01 AI Overview - ALEX
01 AI Overview - ALEX
https://e.huawei.com/en/talent/#/
1 Huawei Confidential
2 Huawei Confidential
AI Overview
Foreword
⚫ Mankind is welcoming the fourth industrial revolution represented by intelligent technology. New
technologies such as AI, IoT, 5G and bioengineering are integrated into all aspects of human society;
driving changes in global macro trends, such as sustainable social development and economic growth.
New kinetic energy, smart city upgrading, industrial digital transformation, consumer experience, etc.
⚫ As the world‘s leading provider of ICT (information and communications) infrastructure and smart
terminals, Huawei actively participates in the transformation of artificial intelligence and proposes
Huawei’s full-stack full-scenario AI strategy. This chapter will mainly introduce AI Overview, Technical
Fields and Application Fields of AI, Huawei's AI Development Strategy, AI Disputes, Future Prospects of
AI.
4 Huawei Confidential
Objectives
5 Huawei Confidential
Contents
1. AI Overview
4. AI Disputes
5. Future Prospects of AI
6 Huawei Confidential
AI in the Eyes of the Society
⚫ People get to know AI through news, movies, and actual applications in daily life. What is
AI in the eyes of the public?
Self-service security check
Spoken language evaluation
Haidian Park: First AI-themed Park in the World StarCraft The Terminator Music/Movie recommendation
II: AlphaStar Beat Professional Players 2001: A Space Odyssey Smart speaker
AI-created Edmond de Belamy Sold at US$430,000 The Matrix Ai facial fortune-telling
Demand for AI Programmers:↑ 35 Times! Salary: Top 1! I, Robot Vacuum cleaning robot
50% Jobs Will be Replaced by AI in the future Blade Runner Self-service bank terminal
Winter is Coming? AI Faces Challenges Elle Intelligent customer service
… Bicentennial Man Siri
… …
News Movies Applications in daily life
AI Applications AI Control over human beings Security protection
AI industry outlook Fall in love with AI Entertainment
Challenges faced by AI Self-awareness of AI Smart Home
… … Finance
…
7 Huawei Confidential
AI in the Eyes of Researchers
"I propose to consider the question, 'Can machines think?'"
— Marvin Minsky
8 Huawei Confidential
What Are Intelligences?
⚫ Howard Gardner's Multiple Intelligences
⚫ Human intelligences can be divided into seven categories:
Verbal/Linguistic
Logical/Mathematical
Visual/Spatial
Bodily/Kinesthetic
Musical/Rhythmic
Inter-personal/Social
Intra-personal/Introspective
9 Huawei Confidential
What Is AI?
⚫ Artificial Intelligence (AI) is a new technical science that studies and develops theories, methods, techniques,
and application systems for simulating and extending human intelligence. In 1956, the concept of AI was first
proposed by John McCarthy, who defined the subject as "science and engineering of making intelligent
machines, especially intelligent computer program". AI is concerned with making machines work in an
intelligent way, similar to the way that the human mind works. At present, AI has become an interdisciplinary
course that involves various fields.
Brain
science Cognitive
science
Computer
science
AI Psychology
Philosophy
Linguistics
Logic
10 Huawei Confidential
Relationship of AI, Machine Learning, and Deep Learning
11 Huawei Confidential
Relationship of AI, Machine Learning and Deep Learning
⚫ AI: A new technical science that focuses on the research and development of theories, methods,
techniques, and application systems for simulating and extending human intelligence.
⚫ Machine learning: A core research field of AI. It focuses on the study of how computers can obtain
new knowledge or skills by simulating or performing learning behavior of human beings, and
reorganize existing knowledge architecture to improve its performance. It is one of the core
research fields of AI.
⚫ Deep learning: A new field of machine learning. The concept of deep learning originates from the
research on artificial neural networks. The multi-layer perceptron (MLP) is a type a deep learning
architecture. Deep learning aims to simulate the human brain to interpret data such as images,
sounds, and texts.
12 Huawei Confidential
Three Major Schools of Thought: Symbolism
⚫ Basic thoughts
The cognitive process of human beings is the process of inference and operation of various
symbols.
