Name: 生成式人工智能簡介 (Introduction to Generative AI)
Uploaded: 2025-07-09T05:01:05.000Z
Duration: 22 min 8 s

Hello, and welcome to Introduction to Generative AI.

大家好，歡迎來到《生成式人工智能導論》。

My name is Dr. Gwendolyn Stripling, and I am the Artificial Intelligence Technical Curriculum Developer here at Google Cloud.

我叫 Gwendolyn Stripling 博士，是谷歌雲的人工智能技術課程開發人員。

In this course, you learn to define generative AI, explain how generative AI works, describe generative AI model types, and describe generative AI applications.

在本課程中，您將學習如何定義生成式人工智能、如何解釋生成式人工智能的工作原理、如何描述生成式人工智能模型類型以及如何描述生成式人工智能應用。

Generative AI is a type of artificial intelligence technology that can produce various types of content, including text, imagery, audio, and synthetic data.

生成式人工智能是一種人工智能技術，可以生成各種類型的內容，包括文本、影像、音頻和合成數據。

Well, since we are going to explore generative artificial intelligence, let's provide a bit of context.

好吧，既然我們要探討生成式人工智能，那就讓我們先了解一下背景情況。

So two very common questions asked are, what is artificial intelligence?

是以，有兩個非常常見的問題是：什麼是人工智能？

And what is the difference between AI and machine learning?

人工智能和機器學習有什麼區別？

One way to think about it is that AI is a discipline, like physics, for example.

一種思考方式是，人工智能是一門學科，就像物理學一樣。

AI is a branch of computer science that deals with the creation of intelligent agents, which are systems that can reason and learn and act autonomously.

人工智能是計算機科學的一個分支，主要研究創建智能代理，即能夠推理、學習和自主行動的系統。

Essentially, AI has to do with the theory and methods to build machines that think and act like humans.

從根本上說，人工智能與製造像人類一樣思考和行動的機器的理論和方法有關。

In this discipline, we have machine learning, which is a subfield of AI.

在這門學科中，我們有機器學習，它是人工智能的一個子領域。

It is a program or system that trains a model from input data.

它是一個根據輸入數據訓練模型的程序或系統。

The trained model can make useful predictions from new or never-before-seen data drawn from the same one used to train the model.

經過訓練的模型可以從用於訓練模型的同一數據中提取新的或從未見過的數據，從而做出有用的預測。

Machine learning gives the computer the ability to learn without explicit programming.

機器學習為計算機提供了無需明確編程的學習能力。

Two of the most common classes of machine learning models are unsupervised and supervised ML models.

最常見的兩類機器學習模型是無監督和有監督 ML 模型。

The key difference between the two is that with supervised models, we have labels.

兩者之間的主要區別在於，有監督模型有標籤。

Labeled data is data that comes with a tag, like a name, a type, or a number.

標籤數據是帶有標籤的數據，如名稱、類型或數字。

Unlabeled data is data that comes with no tag.

無標籤數據是指沒有標籤的數據。

This graph is an example of the sort of problem that a supervised model might try to solve.

這幅圖就是有監督模型試圖解決的問題的一個例子。

For example, let's say you are the owner of a restaurant.

例如，假設您是一家餐館的老闆。

You have historical data of the bill amount and how much different people tipped based on order type, whether it was picked up or delivered.

您有賬單金額的歷史數據，以及不同的人根據訂單類型、取貨或送貨給了多少小費。

In supervised learning, the model learns from past examples to predict future values, and in this case, tips.

在監督學習中，模型從過去的例子中學習，以預測未來的值，在這種情況下，預測的是提示。

So here, the model uses the total bill amount to predict the future tip amount based on whether an order was picked up or delivered.

是以，在這裡，模型使用賬單總金額來預測未來的小費金額，預測的依據是訂單是取貨還是送貨。

This is an example of the sort of problem that an unsupervised model might try to solve.

這就是一個無監督模型可能試圖解決的問題的例子。

So here, you want to look at tenure and income, and then group or cluster employees to see whether someone is on the fast track.

是以，在這裡，你要看的是任期和收入，然後對員工進行分組或聚類，看是否有人在快速發展的軌道上。

Unsupervised problems are all about discovery, about looking at the raw data and seeing if it naturally falls into groups.

無監督問題的關鍵在於發現，在於觀察原始數據，看它是否自然地歸類。

Let's get a little deeper and show this graphically, as understanding these concepts are the foundation for your understanding of generative AI.

讓我們再深入一點，用圖表來說明這一點，因為理解這些概念是理解生成式人工智能的基礎。

In supervised learning, testing data values, or x, are input into the model.

在監督學習中，測試數據值或 x 被輸入到模型中。

The model outputs a prediction and compares that prediction to the training data used to train the model.

模型輸出預測結果，並將預測結果與用於訓練模型的訓練數據進行比較。

If the predicted test data values and actual training data values are far apart, that's called error.

如果預測的測試數據值與實際的訓練數據值相差甚遠，這就是誤差。

And the model tries to reduce this error until the predicted and actual values are closer together.

