* Posts by Rebel Science

45 publicly visible posts • joined 9 Oct 2017

US Declaration of Independence labeled hate speech by Facebook bots

Rebel Science

Book burning Nazis

What Mark Zuckerberg and the other evil masters of social networks are doing is no different than what the Nazis did. It's just book burning by another name.

'Autopilot' Tesla crashed into our parked patrol car, say SoCal cops

Rebel Science

The fundamental problem of self-driving cars is deep learning

Deep learning sucks. Unlike the brain, a deep neural net can only see things it has been trained to detect. IOW, don't wear a Chewbacca costume in front of an autonomous car. Just saying.

Mobileye's autonomous cars are heading to California. But they're not going to kill anyone. At least not on purpose

Rebel Science

Level 5 autonomy is coming but no thanks to Deep Learning

Unlike the brain, a deep neural net cannot see an object it has never seen before. There will be no fully autonomous cars until we solve the AGI problem. We need breakthroughs in instant object detection, a common sense or causal understanding of the world, prediction and planning. Once you have that, you won't even need fancy sensors like Lidar, radar, infrared cameras and such. A simple movable binocular camera will do. Above all, you won't need to test the system for millions of miles. The machine will learn to drive just like humans do.

Autonomous vehicle claims are just a load of hot air… and here's why

Rebel Science

Deep Learning is the problem

Don't ride in a self-driving car that is powered by deep learning. A deep neural network is essentially a rule-based expert system. They all have a fundamental flaw: they fail catastrophically if they encounter a situation that they have never seen before. For example, it may not recognize a pedestrian in a Chewbacca suit because it has never been trained to recognize one. The human brain does not have to be trained for every possible situation, an impossible task. We have common sense. This is why we are still orders of magnitude safer drivers than autonomous cars. The California DMV disengagement data on self-driving cars proves this.

Transportation agencies everywhere should ban all self-driving cars on public streets until they can prove they have common sense.

Fatal driverless crash: Radar-maker says Uber disabled safety systems

Rebel Science

We must ban all self-driving cars on public streets now

In my opinion, the @USDOT should immediately impose a moratorium on all autonomous vehicles on public roads. Deep Learning is not suitable for uncontrolled or open environments where humans can be harmed. A deep neural net is like a rule-based expert system: it will fail catastrophically if it encounters a situation it has not been trained on. The blame for the next fatality will rest on the shoulders of @SecElaineChao.

The UK and other European nations should do likewise because more fatal accidents are coming. Guaranteed.

Uber breaks self-driving car record: First robo-ride to kill a pedestrian

Rebel Science

Is this the first deep learning homicide?

Was the Uber taxi controlled by a deep learning system at the time?

Fun fact of the day: Voice recognition tech is naturally sexist

Rebel Science

Sexist you say?

It's time to ban them on Twitter, FaceBook, Instagram and YouTube for hate speech. Also, detain them at the London Heathrow airport for 3 days and confiscate their passports. But do let the jihadis back in, by all means. LOL

Bots don't spread fake news on Twitter, people do, say MIT eggheads

Rebel Science

Twitter spreads fake news on Twitter.

And blocks news that are right wing. Twitter is a huge propaganda platform for the far left. We all know this is true. Why deny it?

Google assisting the Pentagon in developing AI for its drones

Rebel Science

Re: Google and the military

They've been with the military, DARPA, CIA, etc. from day one. DeepMind was just pretending to be ethical and conscientious just to attract top talent. The mainstream AI industry is not to be trusted.

Super Cali's futuristic robo-cars in focus. Even though a watchdog says tech is quite atrocious

Rebel Science

Don't trust any car that uses Deep Learning

Deep neural nets are notorious for failing catastrophically when they encounter a situation they've never seen before. You've been warned.

Reinforcement learning woes, robot doggos, Amazon's homegrown AI chips, and more

Rebel Science

Only DeepMind's Demis Hassabis believes that deep reinforcement learning is the future of AGI

It's embarrassing, to say the least.

Disengage, disengage! Cali DMV reports show how often human drivers override robot cars

Rebel Science

Re: One intervention needed per 2km....