A human being is a physical symbol system, and so is a computer. Computers, therefore, can be
used to simulate intelligent behavior of human beings.
The core of AI lies in knowledge representation, knowledge inference, and knowledge
application. Knowledge and concepts can be represented with symbols. Cognition is the process
of symbol processing while inference refers to the process of solving problems by using heuristic
knowledge and search.
14 Huawei Confidential
Three Major Schools of Thought: Behaviorism
⚫ Basic thoughts:
Intelligence depends on perception and action. The perception-action mode of intelligent
behavior is proposed.
Intelligence requires no knowledge, representation, or inference. AI can evolve like human
intelligence. Intelligent behavior can only be demonstrated in the real world through the
constant interaction with the surrounding environment.
15 Huawei Confidential
Brief Development History of AI
2016 March: AlphaG o
defeated t he world
2014: Microsoft champion Go player Lee
released the first Sedol by 4-1.
1997: Deep Blue individual intelligent
1985: Decision-
defeated t he world assistant Microsf t
1956: AI was proposed at making tree
1976: Due to failure chess champion Cortana in the world.
the Dartmouth Conference. models with
of projects such as Garry Kasparov.
better
machine t ranslat ion
visualization 2006: Hinton and his
and negat ive impact
ef fect and multi- students st arted deep
of some academic
layer ANNs learning.
reports, the f und for
which broke
AI was decreased 2017 October: The Deep
through t he limit
in general. Mind team released
of early
percept ron. AlphaG o Zero, the
1987: The 2010: The
1959: Arthur Samuel strongest version of
market of LISP era of big
proposed machine AlphaG o.
machines data
learning.
collapsed. came.
1956-1976 1997-2010
First period of boom Period of recovery
The concept and development target 1976-1982 1982-1987 Computing perf ormance
2010-
of AI were determined at t he First period of Second period was improved and Internet
1987-1997 Period of rapid growth
Dartmouth conf erence. low ebb of boom technologies got
Second period of low New-generation
AI suff ered from Expert syst em popularized quickly.
ebb information technologies
quest ioning and capable of logic Technical f ields f aced triggered transformation of
crit icism due to rule inference bottlenecks, people information environment
insufficient and answering on longer focused on and dat a basis. Multi-
computing quest ions of abstract inference, model data such as
capabilities, high specific fields and models based on massive images, voices,
computing went popular symbol processing and texts emerged
complexity, and and fifth- were rejected. continuously. Computing
great difficulty of generation capabilities were improved.
inference computers
realization. developed.
16 Huawei Confidential
Overview of AI Technologies
⚫ AI technologies are multi-layered, covering the application, algorithm mechanism,
toolchain, device, chip, process, and material layers.
Application
Algorithm
Device
Chip
Process
17 Huawei Confidential
Types of AI
⚫ Strong AI
The strong AI view holds that it is possible to create intelligent machines that can really reason
and solve problems. Such machines are considered to be conscious and self-aware, can
independently think about problems and work out optimal solutions to problems, have their
own system of values and world views, and have all the same instincts as living things, such as
survival and security needs. It can be regarded as a new civilization in a certain sense.
⚫ Weak AI
The weak AI view holds that intelligent machines cannot really reason and solve problems.
These machines only look intelligent, but do not have real intelligence or self-awareness.
19 Huawei Confidential
Classification of Intelligent Robots
⚫ Currently, there is no unified definition of AI research. Intelligent robots are generally
classified into the following four types:
"Thinking like human beings": weak AI, such as Watson and AlphaGo
"Acting like human beings": weak AI, such as humanoid robot, iRobot, and Atlas of Boston
Dynamics
"Thinking rationally": strong AI (Currently, no intelligent robots of this type have been created
due to the bottleneck in brain science.)
"Acting rationally": strong AI
20 Huawei Confidential
AI Industry Ecosystem
⚫ The four elements of AI are data, algorithm, computing power, and scenario. To meet requirements of these
four elements, we need to combine AI with cloud computing, big data, and IoT to build an intelligent society.
