It is still a summary of iResearch's research report on China's artificial intelligence industry in 2018. This issue will summarize and share the analysis of the development trend of the artificial intelligence industry in this report.
1. There is still a lot of room for development of complete behavior planning or decision-making of things.
Most of the current artificial intelligence models are currently based on deep learning algorithms. But this also brings a lot of limitations. Here we need to compare the difference between artificial intelligence and human intelligence.
- Artificial intelligence based on deep learning needs to use a large amount of data to build a model and then solve a small problem.
- Human intelligence is learning to solve big problems with very little data.
- Humans can rely on their own observations and judgments to form the final value decision. AI capabilities such as machine speech recognition and computer vision are still difficult to support the understanding and judgment of things at this stage.
Therefore, in the next stage, the important development direction of artificial intelligence technology can be summarized as follows:
- How to reduce the amount of data required to train the model in the process of training deep learning, or use some data that does not require multiple labels
- Continuous learning, or effective self-adaptation to changing environments in the process of deep learning
- At present, many artificial intelligence technologies still need to standardize a specific scene, how to make machines can use their own knowledge to solve problems in different scenes like humans
- Interpretability is still a problem that needs to be solved in the current deep learning process. The existing technology is to visualize deep neural networks.
2. Looking at the entire industry, in addition to the progress of algorithms, the construction of products, services, and markets is still a barrier to the development of the artificial intelligence industry.
The implementation of artificial intelligence technology often involves hardware equipment transformation for specific business scenarios, software integration, and deployment of local computing facilities. The actual efficacy of algorithms and technologies requires targeted development based on a deep understanding of customers' real business scenarios. The important factors affecting the large-scale application of artificial intelligence technology can be summarized into the following five points:
- Academic frontier algorithms - can be judged based on the publication of conferences and journal papers, or the results of related competitions
- Marketing - Resource channel construction and sales system expansion can accelerate technology landing in vertical business fields
- Engineering technology - The implementation of engineering technology covers architecture construction, system coordination, process control, quality monitoring and algorithm optimization that fits customer business scenarios
- Customer service - customer service needs to meet the requirements of pre-sale, middle and post-sale, timely response to demand, comprehensive solutions, and problems can be repaired in time
- New hardware products - ease of use, fluency and stability + meet actual business needs
3. Will artificial intelligence produce the mass unemployment we predicted before?
First of all, it must be acknowledged that the development and upgrading of machine learning and deep learning technology will definitely affect the job market gradually. But the impact here is more focused on repetitive work. But people will also be liberated from repetitive tasks and spend more time and energy on creative activities. In other words, artificial intelligence solves the labor crisis . The following is a summary of the impact of artificial intelligence on the job market:
- Simple, repetitive work, machines will definitely replace humans. (This type of work is mental decision-making that does not require complex thinking or just pure manual labor)
- Human-machine collaborative work will gradually increase, which is also the mainstream relationship for a long time.
- Machines cannot replace humans, for jobs that involve empathy or creativity. At the technical level that artificial intelligence cannot complete understanding or inference, there is no way to replace similar human work for the time being.
4. What is the core value of artificial intelligence
Artificial intelligence is the realization of the effect of human thinking through intelligence. From a macro perspective, this effect is reflected in the intelligent society and intelligent economy. That is, artificial intelligence will greatly improve the labor-intensive, simple and repetitive traditional economic operation that relies on labor creation. model, and rely on this economic model to build an industrial system of interconnection of all things and intelligent collaboration, and create an international leading intelligent society. From a micro level, artificial intelligence will replace traditional labor and bring new production methods to improve production efficiency and reduce costs, thereby improving corporate efficiency and improving people's work and life. And as machines get smarter, we will eventually achieve Humanistic AI, which is humanistic forms through machines and extensions of human intelligence in that form. The specific value can be summarized into the following four points:
- Intelligent society: While building intelligent industries, realize the intelligent upgrading of traditional industries.
- Smart economy: Utilize human intelligence and artificial intelligence to give full play to the combination of physical strength and knowledge, and produce products and services with high knowledge added value and high knowledge density
- Smart commerce: reducing the cost of goods and services
- Can only work: Alleviating the labor crisis and improving the freedom of life and work