You're kidding? We would all be dead if that were true. Without intervention, the self-driving cars would suffer catastrophic accidents.

Rebel Science

Fully Autonomous Vehicles Will Be Science Fiction for the Foreseeable Future

Fully autonomous vehicles are way beyond what current AI technologies can handle. A major breakthrough in AGI must happen before we realize this dream. One thing is certain: it will not happen with Deep Learning. A deep neural net is really an expert system and, as such, it suffers from the same fatal flaw: it fails catastrophically every time it encounters a situation for which it has not been trained. This is unsuitable for real world applications where safety is a must.

To all big time investors: Do not waste money investing in any project using deep learning to achieve full driving autonomy. It's a waste of time and money. Invest in AGI research instead.

Aut-doh!-pilot: Driver jams 65mph Tesla Model S under fire truck, walks away from crash

Rebel Science

Re: The Nasty Little Truth About Deep Learning

Interesting as your comments are, they've got fuck all to do with a Tesla, which does not claim to be, and is not, an autonomous vehicle.

You realize that the Tesla vehicle was on AutoPilot and that 'Auto' means autonomous, don't you?

Rebel Science

Re: The Nasty Little Truth About Deep Learning

I see the DeepMind employees and fans are out in force. Makes no difference. DeepMInd is just a deep learning outfit. Their leader, Hassabis, does not even know that the brain learns without backprop. It's embarrassing.

Rebel Science

The Nasty Little Truth About Deep Learning

The nasty truth is that deep learning is prone to catastrophic failures. It's the same flaw that pretty much doomed expert systems. In fact, despite denials, DNNs ARE expert systems. Unlike human drivers, a deep learning system, the kind used in self-driving cars, cannot see something it has not been pre-trained to recognize. It's a monumental flaw that guarantees that catastrophic failures are unavoidable.

But the deep learning community will continue to hype this technology to death. The AI winter cometh.

Skynet it ain't: Deep learning will not evolve into true AI, says boffin

Rebel Science

Deep learning is not and has never been intelligent

Deep learning systems have no idea what they're seeing. We must give them a label. Even then, they still have no clue. An adult human brain can instantly see and interact with a completely new object or pattern that it has never seen before. And it can do it from different perspectives. A DNN, by contrast, must be given hundreds if not thousands of samples of the object in order to properly detect it in an invariant manner. And you still have to give it a label.

Rebel Science

Deep learning must be discarded like yesterday's garbage in order to solve AGI

A number of us have been saying this for many years. But the AI community, like all scientific fields, is extremely political. Only the famous leaders have influence, even if they are clueless.

Why Deep Learning Is a Hindrance to Progress Toward True AI

Checkmate: DeepMind's AlphaZero AI clobbered rival chess app on non-level playing, er, board

Rebel Science

DeepMind is clueless about how to achieve AGI

DeepMind has never made a breakthrough in AI and never will. They essentially apply well-known techniques invented by others (Monte Carlo search, deep learning and reinforcement learning) to games chosen for their limited number of behavioral options. I would be infinitely more impressed if they made a robot that could walk in any generic kitchen and fix a meal of scrambled eggs with bacon, toast and coffee.

As an aside, Demis Hassabis and his team at DeepMind are on the record for suggesting that the human brain uses backpropagation for learning. They published a peer-reviewed paper on it. I cringe when I think about it.

Elon Musk finally admits Tesla is building its own custom AI chips

Rebel Science

Re: Musk is crazy

Maybe he knows something that we don't.

OpenAI uses cunning code to speed up GPU machine learning

Rebel Science

Re: Deep Learning Is a Hindrance to Progress Toward True AI

Deep learning is certainly useful and will always be useful but not for figuring out AGI.

Rebel Science

Deep Learning Is a Hindrance to Progress Toward True AI

So OpenAI wants to develop AGI but they are still spending talent and busloads of cash on DNNs, connection weights and backpropagation? It's a sure bet AGI will not come from that bunch. Deep learning experts are the least qualified people to work on AGI. Almost everything they know and think is important is wrong. Just saying.

Level 5 driverless cars by 2021 can be done, say Brit industry folk

Rebel Science

Lab rats, the lot of you

Your government is using you as lab rats. They want to find out how many fatal traffic accidents by autonomous cars you are willing to tolerate. Good luck with that.