21 Huawei Confidential
Sub-fields of AI
22 Huawei Confidential
Contents
1. AI Overview
4. AI Disputes
5. Future Prospects of AI
23 Huawei Confidential
Technical Fields and Application Fields of AI
25 Huawei Confidential
Computer Vision Application Scenario (1)
⚫ Computer vision is the most mature technology among the three AI technologies. The main topics of computer vision
research include image classification, target detection, image segmentation, target tracking,
optical character recognition (OCR), and facial recognition.
⚫ In the future, computer vision is expected to enter the advanced stage of autonomous understanding, analysis, and decision-
making, enabling machines to "see" and bringing greater value to scenarios such as unmanned vehicles and smart homes.
⚫ Application scenarios:
26 Huawei Confidential
Computer Vision Application Scenario (2)
Facial verification passed Facial verification failed
Infringement Infringement
Plant Food
Smart album
Image search Infringement Infringement
People Building
27 Huawei Confidential
Voice Processing Application Scenario (1)
⚫ The main topics of voice processing research include voice recognition, voice synthesis, voice wakeup,
voiceprint recognition, and audio-based incident detection. Among them, the most mature technology is voice
recognition. As for near field recognition in a quite indoor environment, the recognition accuracy can reach
96%.
⚫ Application scenarios:
Question Answering Bot (QABot) Voice navigation
28 Huawei Confidential
Voice Processing Application Scenario (2)
Intelligent
education Real-time conference
records
⚫ Other applications:
Spoken language evaluation
Diagnostic robot
Voiceprint recognition
Smart sound box
...
29 Huawei Confidential
NLP Application Scenario (1)
⚫ The main topics of NLP research include machine translation, text mining, and sentiment analysis. NLP imposes high requirements
on technologies but confronts low technology maturity. Due to high complexity of semantics, it is hard to reach the human
understanding level using parallel computing based on big data and parallel computing only.
⚫ In future, NLP will achieve more growth: understanding of shallow semantics → automatic extraction of features and understanding
of deep semantics; single-purpose intelligence (ML) → hybrid intelligence (ML, DL, and RL)
⚫ Application scenarios:
Theme Trend
Public opinion mining analysis Evaluation
analysis analysis
Emotional
Public opinion
analysis
analysis
Hotspot
event Information
distribution
30 Huawei Confidential
NLP Application Scenario (2)
Text
Machine
classification
translation
⚫ Other applications:
Knowledge graph
Intelligent copywriting
Video subtitle
...
31 Huawei Confidential
AI Application Field - Intelligent Healthcare
Virtual assistant: electronic voice medical records, intelligent guidance, intelligent diagnosis, and
medicine recommendation
Medical image: medical image recognition, image marking, and 3D image reconstruction
32 Huawei Confidential
AI Application Field - Intelligent Security
⚫ Security protection is considered the easiest field for AI implementation. AI technologies applied in this field are relatively
mature. The field involves massive data of images and videos, laying a sound foundation for training of AI algorithms and
models. Currently, AI technologies are applied to two directions in the security protection field, namely, civil use and police
use.
⚫ Application scenarios:
Police use: suspect identification, vehicle analysis, suspect tracking, suspect search and comparison, and access control at key places
Civil use: facial recognition, warning against potential danger, and home protective measure deployment
33 Huawei Confidential
AI Application Field - Smart Home
⚫ Based on IoT technologies, a smart home ecosystem is formed with hardware, software, and cloud
platforms, providing users personalized life services and making home life more convenient,
comfortable, and safe.
Control smart home products with Implement home security protection Okay, the
voice processing such as air with computer vision technologies, temperature's set.
conditioning temperature for example, facial or fingerprint Set the temperature to
adjustment, curtain switch control, recognition for unlocking, real-time 26 degrees.
and voice control on the lighting intelligent camera monitoring, and
system. illegal intrusion detection.
34 Huawei Confidential
AI Application Field - Smart City
35 Huawei Confidential
AI Application Field - Retail
⚫ AI will bring revolutionary changes to the retail industry. A typical symptom is unmanned supermarkets. For example, Amazon Go, unmanned
supermarket of Amazon, uses sensors, cameras, computer vision, and deep learning algorithms to completely cancel the checkout process,
allowing customers to pick up goods and "just walk out".