模型會努力減小誤差，直到預測值和實際值更加接近。

Now that we've explored the difference between artificial intelligence and machine learning and supervised and unsupervised learning, let's briefly explore where deep learning fits as a subset of machine learning methods.

既然我們已經探討了人工智能與機器學習、有監督學習與無監督學習之間的區別，那麼讓我們簡要探討一下深度學習作為機器學習方法的一個子集的定位。

While machine learning is a broad field that encompasses many different techniques, deep learning is a type of machine learning that uses artificial neural networks, allowing them to process more complex patterns than machine learning.

機器學習是一個廣泛的領域，包含許多不同的技術，而深度學習是一種使用人工神經網絡的機器學習，使其能夠處理比機器學習更復雜的模式。

Artificial neural networks are inspired by the human brain.

人工神經網絡的靈感來自人腦。

They are made up of many interconnected nodes, or neurons, that can learn to perform tasks by processing data and making predictions.

它們由許多相互連接的節點或神經元組成，可以通過處理數據和進行預測來學習執行任務。

Deep learning models typically have many layers of neurons, which allows them to learn more complex patterns than traditional machine learning models.

深度學習模型通常有多層神經元，這使它們能夠學習比傳統機器學習模型更復雜的模式。

And neural networks can use both labeled and unlabeled data.

而神經網絡既可以使用有標籤的數據，也可以使用無標籤的數據。

In semi-supervised learning, a neural network is trained on a small amount of labeled data and a large amount of unlabeled data.

在半監督學習中，神經網絡是在少量有標籤數據和大量無標籤數據的基礎上進行訓練的。

The labeled data helps the neural network to learn the basic concepts of the task, while the unlabeled data helps the neural network to generalize to new examples.

有標籤的數據有助於神經網絡學習任務的基本概念，而無標籤的數據則有助於神經網絡泛化到新的示例。

Now, we finally get to where generative AI fits into this AI discipline.

現在，我們終於要討論生成式人工智能在這門人工智能學科中的定位了。

Gen AI is a subset of deep learning, which means it uses artificial neural networks, can process both labeled and unlabeled data using supervised, unsupervised, and semi-supervised methods.

Gen AI 是深度學習的一個子集，這意味著它使用人工神經網絡，可以使用監督、無監督和半監督方法處理有標籤和無標籤數據。

Large language models are also a subset of deep learning.

大型語言模型也是深度學習的一個子集。

Deep learning models, or machine learning models in general, can be divided into two types, generative and discriminative.

深度學習模型或一般的機器學習模型可分為兩種類型：生成型和判別型。

A discriminative model is a type of model that is used to classify or predict labels for data points.

判別模型是一種用於對數據點進行分類或預測標籤的模型。

Discriminative models are typically trained on a data set of labeled data points, and they learn the relationship between the features of the data points and the labels.

判別模型通常是在標有數據點的數據集上進行訓練，並學習數據點特徵與標籤之間的關係。

Once a discriminative model is trained, it can be used to predict the label for new data points.

判別模型訓練完成後，就可以用來預測新數據點的標籤。

A generative model generates new data instances based on a learned probability distribution of existing data.

生成模型根據現有數據的概率分佈生成新的數據實例。

Thus, generative models generate new content.

是以，生成模型可以生成新的內容。

Here, the discriminative model learns the conditional probability distribution, or the probability of Y, our output, given X, our input, that this is a dog, and classifies it as a dog and not a cat.

在這裡，判別模型學習條件概率分佈，或者說在輸入 X 的情況下，我們的輸出 Y 的概率，即這是一隻狗，並將其分類為狗而不是貓。

The generative model learns the joint probability distribution, or the probability of X and Y, and predicts the conditional probability that this is a dog, and can then generate a picture of a dog.

生成模型可以學習聯合概率分佈，即 X 和 Y 的概率，並預測出這是一隻狗的條件概率，然後生成一幅狗的圖片。

So to summarize, generative models can generate new data instances, while discriminative models discriminate between different kinds of data instances.

是以，概括地說，生成模型可以生成新的數據實例，而判別模型則可以判別不同類型的數據實例。

The top image shows a traditional machine learning model, which attempts to learn the relationship between the data and the label, or what you want to predict.

上圖顯示的是一個傳統的機器學習模型，它試圖學習數據和標籤之間的關係，或者說你想預測什麼。

The bottom image shows a generative AI model, which attempts to learn patterns on content so that it can generate new content.

下圖顯示的是一個生成式人工智能模型，它試圖學習內容的模式，從而生成新的內容。

A good way to distinguish what is gen AI and what is not is shown in this illustration.

本圖是區分哪些是基因人工智能，哪些不是的好方法。

It is not gen AI when the output, or Y, or label, is a number, or a class, for example, spam or not spam, or a probability.

當輸出或 Y 或標籤是一個數字或一個類別（例如垃圾郵件或非垃圾郵件）或一種概率時，它就不是基因人工智能。

It is gen AI when the output is natural language, like speech or text, an image, or audio, for example.

當輸出是自然語言（如語音或文本）、影像或音頻時，它就是基因人工智能。

Visualizing this mathematically would look like this.

