Can ML/AI predict The Future and how Time-Travel is possible?
― H.G. Wells, The Time Machine
When we think of the phrase “time travel,” we usually ponder of traveling faster than 1 second per second. That kind of time travel sounds like something we’d only see in movies like Predestination (2015),Doctor Strange (2016),Interstellar (2014),Primer (2004),Time Bandits (1981) or science fiction books. Could it be real? Sci-tech says yes!
Well-determined predictions that are attempted and verified, yes. Predictions that are almost surely certain to occur, sure. ANNs (artificial neural networks) are great at modelling and using statistics/back propagations/data manipulations and the like to make a decision. Envisioning the future in a widespread sense.
But, Nathan proposed ethical points when he discussed about creativity, originality, etc. We need a sense of creation to imagine and forecast the future. This “sense of creation” requires subjective experience and dynamical, continuous, and self-organizing learning. AI / Machine learning is yet to implement unsupervised learning. They’re also yet to encompass the problem of consciousness and connotation (those are philosophical problems that hampered the mechanization of cognition into AI for years).
Instead, Google’s Deepmind division — which conducts AI research — is giving its AI algorithms an imagination, so they can predict how a situation might play out. The research is published in two papers which says —
“When placing a glass on the edge of a table, for example, we will likely pause to consider how stable it is and whether it might fall,” Deepmind mentioned in a post.
“If Algorithms are to develop equally sophisticated behaviors, they too must have the capability to ‘imagine’ and reason about the future.”
Deepmind(DMG) made the news recently after developing AlphaGo, an AI machine that defeated some of the world’s best players at the ancient Chinese board game Go.
AlphaGo being a good example of an AI “agent” that can plan for the future events quite well. But it only operated in the confined set of rules of the game Go. Thereafter, Deepmind expects to be able to apply the lessons learned to real world examples too.
What does this mean for time travel? Well, according to the theory, the faster we travel, the slower we experience time. Indian Scientists have done some experiments to prove it.
For example, there was an experiment that used two clocks set to the exact same time. One clock stayed on Earth, while the other flew in an airplane (going in the same direction Earth rotates).
After the airplane flew around the world, scientists compared the two clocks. The clock on the fast-moving airplane was slightly behind the clock on the ground. So, the clock on the airplane was traveling slightly slower in time than 1 second per second.
That has consequently led to the evolution of I2As (imagination-augmented agents), which are designed with a neural network to extract information that might be useful for future decisions. They can adapt a number of imagined possibilities for a particular task and learn different strategies to conduct plans.
To test this out, Google’s Deepmind lets the agents loose on a puzzle game called Sokoban and a spaceship navigation game. For both it’s necessary to forward planning, with levels procedurally generated so that the agents could not simply use trial-and-error. In fact, they could only try each level once and simultaneously.
“For both autonomous tasks, the imagination-augmented agents outperform the imagination-less baselines considerably,” said Deepmind. “They learn with less experience and are able to deal with the imperfections in modelling the environment.”
How do we understand that time travel is possible?
More than 100 years ago, the renowned scientist Albert Einstein came up with an idea about how time works. He called it relativity. This theory says that time and space are linked together. Einstein also said our universe has a speed limit: nothing can travel faster than the speed of light (186,000 miles per second).The theory of relativity usually encompasses two interrelated theories by Albert Einstein: special relativity and general relativity, proposed and published in 1905 and 1915, respectively.
Again, Deepmind said the next step would be to build up the idea to other problems, and design agents that can use imaginations to plan for the future in a diversified scenarios.
As for the algorithms: too many kinds exist. from simple ‘same as today’ models to complex deep learning algorithms, graph analysis and/or probabilistic reasoning models.
The future world is a very broad domain. It would cover economics, geopolitics, technology, sociology, and even philosophy. The future of, say, India will be different from the future of other countries.
All AI systems developed today are domain-specific. They are designed for a specific task to solve specific problems. We don’t (yet) have polymath AI systems that are experts in all fields.
A Different Aspect of AI (Prediction Constraints) -
If we did, have to take into account that the system would make the wrong prediction. The data it has access to could be incomplete, biased or simply wrong. Predicting the future also requires intuition and consciousness. AI systems can simulate intuition but can not have real-time intuition capabilities.
Take the table below. It is a prediction of the future of gender from the American futurist Lawrence (Larry) Taub. No AI system could have made such vital prediction on SEX DIALECTIC because it is based on both subjective personal experience and intuition of Human physiology.
Meanwhile , it’s expected from AI to make discoveries in astronomy (remember that machine learning feeds on data while astronomy produces more data than humans can comprehend). Particle physics is another area where the pattern recognition capabilities of AIs can bring discoveries to light.
To the extent that patterns can be generalized, AIs can be used to see into the future. AIs are used to accurately predict demand for electrical energy. In gaming, AIs can predict an opponent’s next moves. AI is used to predict stock markets ,etc.
Why we need Prediction-Technology ?
Google’s Predictive AI Group announced a big breakthrough at TGIF,” It should now be possible to push that probability far, far into the future — like, decades, even assuming no additional advancements . Because we needed to. We wanted to. We are human. We can do anything. We decided to do this. That is all.”
Artificial Intelligence , Artifical Neural Network , Machine Learning Digest, Time to Change , Time Machine , Time Machine Plus #ML #AI #PredictiveModels #ANNs
