Abstract
As a member of the oldest art family of human beings, drama is also a very important part of human civilization. With the development of technology, more and more types of drama film and television appear. In today’s society, bad films and commercial films are emerging one after another. The intervention of capital has brought a very poor experience to moviegoers and is not conducive to the development of the entire field of drama and film. The further development and scientific promotion of the field of drama film and television has become a relatively important issue. In this context, this paper takes this as the starting point, integrates big data and artificial intelligence algorithms, and aims to analyze the promotion and influence in the field of drama and film. It has existed for a long time in the fields of physics, biology, environmental ecology, and other industries such as military, finance, and communication, but it has attracted people’s attention due to the development of the Internet and information industry in recent years. The article introduces the transition of the mainland film system and the manifestations of conflict in plot art and film and television plots. It uses the convolutional neural network algorithm in deep learning to build the system. This paper adopts an online survey, distributes it to moviegoers nationwide, and collects 4,000 valid samples for analysis of moviegoers’ occupation and monthly income. In terms of the occupational distribution of the audience, the proportion of white-collar workers in the company is the highest at 40%, the student group only accounts for 17.4%, and the retired worker group accounts for the smallest proportion, at 0.6%. To sum up, the structure of Chinese films has shifted from a younger, student-dominated, middle-income group to a mature, workplace-oriented, middle- and high-income group. It has successfully completed the analysis of the promotion and impact of big data and artificial intelligence in the field of drama, film, and television.
1. Introduction
The movie depends more on the “narrative style,” which is all about the “narrative style” of the text. With the advancement of science and technology, the reform of the film system and the improvement of people’s living and cultural level, people have gradually entered the era of consumer culture, which has put forward higher requirements for audio-visual aesthetics. Mainland Chinese movies also follow the global trend, following the development of Hollywood and other movies, and there are more “spectacle” movies that seek visual pleasure. Among them, style films headed by the “fifth generation” swept across China, announcing the arrival of the “spectacle era,” and mainland Chinese films also shifted from narrative films to “spectacle” films. Spectacle movies refer to a type of movie that spends a lot of time and money on special effects. It was first produced by Hollywood in the United States.
Artificial intelligence is widely used in machine vision, fingerprint recognition, face recognition, retina recognition, iris recognition, palmprint recognition, expert systems, automatic planning, intelligent search, theorem proving, games, automatic programming, intelligent control, robotics, language and image understanding, genetic programming, etc.
Artificial intelligence is the use of artificial intelligence way to perform mechanical intelligence or to simulate human intelligence on machines, or to give humans machine intelligence. Since artificial intelligence occurs in machines, it can also be called mechanical intelligence. And because mechanical intelligence simulates human intelligence, it can also be called simulated intelligence. The phrase AI is now used as a designation for the subject of “research into how to achieve in machines the intelligence of mankind.” In this sense, it can be defined as: AI is the study of how to establish smart robots or clever to simulate, extend human smart. Ordinary people think that AI is the research of how to make machines have the functions to solve the problems. In short, it allows machines to perform tasks that require human intelligence or even more intelligence than humans.
2. Related Work
In the progress of technology, the usage of this research has become more and more extensive, and they are widely used in military, medical, environmental monitoring, and other fields. Dubé et al. proposed a converged creative pattern to solve the challenges still facing society in these days. The approach is to build bridges between industries and disciplines around comprehensive goals on both sides of the socio-economic divide, enabling innovation with the goal of creating wealth. But his research did not use more matrices for sample collection [1]. Through the mining of public data and personal data, Chehri et al. attached equal importance to process and results, and combines task learning with autonomous learning. He combed the help of big data and artificial intelligence to the personalized teaching of college teachers, the personalized management of college educators, and the personalized learning of college students [2]. Lan constructed a model of individualized learning for university students, which consists of two main subtopics: the individualized knowledge development model of university students and the individualized education model of university educators. Both models contribute to each other and work together in the education process of university students [3]. Li et al. have produced a big data and artificial intelligence framework for intelligent and personalized air pollution monitoring and health management, aiming to contribute to Hong Kong. His research has been widely used by the government in air pollution monitoring and has contributed to the economic development of Hong Kong. However, his research results cannot be extended to other cities and regions, and have regional limitations [4].
With the development of drama, more and more types of film and television dramas appear. Chen and Yang analyzed the present status of Chinese film and dramas. He studied the creative creativity in film and TV. He conducted a survey on the 20 most popular Chinese movies in 2019. The results of his experiments show that the public approves of the quality of the films in 2019 [5]. Wang et al. expounded the importance of film and television production project management, and provided a solution for film and television network project management by using modern computer technology [6]. MacKenzie studied the classic episodes of the RAF bomber command in the film and television series. However, his research only stays at the level of art form and artistic influence, and does not investigate the influence of content on the dissemination of film and television itself [7].
