Catch up on stories from the past week (and beyond) at the Slashdot story archive


Forgot your password?
DEAL: For $25 - Add A Second Phone Number To Your Smartphone for life! Use promo code SLASHDOT25. Also, Slashdot's Facebook page has a chat bot now. Message it for stories and more. Check out the new SourceForge HTML5 internet speed test! ×

Comment Re:FP16 isn't even meant for computation (Score 1) 55

So, one problem is that there is not always more data. In my field, we have a surplus of some sorts of data, but other data requires hundreds of thousands of hours of human input, and we only have so much of that to go around. Processing all of that is easy enough, getting more is not.

Also, by "effective", I should have made it clear that I meant "an effective overall solution to the problem", which includes all costs of training a wider, lower-precision network. This includes input data collection, storage and processing, all of the custom software to handle this odd floating point format, including FP16-specific test code and documentation, run time server costs and latency, any increased risks introduced by using code paths in training and , etc.

I'm not saying that I don't believe it's possible, I've just seen absolutely no evidence that this is a significant win in most or even a sizable fraction of cases, or that it represents a "best practice" in the field. Our own experiments have shown a severe degradation in performance when using these nets w/out a complete retraining, the software engineering costs will be nontrivial, and much of the hardware we are forced to run on does not even support this functionality.

As an analog, when we use integer based nets and switch between 16-bit and 8-bit integers, we see an unacceptable level of degradation, even though there is a modest speedup and we can use slightly larger neural nets. I'm very wary of anything with a mantissa much smaller than 16 bits for that reason--those few bits seem to make a significant difference, at least for what we're doing. We're solving a very difficult constrained optimization problem using markov chains in real time, and if the observational features are lower fidelity, the optimization search will run out of time to explore the search space effectively before the result is returned to the rest of the system. It's possible that the sensitivity of our optimization algorithm to input quality is the issue here, not the fundamental usefulness of FP16, but I'm still quite skeptical. If this were a "slam dunk", I'd expect to see it move through the literature in a wave like the Restricted Boltzmann Machine did.

Oh, and thank you for the like (great reading) and the thoughtful reply. Not always easy to find on niche topics online.

Comment Re:Exploitative by design? (Score 1) 153

It seems like these systems are exploitative by design, even if exploitation wasn't explicitly the goal. They're designed with every possible algorithm and available data to maximize labor output at the lowest possible cost. Individual workers are operating at extreme information asymmetry and against a system which does not negotiate and only offers a take it or leave it choice.

This is by far the best comment I've ever seen regarding this sort of algorithmic labor management.

Normally I'm all for this sort of thing--my company is a client and uses it to handle large bursts of data processing quickly--but the information symmetry argument is a powerful one. Also, there doesn't seem to be a lot of competition in this space, which might otherwise ameliorate a lot of the problems induced by the "take it or leave it" bargaining approach.

The analysis provided by the article is absurd, but yours seems to lead to the inescapable conclusion that some kind of regulation is necessary to prevent blatant exploitation. Maybe just reducing information asymmetry in some way, or requiring transparency in reports available to the public on the website regarding effective wages paid to workers as a fraction of the minimum and average wages of employees in their respective countries. Surely someone can find an answer to this.

Comment Re:FP16 isn't even meant for computation (Score 1) 55

Accidentally posted as anonymous coward, reposting under my actual name.

So they're all excited about the lowest-precision, smallest-size floating point math in IEEE 754?

FP16 is good enough for neural nets. Do you really think the output voltage of a biological neurons has 32 bits of precision and range? For any given speed, FP16 allows you to run NNs that are wider and deeper, and/or to use bigger datasets That is way more important than the precision of individual operations.

There's a lot of rounding error with FP16. The neural networks I use are 16-bit integers, which work much, much better, at least for the work I'm doing. Also, do you have a good citation that FP16 neural networks are, overall, more effective than FP32 networks, as you've described?

Comment Re:What I want to know is... (Score 1) 368

You make quite an assumption that I enjoy more liberties. Maybe I don't. Maybe I came from a place where everyone had too much freedom and there was chaos and there is no stability and I look at China as a possible alternative. Maybe not the best alternative, but better than the stale poverty that exists. I see India with the biggest democracy in the world, yet also the biggest bureaucracy and heavily corrupted politicians. Where progress occurs but there is slower rate of progress and constant clashes between groups.

And I think about my own home and think, if I could choose one path for the country, I would prefer China's path versus India's. Maybe you disagree, maybe your experiences and values have shaped you so much that you can't even fathom why another person would view differently. But then, I feel sorry for your lack of ability to understand others.

