[Posted on Behalf of Steve King, Director, Cybersecurity Advisory Services at Information Security Media Group (ISMG) ]
"AI Needs to Understand How the World Actually Works"
On Wednesday, February 26th, Clearview AI, a startup that compiles billions of photos for facial recognition technology, said it lost its entire client list to hackers.
The company then quickly stated that it has patched the unspecified flaw that allowed the breach to happen, as if that made everything fine again.
In a statement, Clearview AI's attorney said that while security is the company's top priority, "unfortunately, data breaches are a part of life. Our servers were never accessed."
A part of life indeed.
Is AI ready for prime time yet? There were 32 companies exhibiting at RSAC last week, among them 12 startups or early stage organizations that had received over $100 million each in funding.
One of the industry’s leading AI and cryptography thought leaders, Adi Shamir shared his thoughts last week at RSAC on the two biggest problems with deep learning and neural networks.
“We don’t understand why they're working so well,” Shamir said. “And second, we don’t understand why they’re working so terribly.”
This sounds frighteningly similar to that old John Wanamaker quote about an arguably lesser scientific pursuit, “Half the money I spend on advertising is wasted; the trouble is I don't know which half.”
Shamir says technologists are having a difficult time explaining how neural networks learn to discover patterns in lots of data. These neural networks are so big, it can be like having a million interconnected calculators. So trying to find the few that made the most difference in reaching the conclusion can be a challenge.
Shamir noted how some A.I.-powered image recognition systems can confuse photos that look normal, but have been subtly altered by other A.I. systems, as in the case when Google’s technology mistook an image of a turtle with a rifle. Or how current street populations in China are confusing all of the People’s Party biometric scanners by wearing anti-corona-virus face-masks.
Let’s find out what a true expert in AI and deep learning thinks about where we are today on the cycle of innovation.
In December of 2019, the IEEE Spectrum folks interviewed Turing award-winner Yoshua Bengio, the widely revered AI architect and innovator on a range of topics around the shortcomings of Ai, machine and deep learning and neural networks in general. Here’s a couple of interesting excerpts:
Spectrum: How do you assess the current state of deep learning?
Bengio: In terms of how much progress we’ve made in this work over the last two decades, I don’t think we’re anywhere close today to the level of intelligence of a two-year-old child.
But maybe we have algorithms that are equivalent to lower animals for perception. And we’re gradually climbing this ladder in terms of tools that allow an entity to explore its environment.
One of the big debates these days is: What are the elements of higher-level cognition? Causality is one element of it, and there’s also reasoning and planning, imagination, and credit assignment (“what should I have done?”). In classical AI, they tried to obtain these things with logic and symbols. Some people say we can do it with classic AI, maybe with improvements.
Then there are people like me, who think that we should take the tools we’ve built in last few years and create these functionalities in a way that’s similar to the way humans do reasoning, which is actually quite different from the way a purely logical system based on search does it.
Spectrum: How can we create functions similar to human reasoning?
Bengio: Attention mechanisms allow us to learn how to focus our computation on a few elements, a set of computations. Humans do that—it’s a particularly important part of conscious processing. When you’re conscious of something, you’re focusing on a few elements, maybe a certain thought, then you move on to another thought.
This is very different from standard neural networks, which are instead parallel processing on a big scale. We’ve had big breakthroughs on computer vision, translation, and memory thanks to these attention mechanisms, but I believe it’s just the beginning of a different style of brain-inspired computation.
Spectrum: What other aspects of human intelligence would you like to replicate in AI?
Bengio: We talk about the ability of neural nets to imagine: Reasoning, memory, and imagination are three aspects of the same thing going on in your mind. You project yourself into the past or the future, and when you move along these projections, you’re doing reasoning.
If you anticipate something bad happening in the future, you change course—that’s how you do planning. And you’re using memory too, because you go back to things you know in order to make judgments. You select things from the present and things from the past that are relevant.
Attention is the crucial building block here. Let’s say I’m translating a book into another language. For every word, I have to carefully look at a very small part of the book. Attention allows you to abstract out a lot of irrelevant details and focus only on what matters. Being able to pick out the relevant elements—that’s what attention does.
Spectrum: How is your recent work on causality related to these ideas?
Bengio: The kind of high-level concepts that you reason with tend to be variables that are cause and/or effect. You don’t reason based on pixels. You reason based on concepts like door or knob or open or closed. Causality is very important for the next steps of progress in machine learning.
It’s also related to another topic that is much on the minds of people in deep learning. Systematic generalization is the ability humans have to generalize the concepts we know, so they can be combined in new ways that are unlike anything else we’ve seen.
"Today’s machine learning doesn’t know how to do that."
So you often have problems relating to training on a particular data set. Say you train in one country, and then deploy in another country. You need generalization and transfer learning. How do you train a neural net so that if you transfer it into a new environment, it continues to work well or adapts quickly?
Now, imagine the challenge in a cybersecurity detection and deception environment.
Spectrum: Will any of these ideas be used in the real world anytime soon?
Bengio: No. This is all very basic research using toy problems. That’s fine, that’s where we’re at. We can debug these ideas, move on to new hypotheses.
This is not ready for industry by tomorrow morning.
But there are two practical limitations that industry cares about, and that this research may help. One is building systems that are more robust to changes in the environment. Two: How do we build natural language processing systems, dialogue systems, and virtual assistants?
The problem with the current state of the art systems that use deep learning is that they’re trained on huge quantities of data, but they don’t really understand well what they’re talking about. People like Gary Marcus pick up on this and say, “That’s proof that deep learning doesn’t work.” People like me say, “That’s interesting, let’s tackle the challenge.”
Spectrum: How could chatbots do better?
Bengio: There’s an idea called grounded language learning which is attracting new attention recently. The idea is, an AI system should not learn only from text. It should learn at the same time how the world works, and how to describe the world with language.
Ask yourself: Could a child understand the world if they were only interacting with the world via text? I suspect they would have a hard time.
This has to do with conscious versus unconscious knowledge, the things we know but can’t name. A good example of that is intuitive physics. A two-year-old understands intuitive physics. They don’t know Newton’s equations, but they understand concepts like gravity in a concrete sense. Some people are now trying to build systems that interact with their environment and discover the basic laws of physics.
Spectrum: Why would a basic grasp of physics help with conversation?
Bengio: The issue with language is that often the system doesn’t really understand the complexity of what the words are referring to.
For example, the statements used in the Winograd schema; in order to make sense of them, you have to capture physical knowledge. There are sentences like: “Jim wanted to put the lamp into his luggage, but it was too large.” You know that if this object is too large for putting in the luggage, it must be the “it,” the subject of the second phrase. You can communicate that kind of knowledge in words, but it’s not the kind of thing we go around saying: “The typical size of a piece of luggage is x by x.”
We need language understanding systems that also understand the world. Currently, AI researchers are looking for shortcuts. But they won’t be enough.
And by way of background, The Winograd Schema underlies a test that requires the resolution of an expression whose interpretation depends upon another expression in context (its antecedent or postcedent) and is a classic example of natural language processing tasks which depend on commonsense reasoning.
In short, we ain't there yet.
Comments