GPU, ASIC (TPU, NPU), FPGA, etc.

Machine learning, deep learning, reinforcement learning, transfer learning, etc.

Structured data, unstructured data, etc.

Data collection, data cleaning, data standards, data storage, etc.

symbolism school

connectionist school

behaviorism school

Evolutionary school

Bayesianism

school of analogy

Expert system, knowledge map

Machine learning, deep learning

Behaviorism, Evolutionary Computation

Only specializes in a single task or a group of related tasks and does not have general intelligence capabilities

Have certain general intelligence capabilities and be able to understand, learn and apply them to a variety of tasks

Exceeds human intelligence in almost every aspect, including creativity, social skills, etc.

1940s-1956 Turing Test

Dartmouth Conference 1956

Symbolism 1956-1973

Symbolism (expert systems) 1980-1990

1994-present machine learning, deep learning

Algorithms learn from labeled data sets, i.e. each training sample has a known outcome

Algorithms learn from unlabeled data sets

Combines a small amount of labeled data and a large amount of unlabeled data for training

Learn through trial and error which behaviors are rewarded and which behaviors result in punishment

Deep learning, specifically, deep neural network learning

It is an important branch of machine learning

Deep learning algorithms use more "hidden layers" (hundreds), making them more powerful and allowing neural networks to do more difficult jobs

TensorFlow (Google)

Caffe(BVLC)

Keras(fchollet)

CNTK (Microsoft)

Torch7 (Meta)

PaddlePaddle (Baidu)

MindSpore (Huawei)

Machine learning models with large parameter scales and complex computing structures

The basic core structure of most large models is Transformer and its variants.

The large model that is often mentioned at present is mainly the large language model (Large Language Model).

pre-training

fine-tuning

By use

According to characteristics

by function

Press to switch source

Subscription model

API service model

Platform service model

Customized service model

Advertising and promotion models

Data authorization model

Use artificial intelligence technology to automatically create or generate content

Generated content can include text, code, images, music, videos, and more.

generate text

Vincentian picture

Vincent Audio

Vincent Video

Image recognition, visual recognition, face recognition, video recognition, text recognition, gait recognition...

Voice recognition, voiceprint recognition, speech synthesis, voice interaction...

Information understanding, text proofreading, machine translation, natural language generation...

Home service robots, medical care robots, hotel service robots, industrial robots...

Automated production, intelligent quality inspection, equipment operation and maintenance, supply chain management...

Medical image analysis, gene sequencing, disease prediction, drug research and development, personalized treatment...

Risk management, credit assessment, fraud monitoring, quantitative trading, market forecast...

Personalized learning paths, intelligent tutoring, course recommendations…

Autonomous driving, route optimization, traffic analysis, emergency plans…

Manuscript collection and writing, material creation, text polishing...

Character design, element generation, plot design, special effects production...

AI can automate repetitive and tedious tasks, improve production efficiency and quality, and reduce labor costs.

AI can not only improve governance efficiency, but also bring new business models, products and services, stimulating the economy

AI can not only improve governance efficiency, but also bring new business models, products and services, stimulating the economy

AI can help us complete some tasks and improve our quality of life

AI can also play an important role in disease treatment, disaster prediction, climate prediction, and poverty eradication.

May threaten a large number of human jobs and lead to massive unemployment

AI is used to wage war and deceive (imitate voices or change faces to commit fraud)

Infringement of citizens’ rights (excessive collection of information, invasion of privacy)

If only a few companies have advanced AI technology, it could exacerbate social inequities

AI algorithm bias may also lead to unfairness

As AI becomes more and more powerful, it will also make people dependent on AI and lose their ability to think independently and solve problems.

The powerful creativity of AI may cause humans to lose the motivation and confidence to create.

Surrounding the development of AI, there are also a series of issues such as security (data leakage, system crash), moral ethics, etc.