上海贵族宝贝龙凤楼

ChatGPT, a chat robot based on artificial intelligence technology, has recently become a hot topic. According to the latest survey, 50% of enterprises in the United States are already using ChatGPT, and 48% of them even let it work instead of employees. This news triggered a heated discussion among netizens, and everyone discussed and chatted.

In this digital era, the advantages and values of ChatGPT also appear. As an advanced technology application, ChatGPT can improve the efficiency and quality of enterprises and individuals, save time and cost, and create more value and possibilities. For example, in the field of customer service, ChatGPT can provide users with services and solutions 24 hours a day. In the field of code, ChatGPT can quickly write high-quality code; In the field of copywriting, ChatGPT can generate attractive copywriting according to the target audience and scenes, and so on. Moreover, the increasing popularity of ChatGPT also shows that the demand for intelligent services in human society is growing.

However, we also need to see the problems and challenges existing in ChatGPT. As a robot based on data and algorithms, ChatGPT is not perfect. It may have errors, deviations, loopholes and so on, and it is difficult to control its behavior and consequences. For example, in terms of data security, ChatGPT may disclose or abuse the private information of users or enterprises; In terms of ethics, ChatGPT may violate or destroy social norms and values; In terms of legal liability, ChatGPT may cause or participate in some illegal or infringing acts and so on.

Break a million users in a week and have 100 million users in two months.

Of course, ChatGPT needs to constantly improve its technology and functions in the future development, especially in data security and ethics. ChatGPT developers need to seriously think about chatting.

In this ever-changing era of science and technology, the appearance of ChatGPT has brought many new opportunities and challenges to mankind. As human beings, we need to master the control ability of science and technology, use it with the right attitude and methods, and keep up with the development and changes of science and technology to maintain our competitiveness and innovation ability.

Therefore, ChatGPT is both a friend and an enemy of mankind. It has brought us unprecedented convenience and efficiency, but it also requires us to use and manage it more rigorously and carefully. I don’t know if you are expecting or resisting the pressure that AI will bring us in the future.

"

Ambarella’s CV3 AI domain controller SoC series (hereinafter referred to as "AMBA", NASDAQ: AMBA, a semiconductor company specializing in AI vision chips) was awarded the annual automobile product award in the 2022 Electronics Industry Award, and was highly praised in the annual semiconductor product category.

The Electronic Industry Award is sponsored by British @CIE magazine, which aims to recognize the best professionals, products, projects and companies in the entire electronic industry.

In the annual "Electronics Industry Awards" of "Components in Electronics", a well-known British media (magazine/website), Ambarella won the "2022 Product Award" by virtue of CV3 automotive AI domain controller SoC.

Anba’s award-winning CV3 series realizes centralized AI sensing processing (including high-pixel vision processing, millimeter-wave radar, laser radar and ultrasonic radar), multi-sensor deep fusion and regulation and control of autonomous vehicles by integrating multiple sensors on a single chip. As a result, ADAS L2+ to L4 automatic driving systems have a higher level of environmental awareness.

The selection method of "Electronic Industry Award" is to invite experts in the industry to vote online, and add the votes of an independent expert jury to finally get the winner. The award has entered its fifth year and will soon become an important activity in the industry, opening the door for new business opportunities.

About Anba

Anba’s products are widely used in artificial intelligence computer vision, video image processing, video recording and other fields, including video security, advanced driver assistance system (ADAS), electronic rearview mirror, driving recorder, driver and cabin intelligent monitoring, intelligent car unmanned driving and robot application, etc. Anba’s high-performance, low-power AI processor provides ultra-high definition image processing, video compression and powerful neural network processing. It can extract valuable data from high-resolution video and radar information, and it plays an important role in the fields of intelligent perception, sensor fusion and central domain control processing system. For more information, please visit www.ambarella.com.

How many content creators keep up late to cut videos of liver explosion? Have you ever thought that one day we can get rid of this heavy work and be done by AI instead? In fact, judging from the development status of AIGC industry, the above imagination is entirely possible to become a reality, and it may be in the near future.

From the perspective of Xiaobian, the content editing industry will go through three stages of development, and we are currently in the initial state of the second stage.

The first stage: pure manual editing

In the traditional sense, it is the manual editing of staying up late to burst the liver. The quality of editing depends on the level of editors and the time spent. This stage is characterized by "high labor intensity" and "a lot of repetitive work". It is precisely because of this pain point that with the progress of technology, the pain point of "excessive physical labor" was first solved by AI. After all, what AI is best at is simple and repetitive labor.