3. Methods on the Promotion of Drama and Film
3.1. Plot Art
The creation of film and television works is first and foremost based on texts, especially literary texts [8]. This literature is mainly manifested in the plot. Therefore, the story with twists and turns can steadily attract the attention of the audience and achieve the artistic effect expected by the creator [9]. Then, how to subtly define the plot of the story has become a common problem faced by creators. In drama, dramatic conflict often manifests as: (1) Sharp and intense: in drama, some simple contradictions often form vivid conflicts. The eventual flare-up of the conflict was particularly strong because both sides of the opposition had sufficient leverage. (2) High concentration: dramatic conflict is the manifestation of social conflict within a specific time and space. (3) Intense progression: A dramatic conflict must be exciting and fluctuating, so the audience is always in tension and anticipation. (4) Twists and turns: dramatic conflicts are often complex and changeable. As the saying goes, there is no conflict without contradiction, and no drama without conflict. Many creators pay great attention to the arrangement of the plot, or carefully organize seemingly ordinary events through time or space, and finally form works of art that can express the theme and eliminate ideas [10]. Then, through the execution of characters, the creation of environment and atmosphere, etc., a well-known project is finally completed.
3.2. Roll-Up Neural Networks
The most widely used convergent neural networks (also called convergent networks) in the deep learning field are those specialized for processing data with grid-like structures [11]. Compact neural networks excel in many application domains [12, 13]. Examples are convolutional neural networks for time series data (which can be thought of as a one-dimensional grid formed regularly over time) and image data (which can be thought of as a two-dimensional grid of pixels). It is a trainable multi-layer network structure composed of multiple single-layer convolutional neural networks. In convergent network terminology, the first parameter of convergence is often referred to as the input (i.e., the data to be processed) [14]. The second parameter is called the kernel function (which can be thought of as a processing function), and the output is called the feature map. In a traditional neural network, the output of a neuron is related to the weights of the input nodes and their connections, but the weights are not reused according to the calculation rules of the type of calculation. That is, when multiplying by an element, the input will not be used again. Parameter sharing exists in feature maps. Each feature map in level 2 will detect the same feature map in level 1 [15]. The same idea is represented as a network-specific weight sharing with bound weights (i.e., weight distribution) as shown in Figure 1:
3.3. Optimization Strategy Algorithm
Let denote the arrangement of the collection of points in the surveillance zone. And it uses a pattern in as semi-radius to approach the overlay pattern of the knots. Then, node is denoted as . is the location grid of this knot [16, 17]. In this kind of expression, the choice of mesh edge size is significant, and it is immediately linked to the precision of overlay [18]. Here is an overlay error formula:
The scope of surveillance is classified as knots:
The chain of relationships between the knot axes and the cells can be represented as shown in Figure 2.
Let the event be the grid point covered by the node , and are the possibility of occurrence, using the logical idea [19]. A logic circuit is a circuit that transmits and processes discrete signals, and uses binary as the principle to realize the logic operation and operation of digital signals. is a Boolean feature of yes or no:
As long as any node in the node set covers grid point , grid point is said to be covered by node set. Then, we can get:
Moreover, we get:
The sum of the grid areas covered by the working node set is the sum of the grid areas covered by each node in the node set, denoted as , then:
If the number of outlets is much larger than the parameters required, still using the old method will make the calculation tedious, and the results are meaningless. A lot of manpower and resources will be wasted on the calculation, which is not worth the loss. The following algorithm is improved:
and are knots greatness and knot usage, respectively. The overlay network is simply the application layer network, which is oriented to the application layer and does not consider or rarely consider the network layer and the physical layer. The results formulas of the two is
The objective of the optimization problem of radio signal overlay is to optimize the network footprint while minimum utilizing the knots. Thus, this problem is a combination optimization problem that needs to consider both network coverage and node utilization. To solve such problems, a linear combination of objectives is usually constructed by weighting. The formula is as follows:
A few definitions are introduced here:
Definition 1. The distance between two points and :
Definition 2. Adjacent grids: supposing the whole group of grids be X. Then, the group of adjacent grids is
Definition 3. The point of the grids: setting the measure besides the axis in the lattice, then the matrix is located isSo, the point isThe first index to evaluate the coverage effect of a grids is the positive rate of the grids. Assuming that represents the positive rate of the grid, we can get: stands for every single gravity factor, and the resulting bias differentiation is used as an estimate for the adjustment of the gravity factor.
