Skip to main content

This basic human skill is the next major milestone for A.I.

Remember the amazing, revelatory feeling when you first discovered the existence of cause and effect? That’s a trick question. Kids start learning the principle of causality from as early as eight months old, helping them to make rudimentary inferences about the world around them. But most of us don’t remember much before the age of around three or four, so the important lesson of “why” is something we simply take for granted.

It’s not only a crucial lesson for humans to learn, but also one that today’s artificial intelligence systems are pretty darn bad at. While modern A.I. is capable of beating human players at Go and driving cars on busy streets, this is not necessarily comparable with the kind of intelligence humans might use to master these abilities. That’s because humans — even small infants — possess the ability to generalize by applying knowledge from one domain to another. For A.I. to live up to its potential, this is something it also needs to be able to do.

“For instance, if the robot learned how to build a tower using some blocks, it may want to transfer these skills to building a bridge or even a house-like structure,” Ossama Ahmed, a master’s student at ETH Zurich in Switzerland, told Digital Trends. “One way to achieve this might be learning the causal relationships between the different environment variables. Or imagine that the TriFinger robot used in CausalWorld suddenly loses one finger due to a hardware malfunction. How can it still build the goal shape with only two fingers instead?”

CausalWorld video

A virtual training world for machines

CausalWorld is what Frederik Träuble, a Ph.D. student at the Max Planck Institute for Intelligent Systems in Germany, refers to as a “manipulation benchmark.” It’s a step toward advancing research so that robotic agents can better generalize various changes in an environment’s properties, such as the mass or shape of objects. For example, if a robot learns to pick up a particular object, we might reasonably expect that it can transfer this ability to heavier objects — so long as it understands the right causal relationship.

The kind of virtual training environment we’re used to hearing about in sci-fi movies is the one in, say, The Matrix: a virtual world in which rules don’t apply. In CausalWorld, in which researchers can systematically train and evaluate their methods in robotic environments, it’s just the opposite. It’s all about learning the rules — and applying them. Robot agents can be given tasks similar to the ones kids participate in when they play with blocks to do stacking, pushing and other cause-and-effect play. The researchers can intervene to test the robot’s generalization abilities as it learns. It’s basically a testing environment that will help evaluate how A.I. agents can generalize.

“Most of modern A.I. is based on statistical learning, which is all about extracting statistical information — for example, correlations — from data,” Bernhard Schölkopf, director of the Max Planck Institute, told Digital Trends. “This is great because it allows us to predict one quantity from others, but only as long as nothing changes. When you intervene in a system, then all bets are off. To make predictions in such cases, we need to go beyond statistical learning, towards causality. Ultimately, if future A.I. is to be about thinking in the sense of ‘acting in imagined spaces,’ then interventions are key, and thus causality needs to be taken into account.”

Editors' Recommendations

Luke Dormehl
I'm a UK-based tech writer covering Cool Tech at Digital Trends. I've also written for Fast Company, Wired, the Guardian…
Nvidia is renting out its AI Superpod platform for $90K a month
nvidia hgx 2 supercomputer server nvidiadgx2

Nvidia is looking to make work and development in artificial intelligence more accessible, giving researchers an easy way to access its DGX supercomputer. The company announced that it will launch a subscription service for its DGX Superpod as an affordable way to gain entry into the world of supercomputers.

The DGX SuperPod is capable of 100 petaflops of AI performance, according to the company, and when configured 20 DGX A100 systems, it's designed for large-scale AI projects.

Read more
Google’s LaMDA is a smart language A.I. for better understanding conversation
LaMDA model

Artificial intelligence has made extraordinary advances when it comes to understanding words and even being able to translate them into other languages. Google has helped pave the way here with amazing tools like Google Translate and, recently, with its development of Transformer machine learning models. But language is tricky -- and there’s still plenty more work to be done to build A.I. that truly understands us.
Language Model for Dialogue Applications
At Tuesday’s Google I/O, the search giant announced a significant advance in this area with a new language model it calls LaMDA. Short for Language Model for Dialogue Applications, it’s a sophisticated A.I. language tool that Google claims is superior when it comes to understanding context in conversation. As Google CEO Sundar Pichai noted, this might be intelligently parsing an exchange like “What’s the weather today?” “It’s starting to feel like summer. I might eat lunch outside.” That makes perfect sense as a human dialogue, but would befuddle many A.I. systems looking for more literal answers.

LaMDA has superior knowledge of learned concepts which it’s able to synthesize from its training data. Pichai noted that responses never follow the same path twice, so conversations feel less scripted and more responsively natural.

Read more
How the USPS uses Nvidia GPUs and A.I. to track missing mail
A United States Postal Service USPS truck driving on a tree-lined street.

The United States Postal Service, or USPS, is relying on artificial intelligence-powered by Nvidia's EGX systems to track more than 100 million pieces of mail a day that goes through its network. The world's busiest postal service system is relying on GPU-accelerated A.I. systems to help solve the challenges of locating lost or missing packages and mail. Essentially, the USPS turned to A.I. to help it locate a "needle in a haystack."

To solve that challenge, USPS engineers created an edge A.I. system of servers that can scan and locate mail. They created algorithms for the system that were trained on 13 Nvidia DGX systems located at USPS data centers. Nvidia's DGX A100 systems, for reference, pack in five petaflops of compute power and cost just under $200,000. It is based on the same Ampere architecture found on Nvidia's consumer GeForce RTX 3000 series GPUs.

Read more