⚫ One of the biggest challenges for unmanned supermarket is how to charge the right fees to the right customers. So far, Amazon Go is the only
successful business case and even this case involves many controlled factors. For example, only Prime members can enter Amazon Go. Other
enterprises, to follow the example of Amazon, have to build their membership system first.
36 Huawei Confidential
AI Application Field - Autonomous Driving
⚫ The Society of Automotive Engineers (SAE) in the U.S. defines 6 levels of driving automation ranging from 0
(fully manual) to 5 (fully autonomous). L0 indicates that the driving of a vehicle completely depends on the
driver's operation. The system above L3 can implement the driver's hand-off operation in specific cases, L5
depends on the system when vehicles are driving in all scenarios.
⚫ Currently, only some commercial passenger vehicle models, such as Audi A8, Tesla, and Cadillac, support L2
and L3 Advanced driver-assistance systems (ADAS). It is estimated that by 2020, more L3 vehicle models will
emerge with the further improvement of sensors and vehicle-mounted processors. L4 and L5 autonomous
driving is expected to be first implemented on commercial vehicles in closed campuses. A wider range of
passenger vehicles require advanced autonomous driving, which requires further improvement of
technologies, policies, and infrastructure. It is estimated that L4 and L5 autonomous driving will be supported
by common roads in 2025–2030.
37 Huawei Confidential
AI Will Change All Industries
38 Huawei Confidential
AI: Still in Its Infancy
Ability Example Value Benefits
Capable of storage and Distributed computing and Help human beings store
computing: Machines can neural network and quickly process massive
Computing compute and transfer data, laying a foundation for
information as human beings
Three Phases of AI
40 Huawei Confidential
Contents
1. AI Overview
4. AI Disputes
5. Future Prospects of AI
41 Huawei Confidential
Huawei's Full-Stack, All-Scenario AI Portfolio
AI Applications Application enablement: provides end-to-end
Application
services (ModelArts), layered APIs, and pre-
HiAI Engine ModelArts Enablement integrated solutions.
MindSpore: supports the unified training and
TensorFlow PyTorch PaddlePaddle MindSpore Framework
inference framework that is independent of the
device, edge, and cloud.
Chip
Full Stack CANN Enablement CANN: a chip operator library and highly automated
operator development tool.
IP & Chip
Ascend-Nano Ascend-Tiny Ascend-Lite Ascend Ascend-Mini Ascend-Max IP and Chip
Ascend: provides a series of NPU IPs and chips based
on a unified, scalable architecture.
Consumer Device Public Cloud Private Cloud Edge Computing Industrial IoT Device
Huawei's "all AI scenarios" indicate different deployment scenarios for AI, including public clouds, private
clouds, edge computing in all forms, industrial IoT devices, and consumer devices.
42 Huawei Confidential
Full Stack - ModelArts Full-Cycle AI Workflow
EI Intelligent Twins
ModelArts
AI data framework Visualized workflow Distributed training One-click deployment on Automatic learning AI sharing platform
accelerates data processing management shortens training from device, edge, and cloud enables you to start builds internal and external
by 100 folds. makes development weeks to minutes. supports various deployment from scratch. AI ecosystems for
worry-free. scenarios. enterprises.
43 Huawei Confidential
Full Stack — MindSpore (Huawei AI Computing Framework)
⚫ MindSpore provides automatic parallel capabilities. With MindSpore, senior algorithm engineers and data scientists who
focus on data modeling and problem solving can run algorithms on dozens or even thousands of AI computing nodes with
only a few lines of description.
⚫ The MindSpore framework supports both large-scale and small-scale deployment, adapting to independent deployment in
all scenarios. In addition to the Ascend AI processors, MindSpore also supports other processors such as GPUs and CPUs.