As shown in formula (14):The weight can be judged and calculated by dividing the multiple levels of indicators, and the commonly used methods include AHP, fuzzy method, fuzzy AHP, and expert evaluation method. The calculation of the weight parameter has been a big problem. Calculating layer by layer according to the rules can lead to a large amount of computation, which overloads the computer and also makes the money invested in the study to rise significantly. Here, the calculation can be simplified by back propagation method. In mathematics, the partial derivative of a function of multiple variables is its derivative with respect to one of the variables while keeping the others constant (as opposed to the full derivative, in which all variables are allowed to vary), as shown in formula (15).Below:
3.4. Conflict Forms of Drama
Dramatic conflict is the soul of drama [20]. Dramatic conflict can fully link the history of film and television texts with the characters in the script, making them organically combined. The logical arrangement of the historical rhythm can fully express the relationship between the characters, highlight the characters, and promote the development of the plot. In other words, there is no drama without conflict, this is the law that the Creator will obey [21]. As someone pointed out in the play: “Conflict is one of the most dramatic factors in life, so a lot of scenes, maybe most of them, are actually about some kind of conflict.” Dramatic conflict seems to be an important part of shaping the play factor. In other words, dramatic conflict stems from the contradictions in life, the standardization of opposition, and struggle in life. The fundamental difference between dramatic art and other literary arts are the scene of dramatic conflict. A famous playwright once said: “Without conflict, there is no pressure, no crisis, and no drama.” When creating art, creators do not need to deliberately add contradictions, but realize artistic creation according to the historical priorities. Whether internal or external, loose or serious conflicts can unify the plot. It is also closely related to the real life of the creator. As it is often said, art comes from life, but it is superior to life. In life, only social conflict, dramatic conflict, and stage conflict can enter the dramatic stage [22].
4. Experiment on Influence of Big Data and Artificial Intelligence on Drama Film and Television
4.1. Changes in Audience of Movie Viewing
The audience of the movie is the title of the main group who will go to the cinema to watch the movie when it is shown. Then, the main group of people watching in mainland China has changed again. First, the audience is younger. Walking into the cinema, it is not difficult to find that moviegoers are mostly young and student groups. Because they accept the influence of Western culture faster and more, and grow up in the context of the reform and opening up era, they are more willing to enter the theater for consumption. This is not only closely related to the concept of consumption but also to values.
This time, an online survey was taken and distributed to moviegoers nationwide, and 4,000 valid samples were recovered. Due to the rapid development of China’s film industry, more and more low-income people are entering the cinema to enjoy movies [23]. Figure 3 shows the composition of monthly income of sample viewers in 2021:
From the perspective of the monthly income of the audience, the audience with a monthly income of 6,000–10,000 yuan has the highest proportion, at 27.5%, and the audience with a monthly income of 6,000–20,000 yuan exceeds 50%.
The occupational composition of the audience is shown in Figure 4:
From the perspective of the occupational distribution of the audience, the proportion of white-collar workers in the company is the highest at 40%, the student group only accounts for 17.4%, and the retired worker group accounts for the smallest proportion, at 0.6%. To sum up, the structure of Chinese films has shifted from a younger, student-dominated, middle-income group to a mature, workplace-oriented, middle- and high-income group.
4.2. Cascade Detection Method Based on Classification Network
Table 1 compares the detection results of the classification network-based cascade detection method and other methods on the ICPR 2016 test set, and the last row is the detection results of this paper. Compared with the other methods, the detection method in this paper improves the recall rate. Recall is the number of correct numbers found divided by all correct numbers. That is, reduces the number of samples that are easily missed detection, and the accuracy rate fails to reach the highest. Due to the large improvement in recall, the final F-score outperforms all methods except DRN.
The first column has 5 groups of data, and the brackets indicate the number of images contained in this group of data. It can be seen from the table that there are large differences between different images, and the detection results obtained by the same method are also very different. The accuracy and recall rate of A06 and A09 are relatively high, and the detection effect is better. The accuracy rate of A08 and A16 is better, but the recall rate is very low, resulting in a low final F-score.
Comparing the monitoring effects of the classification network-based cascade monitoring method and other methods on the ICPR2016 test set, the ranking is shown in Table 2 according to the F-score from large to small. I refer to an article from Li and Lin and make Tables 2 and 3 based on this article, Artificial Intelligence-Based Semisupervised Self-Training Algorithm in Pathological Tissue Image Segmentation[24].
