'AI industrialization is in the early stage', Quark Scanner expands To B model

News from Duozhi.com on September 23, "The evolution of comprehensive digitalization of human society is a certain prospect. Every practice, thinking, and discussion will be an organic part of this long road." Recently, in Ali At the "AIGC and Scenario Application Innovation" themed seminar held by Quark App, a subsidiary of the Intelligent Information Business Group, Zhou Xiaopeng, Vice President and Party Secretary of the Alibaba Intelligent Information Business Group, said that under the current technology boom of large models, what needs to be paid more attention to is this. What kind of changes and value will technology bring to society in work and life scenarios.

(Zhou Xiaopeng, Vice President and Party Secretary of Alibaba Intelligent Information Business Group)

Cai Yufeng believes that there is actually still a long way to go from technological breakthroughs to being able to form industrial scale in the market, and not all problems can be solved through technology. He said that the industrialization of artificial intelligence has entered the early stage. The practice of artificial intelligence industrialization is essentially to solve the three major problems of quality, cost and scale in every segmented scenario. A closed loop must be realized to truly realize artificial intelligence industrialization

Through perception, analysis, understanding and reasoning, AIGC has become a consensus on its broad application prospects in providing intelligent auxiliary capabilities and solutions. Especially in the fields of efficient office and online learning, AIGC applications as productivity tools will change traditional working methods and learning models.

In many application fields, scanning tools that are frequently needed by users have become a typical scenario for upgrading intelligent experience and implementing AIGC capabilities.

Huang Ruihua, head of Quark AI vision technology, said that in online office and learning scenarios, AIGC has achieved breakthroughs in recognition accuracy, task types and application scenarios. Taking Quark Scan King as an example, the handwritten character recognition accuracy has increased from 97%-98% to 99%, and the complex formula recognition accuracy has reached 99%.

Taking the recognition of complex formulas as an example, let us take a look at how AIGC helps recognition. Huang Ruihua explained the mystery: For complex formulas, AIGC first splits its structure into several modules, and then lets the model process the output of these three modules respectively. Through this separation of ideas, AIGC can solve the complex process of formula

Huang Ruihua further stated that Quark Scanner has formed a capability matrix of scanning capabilities, image capabilities, content recognition and format understanding capabilities, and learning efficiency tools. "The AI ​​algorithm makes the scanned content more accurate and clearer, while providing powerful editing and management capabilities, bringing users a new information service experience."

Huang Ruihua pointed out that the education industry has an obvious characteristic, that is, the knowledge error tolerance rate is low, so the accuracy of identification is very high. Therefore, Quark Scanner has the above capabilities, which brings more possibilities to its business model: extending from To C to To B business lines, providing some basic capabilities for the education industry, etc.

The above is the detailed content of 'AI industrialization is in the early stage', Quark Scanner expands To B model. For more information, please follow other related articles on the PHP Chinese website!