Tesla launches electric truck it guarantees won't break for a million miles

Rebel Science


I hope to live long enough to see the end of all gasoline powered vehicles.

Stop worrying and let the machines take our jobs – report

Rebel Science

Re: Give us what is ours

US markets were never free. It was never based on inheritance. Only an inheritance-based, free market system whereby the land and its wealth belongs to all is viable. In such a system, all corporations try to make a profit for the people who invest their share of the inheritance (the wealth of the land) into the system. Everybody should receive profits from the corporations in addition to their wages if they work. Socialism and communism (government controlled programs) will never work because there is no incentive for profits and hard work. They also destroy the free market, the only way that goods and services can be properly valued.

In an inheritance-based system, artificial intelligence can simply be used by corporations to make more profits for the people. There would be no disruption because nobody would be depending solely on their work for a living.

We'd better wake up, people, before AI eliminates all jobs and the plutocracy turns the world into a welfare society forced to survive on handouts from a thieving minority.

Rebel Science

Give us what is ours

You're a century or two too late, sunshine. Marx put this idea forward,

Man, give me a break. Marx was a friggin' moron, a mediocre mind. I don't believe in socialism. Socialism/communism is about government programs: free healthcare, free housing, free education, etc. I don't believe in any of that crap. I believe in a purely free market system. Just give us what is ours by right.

Rebel Science

The Centre for Policy Studies is, of course, a Big Brother, plutocrat-funded organization whose job is to BS the masses. In any just society, the people would be delighted to have robots do all their work for them. The fact that we are afraid that automation will take our jobs leaving us without a way to make a living should be a wake up call to the fact that we are slaves in a slave system.

True capitalism is where the people own the corporations because the corporation are funded with capital which represents the wealth of the earth. The wealth of the earth belongs to all. We are being ripped off. Give us what belongs to us.

The Biggest Crime of the Plutocracy Is That They Stole the Capital From the People

Look, ma! No hands! Waymo to test true self-driving cars in US with Uber-style hailing app

Rebel Science

Re: What's the point of "truley human-free" cars?

In the future, most vehicles on the road will be transporting cargo.

Rebel Science

Autonomous cars are based on deep learning, an old, annoyingly inadequate, baby-boomer technology from the last century. Just one little unsupervised learning breakthrough can force deep learning to become obsolete overnight. Suddenly, a bunch of highly paid AI experts are about as valuable as horse buggy mechanics. Not funny, I know. But life can come at you faster than you think.

How we fooled Google's AI into thinking a 3D-printed turtle was a gun: MIT bods talk to El Reg

Rebel Science

This just shows that deep learning is not the answer to AGI. In fact, it's a hindrance. Wake up, people. AGI research money must be applied elsewhere.

Subsidy-guzzling Tesla's Model 3 volumes a huge problem – Wall St man

Rebel Science

No welfare for billionaires. Please.

Tesla share crash amid Republican bid to kill off electric car tax break

Rebel Science

Musk has been living on government welfare all along?

Can you get from 'dog' to 'car' with one pixel? Japanese AI boffins can

Rebel Science

Geoffrey Hinton Is Right. Backpropagation Must Go

If your goal is AGI and you're using backpropagation, you're hopelessly lost. If you're not using spike timing, you're doing it wrong. What should replace backpropagation?

Google: We don't have a quantum computer yet, but we have a compiler

Rebel Science

Re: Is this sarcasm or a troll?

"the qbits will give you probability spreads"

It's just a super expensive and very unreliable analog computer. Instead of solid analog values (as in a slide rule), it's using probabilities. It's crap.

Rebel Science

LOL. Quantum computers are right up there with the phlogiston and dark matter.

1. Opposite states are superposed but only if you're not looking.

2. We can't see them but we know they are there.

3. Trust us. We know what we're talking about.

Right, sure. The crackpottery in Big Science is friggin' hilarious.


ARM chip OG Steve Furber: Turing missed the mark on human intelligence

Rebel Science

The timing of the spikes is what's important

"Depending on its role in the brain, that timing may or may not be significant. It's clear you can't completely ignore it."