THE detection way in this article has a certain improvement in both accuracy and recall, so the F-score is the highest. The two cascaded detection methods in this paper are compared in terms of detection accuracy and detection efficiency [25–27]. Since the segmentation accuracy of SegNet is better than that of FCN, FCN uses upsampling. For example, after 5 convolutions (and pooling), the resolution of the image is reduced by 2, 4, 8, 16, and 32 times in turn. The detection accuracy of the two cascaded detection methods on the ICPR 2016 dataset is shown in Figure 5:
It can be seen that the second method improves the precision, but loses the recall, and the final F-score is the same as the first method.
Finally, the time consumption comparison of the three methods used on the ICPR2016 dataset is carried out, namely, the way in this article, the two classic segmentation networks of FCN and CNN. The Table shows that the cascade detection way based on segmentation and classification shortens the detection time in both steps, I refer to an article about a from Qun and Linlin, and make Table 3 based on this article, Artificial Intelligence-Based Semisupervised Self-Training Algorithm in Pathological Tissue Image Segmentation as shown.
It can be seen that this method takes much less time in the first step than the first two methods, which benefits from the integrated improvement of the algorithm in the first step.
For finding out what the number of concurrent connections is, the message transmission speed reaches the highest level and stabilizes. Table 4 shows the test results of node transmission data.
It is represented by a comparative histogram, as shown in Figure 6:
Table 5 shows the central server synchronization data rate test as follows:
It is represented by a comparative histogram, as shown in Figure 7:
It can be seen that the overall trend of the central synchronization system is roughly the same as that of the node receiving system. But when a persistent connection is used, the speed is slower than the node transmission speed. On the contrary, when a nonpersistent connection is used, the speed is faster than the node transmission speed.
Based on the foregoing experiments on data transportation and acquisition in the transport system, it can be seen that there is a great advantage in persistent connectivity in terms of throughput. However, this benefit starts to drop when the number of simultaneous connects exceeds 500, as large connections tend to consume traffic connections for a long time. The throughput can reach its highest value when the number of parallel joins is approximately 500.
4.3. Server-Side Testing
In this test, the ware was placed on a 3rd-generation Linux OS-based platform. The message publishing quality is QoS = 2, that is, there is one and only one time, which is used to simulate the most stringent test environment. The test results are shown in Figure 8:
As shown in the figure, in such a strict test environment, when the network environment is stable, and the number of concurrency is less than 5000, the data delay time is within 1 second, which is sufficient in some small-scale IoT usage scenarios.
The server tests conducted in this chapter are used to explore the impetus and influence in the field of theatrical film and television. It takes 1000, 2000, 3000 analog data, respectively, and analyzes the influence and promotion degree of TV dramas, digital movies, film movies, and dramas according to the algorithm in this paper, as shown in Figure 9:
As can be seen from the figure, the income statement used for the three simulated data basically hovers in the range of 0–4000. The maximum value is 9286 for digital movies with a column of 3000 items of data, and the minimum value is −2258 for film movies with a column of 1000 items. The reason is that with the advancement of technology and the times, the form of film has become a product of the backward era, while digital films can be watched without leaving home. The experimental results are also more accurate.
5. Discussion
5.1. Network Model Configuration of Artificial Intelligence
There are two important files for training the network model, one is the configuration file that defines the network structure, and the other is the solver file that sets the training parameters. This section describes the network configuration file adjusted for training data. The method in this paper is based on the ResNet 101 level. For network data level (define network input), color images can be converted to lmdb data format with higher read performance as network input. In the experiments, the average of the images will be subtracted for training and testing, which will improve speed and accuracy to some extent. Computing the average is simply computing the average of all training samples. Caffe provides mean values of data in lmdb format. The average file value in binary format, the default input size is 224 × 224, and the training set image size is 101 × 101. It cannot perform the next convolution operation so that the training fails, and the input image can be zero-filled to expand the size to 224 × 224. This paper chooses to modify the size of the convolution stride in one of the residual blocks. This allows the network to fit an input of size 101 × 101.
The output is a discrete category expression, that is, the category to which the output pattern to be recognized belongs is a two-category classification.
5.2. Thinking about Risks of Artificial Intelligence Technology
Threats to social order include unemployment, threats to public safety, and threats of militarization. Threats to social morality include inequality in life, threats to the future survival and development of human beings, and threats to the condition of the human subject. One of the inherent reasons of artificial intelligence technology is the uncertainty and uncertainty of artificial intelligence ontology. The rupture of external factors between technical rationality and social rationality, the impact of social groups on risk perception, etc., both are technical dangers of artificial intelligence. Due to the break between the technology and culture, the development of modern technology has always lacked humanitarian reflection. Therefore, when developing technology, human factors should also be considered, and people should be regarded as the bottom line of technological development.
