AI medicine outlet false fire: bat competes to enter the industry financing record to break again

category:Finance
 AI medicine outlet false fire: bat competes to enter the industry financing record to break again


The logic of investment: Data

Internet giants believe that AI drug research and development is not for no reason. Dr. Pei Jianfeng, a distinguished researcher of the frontier Interdisciplinary Research Institute of Peking University and former director of the national science and technology major special drug design project of new drug creation, told the reporter of science and technology innovation board that in the process of searching for possible landing scenarios of AI technology, drug research and development is easy to be selected because of its wide imagination space.

On the one hand, drug research and development has indeed encountered development bottlenecks, and it needs to continue to develop with the help of new technologies such as AI.

On the other hand, the development of AI technology depends on big data. In this regard, its logic seems to be similar to that of the Internet industry, so it is easy to understand by the investment of Internet giants.

The reporter of science and technology innovation board daily has seen the emphasis on data in the mode of several start-up enterprises. Taking Jingtai technology as an example, one of the reasons why Jingtai technology has won the popularity of Internet venture capital is that its data accumulation is close to Pb level (reporters note: 1PB = 1024tb, 800 human memory is equivalent to 1 Pb). The fund-raising will be used to support the continuous innovation and upgrading of its algorithm with rich real-world data of drug research and development, and challenge more R & D bottlenecks.

Similarly, bat and other Internet giants are proud to enter the Bureau, a large part of the logic is also based on this. They believe that after cutting into the field of new drug research and development with the advantages of algorithms and resources, they can accumulate the advantages of big data, and then iterate the algorithm. In this positive cycle, they will eventually occupy the track; on the contrary, those start-ups that can not quickly open downstream applications in the competition will not be favored by investors due to the lack of strong data support.

Cross border medicine on the Internet: not favored?

However, it is worth noting that this understanding from the Internet industry does not conform to the cognition of the pharmaceutical industry.

Pei Jianfeng pointed out to the reporter of the science and technology innovation board daily that the data is important, but it does not play a fundamental role. In this regard, Yu Xiang, the investment director of Zhongke Chuangxing, agreed. He said that the pharmaceutical industry is serious, and the phenomenon similar to the Internet industry in terms of volume and data should not have existed.

First of all, there is no massive data in the pharmaceutical industry itself. 2C consumer Internet products are easy to access massive data, but the logic of the pharmaceutical industry is completely different. Pei Jianfeng said that every piece of data of drug research and development can only be obtained through experiments by researchers. Due to the progress of the experiment, it is difficult to become massive, and the cost of the experiment is very high.

Dr. Xia Ning, founder and CEO of intelligent chemical technology of industrial enterprises, disclosed the relevant costs to the reporter of science and Technology Innovation Board: the cost of doing an experiment is about 500 US dollars, and the cost of synthesizing a compound is about 1000 dollars. In addition, Pei Jianfeng pointed out that many data are privately owned by enterprises and are not open to the outside world.

Whats more, drug research and development is very complex. For a long time in the future, AI alone will not completely solve the industry bottleneck. Drug R & D is a three-dimensional dimension. Although AI is very important, it is only one dimension. Pei Jianfeng said that if we compare drug research and development to a boat in the sea, AI is a motor, but wind speed, current, course and other factors are also very important.

Based on the above reasons, pharmaceutical practitioners hold negative views on cross-border drug research and development on the Internet. As for the market phenomenon of hot speculation AI drug research and development, Yu Xiang commented to the reporter of Kechuang ban daily that the industry is easy to learn but difficult to master, and chasing after the wind will only damage the reputation of the industry.

In fact, the threshold of drug research and development is far beyond that of the Internet industry. Therefore, this cross-border is not across but climbing stairs. One evidence is the age of the practitioners. Experts with little success in drug research and development are generally over 40 years old, while in the Internet industry, 20-30 years old is the golden age. Pei Jianfeng said.

Hidden worries behind the investment boom

In fact, before the AI boom, using algorithmic tools to improve drug R & D efficiency was a cutting-edge technology in the industry. Pei Jianfengs team began to try AI drug research and development in 2014. Previously, his team used other algorithms to solve the same problem in scientific research institutes.

There are ways to avoid the problem of data volume. Pei Jianfeng said, for example, combining physical chemistry and mathematical principles, summed up some regular laws. Newtons law of gravitation solved a lot of physical problems, but Newton did not sum up the law by big data. What he did was an abstract and conceptual thinking in philosophy, as well as in the field of drug research and development.

On the contrary, those start-ups that focus on AI data have unclear business model and unclear moat.

Huang Li, an investor in the pharmaceutical industry, described what he saw, heard and thought to the reporter of kechuangboard Daily: there are too many enterprises doing AI drug research and development, and I have seen no less than 10 bp. The homogenization competition in the industry is serious and the core value of the enterprise is missing. This reaction will turn into a price war in the business. Moreover, AI is still only an auxiliary tool for scientific researchers, which is helpful to the industry at this stage Limited.

As for when he will consider investing in AI drug R & D enterprises, Huang Li said n years later.

AI pharmaceutical industry is still in its infancy

Science and technology innovation board daily reporter learned that Yingfei Zhiyao is this mode.

We have established the intelligent drug brain platform, which integrates cutting-edge computer-aided drug design technology, new generation AI drug design technology and new drug research and development expert experience, and is a highly integrated human-computer interactive AI. At present, smart drug brain has more than 30 module technologies, such as structure-based three-dimensional molecular generation and optimization, AI molecular inverse synthesis path planning, AI target and molecular drug type evaluation, allosteric drug design and ultra-high-throughput virtual screening, forming a number of mature and reliable processes, providing practical solutions for new drug research and development. Pei Jianfeng said. In the past year, Infineon provided target confirmation, new molecular discovery and optimization services for four pharmaceutical companies and scientific research institutes, and carried out two self-developed candidate drug research and development. However, all the above respondents believe that AI drug research and development is still in the early stage of the industry, and there is still a long way to go before the fruits are picked. In addition, Pei Jianfeng believes that the healthy development of the industry can not be separated from two forces: first, it is necessary to call on the state to attach importance to the construction of data centers, so as to further improve the data and information needed for drug research and development; second, new breakthroughs are needed in the field of AI algorithm. The next generation strong AI algorithm needs to be deeply integrated with domain knowledge in the field of new drug creation, so as to get rid of the dependence on big data alone, says Pei Jianfeng. Source: Science and technology innovation board daily editor: Zhang Mei_ NF2100

We have established the intelligent drug brain platform, which integrates cutting-edge computer-aided drug design technology, new generation AI drug design technology and new drug research and development expert experience, and is a highly integrated human-computer interactive AI. At present, smart drug brain has more than 30 module technologies, such as structure-based three-dimensional molecular generation and optimization, AI molecular inverse synthesis path planning, AI target and molecular drug type evaluation, allosteric drug design and ultra-high-throughput virtual screening, forming a number of mature and reliable processes, providing practical solutions for new drug research and development. Pei Jianfeng said.

In addition, Pei Jianfeng believes that the healthy development of the industry can not be separated from two forces: first, it is necessary to call on the state to attach importance to the construction of data centers, so as to further improve the data and information needed for drug research and development; second, new breakthroughs are needed in the field of AI algorithm. The next generation strong AI algorithm needs to be deeply integrated with domain knowledge in the field of new drug creation, so as to get rid of the dependence on big data alone, says Pei Jianfeng.