There is just too much research on machine learning and artificial intelligence (AI) for anyone to read. The goal of this column is to collect and explain some of the most important recent findings and studies.
This week, scientists ran an intriguing artificial intelligence experiment to forecast how platforms such as food delivery and ride-sharing services influence the economy.
A group of academics tested a system that can read tree heights from satellite photos and forecast startup success using public online data, while a team from ETH Zurich built a system that can read tree heights from public web data.
The work on the market-driven platform builds on the open source research environment for figuring out how artificial intelligence might help with economic policy. A report released in March described the new research in full.
The purpose was to look at two-sided marketplaces such as Amazon, DoorDash, and TaskRabbit, which have more market strength as demand and supply grow.
The researchers used a computer programme to teach a system to recognise the impact of platform-to-consumer interactions.
Reinforcement learning is a technique for predicting how a platform will perform under various design scenarios.
During economic downturns, the researchers discovered that a platform built to maximise revenue raises fees and extracts more revenues from consumers and sellers at the price of societal welfare.
When platform fees are fixed, they discovered that a platform’s revenue-maximizing incentive is in accordance with the larger economy’s welfare concerns.
The technology they intend to open source might serve as a basis for a business or policymaker to examine the platform economy under various situations, designs, and regulatory concerns.
Skopai researchers claim to be the first of its type to be able to estimate a startup’s capacity to attract investors based on publicly accessible data.
Data from startup websites, social media, and business registers, according to the coauthors, may be used to forecast outcomes.
To read our blog on “Using an AI agent to solve puzzles, gamify the known and unknown” click here.
