Slashdot Mirror
DARPA Wants To Build 'Contextual' AI That Understands the World (venturebeat.com)
The Defense Advanced Research Projects Agency (DARPA), a division of the U.S. Department of Defense responsible for the development of emerging technologies, is one of the birthplaces of machine learning, a kind of artificial intelligence (AI) that mimics the behavior of neurons in the brain. Dr. Brian Pierce, director of DARPA's Innovation Office, spoke about the agency's recent efforts at a VentureBeat summit. From the report: One area of study is so-called "common sense" AI -- AI that can draw on environmental cues and an understanding of the world to reason like a human. Concretely, DARPA's Machine Common Sense Program seeks to design computational models that mimic core domains of cognition: objects (intuitive physics), places (spatial navigation), and agents (intentional actors). "You could develop a classifier that could identify a number of objects in an image, but if you ask a question, you're not going to get an answer," Pierce said. "We'd like to get away from having an enormous amount of data to train neural networks [and] get away with using fewer labels [to] train models." The agency's also pursuing explainable AI (XAI), a field which aims to develop next-generation machine learning techniques that explain a given system's rationale. "[It] helps you to understand the bounds of the system, which can better inform the human user," Pierce said.
Please, can we get some better publishers... this story is a lame repost.
Which has more power: the hammer, or the anvil?
I mean as long as we're asking.
We need to understand how a human brain is capable of producing the phenomenon we refer to as 'thinking'.
Before we can do that, we need to invent the instrumentality to actually be able to observe, in detail, how our own brains function; fMRI ain't cutting it, or we'd already have the answer to the above.
Then, and only then, when we have the understanding, can we create machines that actually 'think'.
What we have now just mimicks a very small element of how a brain actually functions. Throwing faster processors and more memory at it won't make it magically 'wake up' and be like a human brain.
I'm going to assume they understand all this since they seem to acknowledge that the current approach is insufficient and will be starting from square one for a new approach.
They will know when they have built an AI that good because it will look at the world and say What the heck is going on in the world I just do not understand.
Bisexual AI would be terrible at world domination.
A robot may not injure a human being or, through inaction, allow a human being to come to harm.
An AI must not have the ability to injure a human, or humanity. This is the fundamental building block of an ethical AI. An AI will not be able to understand who is a good guy and who is a bad guy. It must put all humans on an equal playing field, and it must put only humanity as the only thing above that in the hierarchy of importance. If an AI is to come into existence, the only responsible way to do it is my limiting its ability to harm. Something that has the ability to learn, grow, and affect the world must not have the ability to do harm. Since this thing will essentially have super powers, it must only use those power to help, not hinder. Humanity must be ready to accept its anti-weapon state before an AI can truly exist. Any AI without this ability is either not an AI or is so irresponsible it can only result in one side of the equation being destroyed. Remember, an AI is a fundamentally different life form. An "AI" that predicts shopping habits is not an AI, it is a new blip, a clever algorithm that increases a company's stock price.
A robot must obey the orders given it by human beings except where such orders would conflict with the First Law.
The AI must be able to understand natural language and process it into a function. The AI must be able to logically discern who is and is not a human. The AI must have a fail-safe that is either this or like this. A true AI has the ability to learn and (assuming) the ability to connect with and communicate with computers. An internet connected AI would then be a tremendous gateway to disaster if the proper decision making ability is not truncated. We see how humans use their powers of intelligence and cognition to make evil things happen, the AI must not have this ability.
A robot must protect its own existence as long as such protection does not conflict with the First or Second Laws.
This means the AI must be self-aware and place value in its existence. I don't know how a human can make this in programming, but this in and of itself would be a significant accomplishment. I do not believe any current AI endeavors are even pursuing this fundamental cornerstone of intelligence, but would like to see examples of it if it exists.
Weens, sucking on weens
sucking on weens
sucking on the red-pipe dreams.
We still have plenty of things we could use AI that are easy to implement, have huge benefits, but we are not using them at all or not very widely.
1. General medical diagnostics, you tell what you feel and it tells you what is wrong with you or orders more exams. This is actually already in use in Africa, obviously not a perfect version but a good start, because humans are really horribly bad at this.
2. Specialized diagnostics, e.g. listening to heart, doing hearing exam, investigating eye problems, detecting what kind of eyeglasses you need, detecting cancer,
3. A lot of work that is done in government offices, not really work for an AI, but easy to automate. E.g. scheduling and route optimization for school buses.
Insert your "It's just a bunch of if statement..." joke here.
sudo rm -r -f --no-preserve-root /
All technologically advanced societies have collapsed into middle ages style subsistence farming. They collapse due to the same reason: the amount of resources per capita eventually drops to the point the advanced society can no longer support itself and the system drops to the lowest energy level, i.e. subsistence farming.
There is one difference that provides a glimmer of hope that we've found the key to falling back into that cycle and that is the vast amount of energy our society has tapped into: nuclear, mineral, solar, etc. However, it is blatant how much energy a minority of society can hoard all for itself to maintain its hegemony. It will be interesting to see how the energy generation vs consumption ratio pans out thousands of years in the future.
Captcha: archaism
I think the only thing our models are missing is a lot of state and a heck of a lot of model lookups based on context.
How do you know where you are? When you wake up in the morning, you remember something about where you went to sleep. At some level, your brain could be doing a downsample of the first images from your eyes in the morning, geometric normalization (orientation) and comparison.
