AI has been in the spotlight over recent years. It is also changing the business environment and the competitive landscape. According to a survey by McKinsey in 2017, the industries with a high level of digitalization, such as IT, telecommunication, automobile, assembly and financial services, will achieve greater success if they become early adopters of AI. An analysis of 800 occupations with 2,000 jobs found that 60% of the occupations could automate at least 30% of tasks but less than 5% of occupations could achieve full automation. Labor work in a highly stable and predictable environment and jobs such as data collection are the most likely to be impacted by AI.

Crowdmining Platform is the New Trend in AI

In the past, AI focused on problem-oriented software and tools. In addition to the combination with domain knowledge and systems, it is necessary to write rule-based programs as solution. As the technology moving on, algorithms replace handcrafted codes to achieve more creative solutions.

Organizations and corporates have limited resources at their disposal in terms of the internal development of algorithms. However, the recruitment of external resources can enable access to creative algorithms from all parties in an endless back-and-forth cycle, in order to derive the solutions with infinite creativity. This is why platforms for crowdmining have begun appearing in the world of AI.

The process of crowdmining involves posting problems with relevant data on the platform in order to invite innovative ideas on algorithms. Interested parties can come up with solutions, and the optimal proposals can be determined on the basis of an objective evaluation mechanism. The persons who came up with the best solutions will receive incentives or transfer the ownership of their algorithms to the parties who need to develop feasible approaches to the problems addressed. The biggest advantage for the problem owners or companies is that they can find solutions by relying on external talent, at a lower cost and in a shorter timeframe. In fact, they can even integrate all the solutions via community sharing and collaboration to achieve the best solutions.

Crowdmining has contributed to the development of AI algorithms, the expansion of the start-up ecosystem and the exploration of new business models. For example, TopCoder, which was founded in 2001, is an online competition platform for developers and programmers. Kaggle, a crowdmining platform, offers a mechanism to host competitions in data science. The company also offers consulting services such as data preparation and problem definition. The platform provides functions such as submission evaluation, intellectual property management, job referrals, community operations and cloud computing. Competitions can be held as commercial or not-for-profit. The purpose is for companies and organizations to identify specialized algorithms for thorny problems. Data scientists who solve problems can access data sets and find ways to commercialize their algorithms. Both parties can benefit from the sourcing of talent and business leads. Therefore, Kaggle has been hugely popular with financial, new economy and software companies.

Crowdmining Platform to connect industry players and academics

According to an analysis of TopCoder by Forrester, an international research firm, from historical data, it took an average of 10.8 months to recover investments and the return on investment was 113%. More importantly, the platform allows each software development manager to handle up to 30 projects per annum (vs. 10 in the past). The development leadtime also went down from an average of nine months to three months. In fact, Kaggle claims that the algorithms achieved with collective efforts can improve the performance by 40% compared with original algorithms. In sum, it is beneficial to derive better algorithms and solutions through the crowdmining platforms.

With the support from the Department of Industrial Technology, Ministry of Economic Affairs, the Industrial Technology Research Institute (ITRI) has been thinking about the positioning, mechanism and functionality of a crowdmining platform in Taiwan. A survey conducted by the ITRI on industries and academic institutions suggests there is a strong demand from industries to use AI to enhance competitiveness. However, the academic teams specialized in AI algorithms have trouble commercializing their intellectual property. Therefore, the crowdmining platform could be a way to allow companies to articulate their requirements, provide data and invite the best models and solutions from the academics. 

In summary, the crowdmining platform in Taiwan should serve as a marketplace for bespoke AI algorithms. This will provide a venue for made-to-order algorithms of the corporate world. It will allow more companies to use AI to resolve real business issues and encourage the AI industry to stand in its own right.