The second stage: manual +AI collaboration stage

AI is good at simple and repetitive movements, but AI is not only good at simple and repetitive movements. So the development of video editing industry has entered the stage of "manual +AI collaborative editing". From the clipping used in our daily life to some commercial clips in the toB field, we can see that AI smart clips are active among them.

Watson, an artificial intelligence system of IBM, edited the trailer for the sci-fi thriller movie Morgan as early as 2016. The technical team collected 100 thriller movie trailers for Watson to machine learn. Watson’s unique training model tagged the details of these trailers, such as picture, audio, scene connection, sound configuration, etc., decomposed and learned one by one, and assigned them to its own creative model. Finally, Watson successfully used the trained data model to select materials from 90-minute movies, and edited them into a 6-minute trailer, which was successfully put into use for film announcement.

Up to now, this kind of intelligent matching tool platform has evolved to the extreme, and ordinary users can simply use AI function for intelligent editing. On the local AI intelligent graphic-to-video platform, one frame per second is the highest, and the user inputs the edited copy. AI will automatically perform intelligent picture matching from the platform’s powerful online material library, presenting its own intelligent video generation results in seconds, and matching all-round details such as pictures, dubbing, subtitles, etc. in one stop. Users can directly export and use it with simple confirmation, especially in the production of popular science content and simple sharing content with low editing difficulty.

At the same time, AI, as an auxiliary tool of intelligent editing, is also blooming in the development direction. At present, all kinds of AI-enabled editing software emerge in an endless stream, some of which support intelligent graphic-to-video, some of which focus on recording and cutting, and some of which can simulate the voice of the user for dubbing modification and adjustment. The "scoop" can also be quickly corrected through AI. In short, the current AI intelligent editing industry is in a thriving development trend. It can be said that with the progress and integration of technology, AI will be fully involved in the field of intelligent editing, and our video editing work will be greatly released.

The third stage: pure AI editing stage

In the future, AI will run video editing in the whole process, even the currently difficult "audit" work. This is a bold idea put forward by Xiaobian, which has certain subjectivity. The main idea of this view is that the compliance and rigor of AI creation are unmatched by human editing. At present, AI editing is still in the primary stage, so human assistance is needed in content review and sensitive screen screening. However, the speed and capacity of AI learning is far beyond that of human beings. With a large number of legal provisions and review rules being learned by AI, the whole process of video editing will be completed by AI.

AI can even avoid problems such as audit omission caused by human fatigue and negligence. Even, while recording and cutting, real-time synchronization, the completion of shooting is a piece of export, which is not completely impossible. In addition to the research and development of AI automatic editing software, there may be AI shooting intelligent hardware devices, workstations and other products in the future, so as to maximize the convenience of video production. The development of AI technology makes us more daring to imagine its bright future.

In addition, what really makes Xiaobian feel that the intelligent editing industry will eventually rise vigorously is that the existing problems "seem" are not problems, and AI will eventually break them one by one with strong technical strength. When the AI industry was just beginning to develop, we were worried that AI would replace some human jobs. At that time, the prediction given by the technology industry was that "art, creativity and other jobs that need to exert human imagination will not be replaced".

However, at present, AI’s original painting has become a reality, and the field of artistic creativity has become the first industry to be impacted. It can be seen that AI’s "imagination" is far beyond our imagination, and the speed of technology iterative update is much faster than expected. The future is really coming.

During our daily tabulation, it is often used.Find the work of reference related data.

In fact, this is also one of most small partners in touch.

For example, according to the corresponding results of the employee name, according to the number of item numbers in one item, the corresponding sales, etc.

Today, I will introduce you to a very powerful function combination to achieve this feature.

They are INDEX + MATCH combination.

The specific application scenario is as follows:

01 正 正 查

The correct point here is:Find reference target values ??from left to right.

That is, the lookup value is on the left and the return value is on the right.

As shown below, it is the performance watch of the company employee in the first quarter. We now want to find the corresponding performance according to the name.

Write the following formula in [F2] cell:

=INDEX (C2: C7, Match (E2, B2: B7, 0)))

Formula analysis:

First look up Liu Yan in the [B2: B7] in the [B2: B7], return to the number 2;

This number 2 is then used as the second parameter of the index function, meaning returning to the value of the first parameter in the cell area in the cell area in the cell region.

There are three parameters: Match functions:

Match (lookup value, find area, lookup type)

The MATCH function is specifically used to find a value, where you are in a region or group, and return a number.

This number that is usually returned as a parameter of other functions.

Here, the returned numbers are used as the parameter of INDEX.

If the lookup type of the third parameter is 0, it means an exact lookup.

INDEX usually has three parameters:

Index (Returned area, line number, list number)

Since the first parameter is a single line or a single column, the third parameter can be omitted.