MindSpore
Unified APIs for all scenarios
45 Huawei Confidential
Full Stack — CANN
CANN:
A chip operators library and highly automated operator
development toolkit
Optimal development efficiency, in-depth optimization of the
AI applications common operator library, and abundant APIs
Operator convergence, best matching the performance of the
Ascend chip
HiAI Service General APIs Advanced APIs Pre-integrated Solutions
Application
HiAI Engine ModelArts enablement
CANN
Compute Architecture for Neural Networks
Full MindSpore TensorFlow PyTorch PaddlePaddle … Framework
stack
FusionEngine
Processor
CANN enablement
TBE operator CCE Operator
development tool Library
Ascend- Ascend- Ascend- Ascend- Ascend-
Nano Tiny Lite Ascend Mini Max
IP and Chip
CCE Compiler
Public Private Edge Industrial
Consumer device
cloud cloud computing devices
All scenarios
46 Huawei Confidential
Full Stack — Ascend 310 AI Processor and Da Vinci Core
47 Huawei Confidential
Ascend AI Processors: Infusing Superior Intelligence for
Computing
FLOPS
256T
4
3
125T
Ascend 310 Ascend 910 2 90T
45T
AI SoC with ultimate Most powerful AI 1
energy efficiency processor
Ascend-Mini Ascend 910
Architecture: Da Vinci Ascend-Max
Architecture: Da Vinci
Half-precision (FP16): 8 TFLOPS
Integer precision (INT8): 16 TOPS Half-precision (FP16): 256 TFLOPS
16-channel full-HD video decoder: H.264/265 Integer precision (INT8): 512 TOPS
1-channel full-HD video encoder: H.264/265 128-channel full HD video decoder: H.264/265
Max. power: 8 W Max. power: 310 W
48 Huawei Confidential
Atlas AI Computing Platform Portfolio
Internet, security, finance, transportation, power, etc.
Atlas intelligent edge platform Atlas deep learning platform
Application
Enablement Industry SDK/Container Cluster management/Model
engine/Basic service repository management/Data pre-processing
TensorFlow/PyTorch/Caffe/MxNet Common
Framework MindSpore components
Framework
Framework Adapter
Adapret
AscendCL
Safety subsystem
Driver
subsystem
Atlas 800 AI inference server
Atlas 300 inference
Atlas 200 512 TOPS INT8
accelerator card
16 TOPS INT8 64 TOPS INT8
Atlas 500 Atlas 900
Chips & 16 TOPS INT8 256–1024 PFLOPS FP16
Hardware Atlas 800 AI training server
Atlas 300 training card
Atlas 200 developer kit 256 TFLOPS FP16 2 PFLOPS FP16
Da Vinci
Architecture
50 Huawei Confidential
HUAWEI CLOUD AI and HUAWEI Mobile Phones Help RFCx
Protect the Rainforest
51 Huawei Confidential
Contents
1. AI Overview
4. AI Disputes
5. Future Prospects of AI
52 Huawei Confidential
Algorithmic Bias
⚫ Algorithmic biases are mainly caused by data biases.
⚫ When we use AI algorithms for decision-making, the algorithms may learn to discriminate an individual based on existing
data including race and gender, and therefore create unfair outcomes, such as decisions that are discriminatory based on
race, sex or other factors. Even if factors such as race or gender are excluded from the data, the algorithms can make
discriminatory decisions based on information of names and addresses.
53 Huawei Confidential
54 Huawei Confidential
Privacy Issues
⚫ The existing AI algorithms are all data-driven. In this case, we need a large amount of data to train
models. We enjoy the convenience brought by AI every day while technology companies like
Facebook, Google, Amazon, and Alibaba are obtaining an enormous amount of user data, which
will reveal various aspects of our lives including politics, religions, and gender.
55 Huawei Confidential
Seeing = Believing?
⚫ With the development of computer vision technologies, reliability of images and videos is decreasing. Fake
images can be produced with technologies such as PS and generative adversarial networks (GAN), making it
hard to identify whether images are true or not.
⚫ Example:
A suspect provided fake evidence by forging an image in which the suspect is in a place where he has never been to or
with someone he has never seen using PS technologies.
In advertisements for diet pills, people's appearances before and after weight loss can be changed with PS technologies
to exaggerate the effect of the pills.
Lyrebird, a tool for simulating voice of human beings based on recording samples of minutes, may be used by criminals.
Household images released on rent and hotel booking platforms may be generated through GAN.
Example: https://www.whichfaceisreal.com/
56 Huawei Confidential
AI Development = Rising Unemployment?
⚫ Looking back, human beings have always been seeking ways to improve efficiency, that is, obtain more with
less resources. We used sharp stones to hunt and collect food more efficiently. We used steam engines to
reduce the need for horses. Every step in achieving automation will change our life and work. In the era of AI,
what jobs will be replaced by AI?