Strengthening the exchanges between humanities and science and technology can be carried out in the following ways: first, when formulating science and technology policies, it is necessary to combine the suggestions of people in different fields, such as the suggestions of people in the field of humanities. Technologists may expect greater use of technology to bring about social change or gain some power through technology, ignoring the dangers that technology can bring to society. Humanities practitioners can analyze technology ahead of time before researching and developing it. For example, philosophers discussed the philosophical aspects of the dangers of artificial intelligence technology, providing references for the research and application of artificial intelligence technology. Second, by enhancing the humanitarian impact of technicians, AI personnel can impart humanitarian knowledge in the research and development of technological applications, enhance humanitarian thinking, and help reduce technicians developing products that pose huge risks to humans. While the power of these proposals to change this reality may be limited, the role of humans in AI technology will guide its development.
5.3. Relationship between Artificial Intelligence and Technological Paradigms
A technological paradigm refers to a mode of solving selected technological and economic problems, and the solutions to these problems are based on the principles of the natural sciences. It must start with technical examples, and since the concept comes from “examples,” we have to start with examples. When scholars put forward the concept of “example,” the definition of the concept itself is not clear. He has tried several times to correct this flaw, but with no good results. So much so that the word “example” can be used to describe both some controversy in the category of metaphysics and the use of different techniques. So let us take this example as “a research tradition that dominates or dominates a particular discipline” for the time being. That is, the example is the supervision of researchers in a certain field as the core content of the community, and it is also the theoretical method of maintaining the discipline as an independent scientific field. In fact, it is the intrinsic quality norm of a discipline. Technology examples are references to example technology developments. Therefore, technology examples can be seen as the guiding existence of a specific technology in the development process. It not only retains the experience content accumulated during the development process, but also applies to the latter, providing direction for growth. In a word, the technological instance is the basic technological theory for the technological community to maintain the content of technological production in the process of technological development.
In terms of characteristics, the characteristics of examples are universal, scientific, historical, regional, and subjective. Because the technological paradigm is the development of the paradigm in the economic technology field, the characteristics of the paradigm also apply to the technological paradigm. But the conditions for this formation still need the following points: (1) An independent subject area; (2) To build a technical community, and the introduction of the subject is also the introduction of the basic theory of the subject.
5.4. Reflections from “Popcorn Movie”
The advancement of film technology has brought us an audio-visual feast, but unknowingly hidden a crisis. In recent years, both the Chinese film industry and Hollywood have been greedy for the cinema. They tend to ignore the humanitarian concerns of movies, stay away from real life, blindly emphasize the excitement of audio-visual dramas, and showcase their techniques. This has put Chinese films on an artistic path of no return. Most of these films are made for registration, but even if they cheat registration, they lose their reputation. For example, various early film promotions earned enough credit for the film’s release. Everyone went to the cinema, and in the end, in exchange for any criticism, everyone felt like they bought a ticket to the cinema but saw a movie that was not worth seeing. In the end, all employees apologise to the public, which is notorious. Even if such a film hits the screen, it cannot win audiences. This misunderstanding deserves our vigilance and reflection. At present, China’s “bad movies” mainly have the following aspects: empty plots, special effects that burn money, and strong commercial atmosphere. All of this requires reflection.
6. Conclusion
Reception theory, as a new paradigm in research, has enabled Western film and television drama research to shift from “author-centered” and “work-centered” to “reader-centered.” In other words, film and television drama works show long-term vitality and vitality through their interdependence with the readers. The article combines big data and artificial intelligence theory with drama film and television, and aims to use this article to contribute to the promotion and development of drama film and television. This paper uses the sample collection method and the convolutional network algorithm, designs a cascade detection method based on the classification network, conducts the evaluation results of the test set, and compares the results of different methods in ICPR 2016. It explores the role and impact of server testing in the field of drama and film. It takes 1000, 2000, 3000 analog data, respectively, and analyzes the influence and promotion degree of TV dramas, digital movies, film movies, and dramas according to the algorithm of this paper. The results show that the income statement used for the three simulated data basically hovers in the range of 0–4000. The maximum value is 9286 for digital movies with a column of 3000 items of data, and the minimum value is −2258 for film movies with a column of 1000 items. The reason is that with the advancement of technology and the times, the form of film films has become a product of the backward era, while digital films can be watched without leaving home. The experimental results are also more accurate. Follow-up experiments can be conducted in multiple dimensions, such as the influence of audiences of different ages on the promotion of film and television dramas.
Data Availability
The data that support the findings of this study are available from the author upon reasonable request.
Conflicts of Interest
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.