如果用數學方法來直觀地理解，可以這樣理解

If you haven't seen this for a while, the Y is equal to f of X equation calculates the dependent output of a process given different inputs.

如果你很久沒有見過這個公式了，Y 等於 X 的 f 這個公式計算的是一個過程在不同輸入條件下的依存輸出。

The Y stands for the model output, the F embodies the function used in the calculation, and the X represents the input, or inputs used for the formula.

Y 代表模型輸出，F 代表計算中使用的函數，X 代表輸入或公式中使用的輸入。

So the model output is a function of all the inputs.

是以，模型的輸出是所有輸入的函數。

If the Y is a number, like predicted sales, it is not gen AI.

如果 "Y "是一個數字，比如預測銷售額，那麼它就不是基因人工智能。

If Y is a sentence, like defined sales, it is generative, as the question would elicit a text response.

如果 Y 是一個句子，比如定義的銷售額，那麼它就是生成性的，因為問題會引出一個文本回答。

The response would be based on all the massive large data the model was already trained on.

響應將基於模型已經訓練過的所有海量大數據。

To summarize, at a high level, the traditional classical supervised and unsupervised learning process takes training code and label data to build a model.

總而言之，從高層次來看，傳統的經典監督和非監督學習過程需要利用訓練代碼和標籤數據來建立模型。

Depending on the use case or problem, the model can give you a prediction, it can classify something, or cluster something.

根據不同的用例或問題，模型可以給出預測、分類或聚類。

We use this example to show you how much more robust the gen AI process is.

我們通過這個例子向大家展示基因人工智能流程的強大之處。

The gen AI process can take training code, label data, and unlabeled data of all data types and build a foundation model.

Gen AI 流程可以獲取各種數據類型的訓練代碼、標籤數據和未標籤數據，並建立基礎模型。

The foundation model can then generate new content.

然後，基礎模型可以生成新的內容。

For example, text, code, images, audio, video, et cetera.

例如，文本、代碼、影像、音頻、視頻等。

We've come a long way from traditional programming to neural networks, to generative models.

從傳統編程到神經網絡，再到生成模型，我們已經走過了漫長的道路。

In traditional programming, we used to have to hard code the rules for distinguishing a cat.

在傳統編程中，我們過去不得不硬編碼區分貓的規則。

In the wave of neural networks, we could give the network pictures of cats and dogs and ask, is this a cat?

在神經網絡浪潮中，我們可以給網絡提供貓和狗的圖片，然後問：這是貓嗎？

In the generative wave, we, as users, can generate our own content.

在生成浪潮中，作為用戶，我們可以生成自己的內容。

Whether it be text, images, audio, video, et cetera.

無論是文本、影像、音頻還是視頻，等等。

For example, models like POM, or Pathways Language Model, or Lambda, Language Model for Dialogue Applications, ingest very, very large data from the multiple sources across the internet and build foundation language models we can use simply by asking a question.

例如，像 POM（Pathways Language Model）或 Lambda（Language Model for Dialogue Applications）這樣的模型，可以從互聯網上的多種來源獲取非常非常龐大的數據，並建立基礎語言模型，我們只需提出一個問題就能使用這些模型。

Whether typing it into a prompt or verbally talking into the prompt itself.

無論是在提示符中鍵入，還是在提示符中口述。

It can give you everything it has learned about a cat.

它可以向你提供關於貓的一切資訊。

現在我們來看看我們的正式定義。

Generative AI is a type of artificial intelligence that creates new content based on what it has learned from existing content.

生成式人工智能是一種人工智能，它能根據從現有內容中學到的知識創建新內容。

The process of learning from existing content is called training and results in the creation of a statistical model.

從現有內容中學習的過程稱為訓練，其結果是創建一個統計模型。

When given a prompt, gene AI uses the model to predict what an expected response might be.

當收到提示時，基因人工智能會使用模型來預測預期的反應。

Essentially, it learns the underlying structure of the data and can then generate new samples that are similar to the data it was trained on.

從本質上講，它可以學習數據的底層結構，然後生成與訓練數據相似的新樣本。

As previously mentioned, a generative language model can take what it has learned from the example it's been shown and create something entirely new based on that information.

如前所述，生成式語言模型可以從所展示的示例中吸取經驗教訓，並根據這些資訊創造出全新的內容。

Large language models are one type of generative AI since they generate novel combinations of text in the form of natural sounding language.

大型語言模型是生成式人工智能的一種，因為它們能以自然語言的形式生成新穎的文本組合。

A generative image model takes an image as input and can output text, another image or video.

生成式影像模型以影像為輸入，可輸出文本、另一幅影像或視頻。

For example, under the output text, you can get visual question answering while under output image, an image completion is generated and under output video, animation is generated.

例如，在輸出文本下，您可以獲得可視化問題解答；在輸出影像下，會生成影像補全；在輸出視頻下，會生成動畫。

A generative language model takes text as input and can output more text, an image, audio or decisions.

生成式語言模型將文本作為輸入，並可輸出更多文本、影像、音頻或決策。

For example, under the output text, question answering is generated and under output image, a video is generated.

例如，在輸出文本下生成問題解答，在輸出影像下生成視頻。