of deep learning

Deep neural network (deep learning) is a kind of neural network with at least one hidden layer, that is, there are many hidden layers.

# Differences between deep learning and traditional methods

# Supervised learning

Supervised learning in deep learning includes**Convolutional neural network, cyclic neural network**Wait.

# Unsupervised learning

Unsupervised learning in deep learning includes deterministic self encoder method, contrast divergence method based on probability constrained Boltzmann machine and so on.

## Common methods of deep learning

- Autoencoder
- Convolutional neural network
- Cyclic neural network

## Deep learning unsupervised self encoder

Self encoder can be used as a feature dimension reduction method.

When we use four values to represent four categories:

Four values indicate that the four categories are not compact and exist**Compressed representation**For example, two values can represent these four different numbers.

## Deep learning supervised convolutional neural network

## Deep learning supervised convolutional neural network

## Supervised method of deep learning — cyclic neural network

The source of recurrent neural network is to characterize**The relationship between the current output of a sequence and previous information**。 In terms of network structure, the recurrent neural network will remember the previous information and use the previous information to affect the output of later nodes. That is, the of recurrent neural network**Nodes between hidden layers are connected**, the input of the hidden layer includes not only the output of the input layer, but also the output of the hidden layer at the previous time.

“It’s very hard to train in Australia. I’m dying. It’s better to live than to die,” Fu Yuanhui said. It may be angry in words. “Who knows what I’ve been through, I’m too tired”. Although the words are hard, the facial expression, voice and emotion are not, so I’m still happy to sum up.

# Introduction to strong chemistry, aiphago and transfer learning

## Reinforcement learning

Don’t study, watch TV – parents scold and get beaten

Study hard – reward lollipop

## AIphaGo

## Transfer learning

# Multiple application scenarios for in-depth learning

## Security monitoring

## smart city

## Medical health

## Smart home

# Application method of deep learning in intelligent operation and maintenance

## Development process of intelligent operation and maintenance

## KPI anomaly detection algorithm

## The self encoder and clustering algorithm are used to cluster KPIs quickly

**Consistent pattern**

**Jittery mode**

**Abnormal pattern**

The common KPI data in operation and maintenance is a kind of time series data, which has the characteristics of many data instances and high dimensions. In order to reduce the cost of data analysis and improve the analysis efficiency, we hope to divide the massive time series data curves into several categories, so as to reduce the number of curves to be investigated.

Therefore, it is necessary to label large-scale auxiliary KPIs and build fault propagation chain.

## Use LSTM to forecast KPI trend

# Write at the end

In recent years, with the rapid development of aiops, the urgent needs of it tools, platform capabilities, solutions, AI scenarios and available data sets burst out in various industries.**Based on this, cloud intelligence released the aiops community in August 2021.**

**Community order** **Open Source** **The visual data arrangement platform flyfish and operation and maintenance management platform are provided** **OMP** **, cloud service management platform – Moore platform** **Hours** **Algorithms and other products.**

**Visual orchestration platform flyfish:**

Project introduction:https://www.cloudwise.ai/flyFish.html

GitHub address:https://github.com/CloudWise-OpenSource/FlyFish

Gitee address:https://gitee.com/CloudWise/fly-fish

Industry case:https://www.bilibili.com/video/BV1z44y1n77Y/

**Some large screen cases:**