Are you kidding? You guys need to completely forget about spiking rate. Timing is everything. EVERYTHING. Spiking rate is a red herring, a complete waste of time (no pun intended). It is true that the retina uses rank order encoding to compress visual information (~200 to 1) but the cortex is entirely driven by the precise timing of the spikes.

Fast Unsupervised Pattern Learning Using Spike Timing

Rebel Science

Efficient, unsupervised spiking neural nets are the future of AGI

Great article.

"And so we built a hardware-software system that has good support for sparse connectivity. We're very focused on spiking networks whereas machine learning almost completely ignores spikes."

Wonderful. Now that deep learning guru Geoffrey Hinton has finally acknowledged that we must abandon backpropagation and start over, it is time to promote the correct paradigm that will replace backpropagation. Deep neural nets will soon become obsolete. The future of machine learning will be based on the precise timing of discrete sensory signals, aka spikes. Welcome to the new age of unsupervised spiking neural networks.

Unsupervised Machine Learning: What Will Replace Backpropagation?

How DeepMind's AlphaGo Zero learned all by itself to trash world champ AI AlphaGo

Rebel Science

Re: FFS, it isn't AI ...

@Simon Ward

You're absolutely correct. In spite of all the hype, denials and posturing, they're still doing GOFAI, the baby boomer AI of the last century. They just got faster machines and more memory to play with.

Rebel Science

Very impressive indeed. But I would be orders of magnitude more impressed if it could also walk to the kitchen and make me a cup of coffee.

El Reg was invited to the House of Lords to burst the AI-pocalypse bubble

Rebel Science

Re: nice

I agree. Unfortunately, the Deep Learning people don't want to hear anything that contradicts their paradigm.

Ghost in Musk's machines: Software bugs' autonomous joy ride

Rebel Science

There are two problems: software unreliability and the brittleness of deep neural nets

Software unreliability is proportional to complexity and is a direct result of our current computing paradigm which is based on the algorithm. The solution is to stop using the algorithm as the basis of programming and adopt a signal-based, reactive programming model. Essentially, software should work more like electronic circuits.

The second problem is that, in spite of the loud denials from the AI community, their biggest success, deep learning, is just GOFAI redux. A deep neural network is actually a rule-based expert system. AI programmers just found a way (gradient descent, fast computers and lots of labeled or pre-categorized data) to create the rules automatically. The rules are in the form, if A then B, where A is a pattern and B a label or symbol representing a category.

The problem with expert systems is that they are brittle. Presented with a situation for which there is no rule, they fail catastrophically. Adversarial patterns prove this in neural nets and Tesla Motors found out about it the hard way. The car's neural network failed to recognize a situation and caused a fatal accident. This is not to say that deep neural nets are bad per se. They are excellent in controlled environments, such as the factory floor, where all possible conditions are known in advance and humans are kept at a safe distance. But letting them loose in the real world is asking for trouble. Obviously, we will need a better solution.

Here are a few relevant links for those who care:

Why Software Is Bad and What We can Do to Fix It

The World Is its Own Model or Why Hubert Dreyfus Is Still Right About AI

In Spite of the Successes, Mainstream AI is Still Stuck in a Rut

Why Deep Learning Is A Hindrance to Progress Toward True Intelligence

Calm down, Elon. Deep learning won't make AI generally intelligent

Rebel Science

Re: Bishop Bollocks

The little brain-dead materialist lies like a rug.

Rebel Science

Re: Bishop Bollocks

Another brain-dead superstitious materialist heard from. Here's what your little materialist cult believes in: the universe created itself by some unknown magic; machines are or will be conscious by some unexplainable magic called emergence; lifeforms created themselves from dirt. I could go on but then I would barf out my lunch.

Rebel Science

Deep Learning is not even in the same ballpark as AGI

I'm glad to see the deep learning hype is finally subsiding. I and many others have been saying this for years. The success of deep learning has been a disaster to AGI research. Geoffrey Hinton, one of its leading pioneers, has finally admitted that they need to scrap backpropagation and start over.

AI Pioneer Now Says We Need to Start Over. Some of Us Have Been Saying This for Years