Will Open Innovation be the Future of Artificial Intelligence
- By 2030, China intends to become the world’s premier AI innovation center and its AI industry is targeted to exceed 148 billion dollars. To support continued primacy in this industry, China plans to create leading innovation and personnel training bases, while constructing more comprehensive plans for legal, regulatory, ethical, and policy frameworks.
- The Canadian Government in March this year announced it would invest 125 million dollars in Pan-Canadian Artificial Intelligence Strategy to mark its position as world leader in AI.
- In India, nearly 300 start-ups use some form of AI, according to Tracxn, a start-up tracker. Among dedicated AI-only Indian start-ups, 23% are working on providing solutions to multiple industries, 15% are in e-commerce, 12% in healthcare, 11% in education, 10% in financial services, and the rest in fields such as retail and logistics, according to a 2017 report by Kalaari Capital, a venture capital firm.
Slowly, there is an enormous cultural shift in the last few years where machines are replacing man. For instance, in 2015, a 120-year-old biological mystery was solved by a machine. This AI developed a new scientific theory on the regeneration process of the flatworm. The scientists and biologists at Tufts University programmed the AI to analyse data using trial and error method. For Artificial Intelligence to productively interact with its environment, they heavily rely on the programmers capability. For an AI to successfully run in the ecosystem, there are three main requirements.
The fundamental challenge in AI is to write algorithms that work. Data scientists specialized in programming these are few in number. There are open AI platforms, one by Elon Musk and lots of these algorithms are available open-source. Apache Mahout is a library of scalable machine learning algorithms. Tensorflow is an open-source software library for numerical computation Intelligence. H20 is an open-source artificial intelligence tool which is business-oriented to help you make decisions from data and enables the user to draw insights. The real challenge is, even with easily available Artificial Intelligence programmes, a lack of ability to modify them and produce optimum functionality.
AI techniques extract more value for huge data. For AI to detect patterns and record behaviours, it needs to be fed magnanimous amounts of data. In the current digital economy, data is more valuable than ever. Companies like Amazon, Google, Facebook, Uber, Tesla, Microsoft amongst others generate this data and have it readily available for AI. While other companies have smart algorithms they lack data, consequently, precise results are far-fetched. Future predictions suggest that data could be traded similar to oil. If so, there will be data market places which will rule the future where the data provider and data processors will be the beneficiaries.
3. Computational Power:
To an extent, this issue is alleviated by the cloud. The computational power is being increased every day as per Moore’s law. Recently Amazon released Amazon Machine Learning. This will help with access to machine learning for medium and small enterprises. Unlike CPU for normal computing, deep learning with a lot of data requires GPUs to solve complex problems. If the cloud can support mainframe AI frameworks like TensorFlow or integrate with PaaS services, we will be looking at cloud computing powered by Artificial Intelligence.
AI has great use cases namely personalization, prediction, optimization, anomaly detection and unstructured language processing. Use cases are applicable to manufacturing, automotive, healthcare, energy, and other industries. Unless AI is democratized the usage, over a period of time will prove to be expensive.
In order for this to be successful in the future, it is important to democratize data availability. Open Innovation in AI is inevitable. Algorithms, data and computational power have to become an open commodity. The next wave of Artificial Intelligence will hit the smaller SMBs and Enterprises. The aggregate of the data from these enterprises will only help them to add value to their experience.
October 26, 2018
September 17, 2018
August 31, 2018