To machines, you can teach the most complex computing algorithms. But, can you teach them what common sense is? To many, this could be a straightforward “Yes.” And we don’t deny that freedom. However, putting common sense into AI and machine learning is still a distant idea, especially when you are dealing with kids, patients, cars, and of course, the peer autonomous robots. To save humans from the fallouts with machines, Allen Institute for Artificial Intelligence (AI2) introduced the ambitious Project Alexandria. This is a new research initiative at AI2 to focus on the ‘common sense’ aspects of dealing with AI and machine learning.
The Purpose of Project Alexandria
AI2 has confirmed that Project for Common Sense in AI is a long-term objective. The purpose of Project Alexandria is introducing a set of standards to help AI devices leverage common sense capabilities. Why Alexandria?
Well, Project Alexandria could become the ‘lighthouse’ for AI paths in the future.
It could be not playing with fire, holding a baby softly, drying the towel using a rope, or punching a hole into the wall using a jackhammer! The end results— keep the world of AI safe, and risk-free for human movement and AI-human interactions.
What’s the Engine Driving Project Alexandria?
Project Alexandria would be developed further on novel Crowdsourcing Methods to acquire natural common sense capabilities in AI and machine learning techniques. Currently, Alexandra would work on bringing the common sense capacities in AI by integrating Machine Reading, NLP, computer vision and crowdsourcing within a single, go-to statistical platform.
The data feed would come from three available sources:
- Crowdsource locomotion/ common sense statements
Functions that Would Benefit from Project Alexandria
- Healthcare, pathology and baby care
- Book-keeping, accounting, and mathematical research
- Mining and shipping industries that involve working in precarious, life-endangering environments
- Fire-safety, rescue missions
- Counter-terrorism squads in heavy civilian areas
- Drones and unmanned submarines
The potentials of working with machines displaying common sense are limitless. We would keep an eye on how Project Alexandria could make working with AI fun, smart and ‘sensible’.