If you don’t illegally, the formula can be written as this:

=INDEX (C2: C7, Match (E2, B2: B7, 0), 1)

02 reverse lookup

The reverse refer to it here is:Find reference target values ??from right to left.

That is, the lookup value is on the right, the return value is on the left.

Also as examples above, for example, we want to find Liu Yan’s employee number.

The formula is as follows:

=INDEX (A2: A7, MATCH (E2, B2: B7, 0)))

Formula analysis:

The formula of this formula is similar to the formula of the forward look, but the first parameter of the index function will change the returned area into the [A2: A7] area.

It can be seen that whether it is looking for a look, it is the same.

Just change the area of ??the first parameter of the index function into the area that needs to return.

03 two-way lookup

The two-way refers herein is:Find the reference target value from the horizontal and vertical two directions.

As shown below: It is a quarter employee payment message, and now I need to find how much Liu Yan’s salary in March is.

The formula is as follows:

=INDEX (B3: D8, Match (G3, A3: A8, 0), Match (H2, B2: D2, 0))

Formula analysis:

First look up Liu Yan in [A3: A8] in the vertical position in [A3: A8], return to the number 2;

Then use the second Match function to find the horizontal position in [B2: D2] in this area, return to the number 3;

Finally, the two numbers are respectively used as the second and third parameters of the INDEX function, meaning returning to the INDEX first parameter [B3: D8] zone "Chapter 3 of the third row in the area of ??this intersection cell, namely: 8606.

Because the two-way lookups here need to involve the position in both directions, we are hereUse two MATCH functions to locate the lines and columns.

04 more conditions

Multi-conditional lookup refers to: Find back to the corresponding target value based on 2 or more conditions, which is often used in our daily work.

As shown below: It is a sales list of all item numbers in all regions. We want to find how much the number of item numbers BH003 in Nanjing is.

Condition 1: Nanjing;

Conditions 2: The item number is BH003.

Write the following formula in the [G2] cell:

=INDEX (C2: C7, Match (E2 & F2, A2: A7 & B2: B7, 0))

Formula analysis:

The basic frame of the formula is the same as above, just first connect two conditions into a new condition form with a "&" text connection;

Then follow the ideas looking forward to the looks to the lookup and return to the sales volume in the [C2: C7] area.

Match function Find the location of "Nanjing BH003" in the newly generated array:

Match ("Nanjing BH003", 0)

Returns the number 3, then this number 3, as the second parameter of the index function, means returning the INDEX first parameter [C2: C7] The value in the cell in the cell area, that is, 72.

?

Because this is an array formula, the formula needs to be ended in the [Ctrl + Shift + Enter] three key. The braces outside the formula will be automatically generated. For small friends who use Office 365, you can end up with a three-key.

Of course, the multi-conditional lookups here can also be applied at both horizontal and vertical two directions.

In addition, no matter how many conditions, just connect multiple conditions to a new condition with a "&" text connector first, then apply the above routine.

05 interval lookup

The interval lookup refers to:If the lookup value is in a certain range, the corresponding target value is returned.

As shown below, the number of bonuses from the employee is required according to the performance interval.

Write the following formula in the [G2] cell:

=INDEX (J2: J5, Match (F2, {0; 41; 81; 121}, 1)))

Formula analysis:

This formula is especially the application of the second and third parameters of the MATCH function.

If the third parameter is 1 or omitted, the MATCH function will look for the maximum of less than or equal to the first parameter, andThe value in the second parameter must be sorted in ascending order.

The maximum value of less than 65 is 41, 41 in this constant array ({0; 41; 81; 121}) is 2, that is, the number 2 is returned to the number 2, and then the index second parameter is finally returned 2000.

Of course, here we can also write this constant number in an auxiliary column, which may be better understood.

As shown below:

06 blurry look

The fuzzy look is to realize the target value by combining with the wildcard ("?" Question mark and "*" asterisk).

As shown below: We want to find a surname Xu to open his head, and how much is the employee number of three characters.

The formula is as follows:

=Index (A2: A7, Match (E2 & "", B2: B7, 0))

Formula analysis:

Here is the key to the Match function supports the usage of wildcard.

Everyone must pay attention to:

"?" Question mark wildcard, represent any single character.

"*" Asterisk wildcard, represents any number of characters.

Here we used two "??" question marks, represent any two characters.

07 summary

Today, we have learned two powerful function combinations (INDEX + MATCH)Six kinds of findings commonly used in work:

? Positive lookup;

? Anti-(reverse) lookup;

? Two-way lookup;

Multi-conditional lookup;

Area lookup;

? Fuzzy lookup.

What is the most look for you usually use? What do you have any questions about VLOOKUP?

Welcome to usmessageOh ~