⚫ The answer is repetitive jobs that involve little creativity and social interaction.
58 Huawei Confidential
Contents
1. AI Overview
4. AI Disputes
5. Future Prospects of AI
59 Huawei Confidential
Development Trends of AI Technologies
⚫ Framework: easier-to-use development framework
⚫ Algorithm: algorithm models with better performance and smaller size
⚫ Computing power: comprehensive development of device-edge-cloud computing
⚫ Data: more comprehensive basic data service industry and more secure data sharing
⚫ Scenario: continuous breakthroughs in industry applications
60 Huawei Confidential
Easier-to-Use Development Framework
⚫ Various AI development frameworks are evolving towards ease-of-use and omnipotent,
continuously lowering the threshold for AI development.
61 Huawei Confidential
Tensorflow 2.0
⚫ TensorFlow 2.0 has been officially released. It integrates Keras as its high-level API, greatly
improving usability.
62 Huawei Confidential
Pytorch vs Tensorflow
⚫ PyTorch is widely recognized by academia for its ease of use.
63 Huawei Confidential
Algorithms Model with Better Performance
⚫ In the computer vision field, GAN has been able to generate high-quality images that cannot be
identified by human eyes. GAN-related algorithms have been applied to other vision-related tasks,
such as semantic segmentation, facial recognition, video synthesis, and unsupervised clustering.
⚫ In the NLP field, the pre-training model based on the Transformer architecture has made a
significant breakthrough. Related models such as BERT, GPT, and XLNet are widely used in industrial
scenarios.
⚫ In the reinforcement learning field, AlphaStar of the DeepMind team defeated the top human
player in StarCraft II.
⚫ ...
64 Huawei Confidential
Smaller Deep Learning Models
⚫ A model with better performance usually has a larger quantity of parameters, and a large model
has lower running efficiency in industrial applications. More and more model compression
technologies are proposed to further compress the model size while ensuring the model
performance, meeting the requirements of industrial applications.
Low rank approximation
Network
architecture
Network pruning design
65 Huawei Confidential
Computing Power with Comprehensive Device-Edge-Cloud
Development
⚫ The scale of AI chips applied to the cloud, edge devices, and mobile devices keeps increasing,
further meeting the computing power demand of AI.
66 Huawei Confidential
More Secure Data Sharing
⚫ Federated learning uses different data sources to train models, further breaking data bottlenecks
while ensuring data privacy and security.
67 Huawei Confidential
Continuous Breakthroughs in Application Scenarios
⚫ With the continuous exploration of AI in various verticals, the application scenarios of AI
will be continuously broken through.
Mitigating psychological problems
Automatic vehicle insurance and loss assessment
Office automation
...
68 Huawei Confidential
Mitigating Psychological Problems
⚫ AI chat robots help alleviate mental health problems such as autism by combining psychological
knowledge.
69 Huawei Confidential
Automatic Vehicle Insurance and Loss Assessment
⚫ AI technologies help insurance companies optimize vehicle insurance claims and complete vehicle
insurance loss assessment using deep learning algorithms such as image recognition.
70 Huawei Confidential
Office Automation (Robotics Process Automation)
⚫ AI is automating management, but the different nature and format of data makes it a challenging
task. While each industry and application has its own unique challenges, different industries are
gradually adopting machine learning-based workflow solutions.
72 Huawei Confidential
Summary
⚫ This chapter introduces the definition and development history of AI, describes the
technical fields and application fields of AI, briefly introduces Huawei's AI
development strategy, and finally discusses the disputes and the development trends
of AI.
73 Huawei Confidential
Quiz
B. Smart healthcare
C. Smart city
D. Smart education
2. (True or False) By "all AI scenarios", Huawei means different deployment scenarios for AI, including
public clouds, private clouds, edge computing in all forms, industrial IoT devices, and consumer
devices.
A. True
B. False
74 Huawei Confidential
More Information
75 Huawei Confidential
Thank you. 把数字世界带入每个人、每个家庭、
每个组织,构建万物互联的智能世界。
Bring digital to every person, home, and
organization for a fully connected,
intelligent world.