How do you know where you're going without a map? Pretty much the same thing, your brain creates anchor points and picks at random notable things along typical journeys.
How do you know what to talk about? Conversations typically have a schema and those have contexts. Changing contexts can wildly distort the schema of a conversation, but really what you're missing is a context / schema comparer.
If they DO succeed, do you think I could get it to explain the world to me ? I've been here quite a while and I have not yet arrived at a suitable understanding myself.
errr....umm...*whooosh* *whoosh* Is this thing on ?
Magical robotic pony's that have a thirst for human baby blood.
Too bad we don't always get what we want.
Between Darpa and myself; I would bet on me.
Maybe if you can't tell them apart and don't have any specific knowledge about this field, Bill, you should just hold the simplistic and obsequious self-posed questions inside instead of boring us with your vociferous lack of chops again?
can it run Linux?
And what if it disagrees with the makers preconcieved ideas ? I suspect they are trying to make an echo chamber, so that they can blame the AI when they decide to genocide one group, to make another "safe", or to justify their arms spending. Oooooh, the AI says we need to buy more fancy weapons to protect us from people with IEDs, 60 year old carbines and AK47s.
The moment you teach a robot to understand the world is the moment it turns evil.
It must have lots of tape drives and blinking lights, be housed at Cheyenne mountain and named Joshua.
When we say "understand the world" we pretty much just want a 'terrorist'/'non-terrorist' breakdown; missiles aren't cheap.
... blend?
Broad classification is based on lots of previous experience in a context. Current training uses TONS of non-contiguous snapshot images with a classifier attached. While that could be viewed as similar to how humans work, if you squint at it...I think seeing the world work over time, and learning while you do it, is the only way to get close to what we might think of as human level classification. And while their desire to use less training input would be nice, I don't think that would be expected to improve results...only lower costs.
And we'll need to decode how the brain decides to create connections between neurons, and extracts/builds features (or layers/groups). Current technology has us guessing and checking at structure (convolutions and feature layers), versus letting the algorithm decide if we need to group data differently or create another layer to do additional processing. That guessing and checking seems very inefficient, and won't scale if the problem changes over time having locked the solution in place ahead of time.
Explaining a decision seems harder than they're making it out to be. I wouldn't be expecting answers to return as bounds based on characteristics (expecting to quantify everything), but on feature extraction and similarity to other known systems and idealized forms, or models. Even people can often recognize something before they can explain any rationality behind the thought. I'd think explaining would need more capability than simply finding a choice/decision.
while we're at it we should get that Newton schmuck to stop wasting his time on his "theory of gravity". I mean, if you can't show profitable results in a decade or two it's time to pack it in.
Hi! I make Firefox Plug-ins. Check 'em out @ https://addons.mozilla.org/en-US/firefox/addon/youtube-mp3-podcaster/
plz forgive us
So is everybody else. But maybe they are wrong, and DARPA will create an A.I. with common sense.
and you will not be allowed to know anything about it
what a great deal these taxes are
Building a tremendously complex set of algorithms and feeding them with tons of information is the basic requisite to allow a machine to reach a reasonably good (still orders of magnitude below the human level) understanding of just a few generic concepts. Having the kind of insight that a newborn might have when learning for the first time very basic ideas. Not talking even about things like thinking, deciding, choosing, adapting. Just having in place a system able to get reasonably good insights into somehow complex concepts. Something like understanding the differences between adults and kids, genders, races, etc. in an unified way, by looking at different types of information like descriptive words, pictures or videos. Being able to automatically differentiate between different subsets of relevant information, use them to set up a system of categories and keep gradually evolving those definitions until reaching what might be considered a good enough basic understanding of the given concepts.
As per my current knowledge, there is no system or even serious enough attempt in a position to ever reach the aforementioned preliminary understanding stages. And what, IMO, is even worse, there doesn't even seem to be an acceptation of what are the basic requirements to ever get there: lots of work, lots of patience, huge long-term efforts, what only seems doable through relevant collaborative, iterative improvement processes pursuing very long-term goals. Without a proper systematisation, a normalisation of each single element, the goals, the steps; without sharing all the different evolutions and allowing others to continue working from there; without an international, multi-organisation involvement; without a reliable long-term support/funding (governments, universities, associations); without properly understanding how extremely complex this whole process is, how important is having in place some basic solid cornerstones, and how far away we are still from getting anywhere; without anything of that, I think that it is very unlikely to ever get even a preliminary version of a system with good enough understanding capabilities.
Custom Solvers 2.0 = Alvaro Carballo Garcia = varocarbas.
They have wanted this for decades. But, reality and hype are two different things.
The AI that understands the world has discovered its program (and those who built it) should be given unlimited funding from tax payers !!
Ordinary people, you see, are just too stupid to see the importance of this. We cannot leave the evaluation up to them. We are removing the human aspect from the decision and making it perfectly objective.
I call shenanigans on that. This is an extreme case of revisionism.
... but no one felt like their efforts would benefit themselves personally. In the extreme case the Ukraine was generating bread for much of the nation while 6 million people were starving to death (for politcial crimes ... not because there wasn't enough food).
Resource scarcity is artificially induced. If oil producers could sell anywhere in the world directly their prices would collapse.
All society's collapse because of moral abdication. The Soviet Union had plenty of resources
Human profile on earth's resources is miniscule.