Comment Re:What I want to know is... (Score 1) 368

Many here would think anyone that supported the Chinese government's positions are brainwashed, yet they can't conceive of themselves being heavily biased or maybe they don't have the intellectual ability to see from other perspectives. My post is not flamebait. What it is, is a view contrary to your own from someone who has different experiences and different values. Maybe your biases have shaped you so much that you can't see the world in all its many forms.

Comment Re:What I want to know is... (Score 0, Troll) 368

It's like being friends with a bully and joining the gang. You go around and help the bully. And then one day, the bully smacks you down in front of others reminding you he's the leader. Only then do you call the bully out for his aggressive nature. It's complete fucking BS and the fact that so many are buying the story is pathetic.

However, having said that, I'll add that I am one of the few that think China is in the right. China still have a lot of poor people but they're developing rapidly. The last thing it needs right now is a bunch of tea-bagger like idiots protesting the government. The people don't need others to cause unrest or impede growth. Contrary to most Americans, most Chinese value stability and economic progress to certain liberties. Obviously, with as big a population as China has, even a small percentage of people that prioritize liberty would still number in the tens of millions. But life isn't fair and it sucks to be them. Hopefully they can find some form of happiness in their lives.


Sun's Project Darkstar Game Server Platform No More 82

sproketboy writes "Project Darkstar, an open source software platform from Sun labs that simplifies the development of horizontally scalable servers for online games, is being discontinued as of the Oracle acquisition. This project, mentioned a couple of years back on Slashdot, was a unique concept for building an application server specific to on-line gaming. Sadly they were so close at version 0.9.11 (which is still very stable). Hopefully the open source community can get involved and help continue work on this project."

Comment Re:Other venues... (Score 1) 128

What the fuck is with this need to bash countries? Especially when it's so far off the mark.

Why is stating fact considered, "bashing"? The FACT is, the stated fact is exactly how car companies advertise to consumers because of exactly that fact. Hell, in the US, car purchases are actually considered an impulse purchase (80%+ of all vehicle sales). Its an impulse purchase because its about both the car and the salesmen stroking the consumer's ego rather than the merits of the vehicle. The really sad thing is, few consumers actually realize they purchased a vehicle because they are a complete idiot rather than the merits of the vehicle. Because you consider the stated fact to be bashing, it strongly suggests you easily fall into the ignorant 80% category.

Perhaps the realization you purchase vehicles out of looks and ignorance rather than merit, function, and suitability, is a crushing blow to your ego and that's why you believe he's bashing? Its a rhetorical question - frankly I don't care one way or the other about why you purchase vehicles. I just wanted to make it clear, the GP's factual statement is not bashing and that's the facts. Regardless, you're never going to get away from small penis envy where people over compensate, whereas that's true around the world, but singling out G7 countries is hardly bashing.

Did you even read what I wrote? Do you lack reading comprehension skills and just love to rant?

I stated that Asian countries consider image more than the US. So for the commenter to pick out the US and its people is a form of bashing when most humans behave that way and Asian cultures more so than their American counterparts.

Second, form and style is one of the criteria of "merit". For you, it may be lower on your value and priority scale than it is for someone else. But to automatically label something penis envy is just as stupid as to buy an SUV because you think it's safer.

Comment Re:Other venues... (Score 1) 128

What the fuck is with this need to bash countries? Especially when it's so far off the mark.

What the fuck is with this habit to slap "ugly" on everything what is purely functional?

I live in Nepal and you would be hard pressed to find one person to buy that based upon its looks.

That was precisely my point.

Maybe you misunderstood. It is ugly. And that is why people in Nepal would not buy it. Looks matter A LOT. However, even if something is ugly, if it can raise your status and image because it is expensive or rare, then people might buy it.

Comment Re:Great (Score 1) 128

There are lots of motorbikes in India that are under $1000 USD. And they use regular petrol/gas so you can use them like any other vehicle. Pay a few hundred more and the bikes can go fast enough for highways. Bajaj Pulsar is one of the most popular models. This site has prices as well. But it's in Indian currency. You can use Google to convert INR to USD.

Comment Re:Other venues... (Score 1) 128

What the fuck is with this need to bash countries? Especially when it's so far off the mark.

People are MUCH more superficial and image conscious in southeast Asia than the US. I live in Nepal and you would be hard pressed to find one person to buy that based upon its looks. In Nepal, cars are damn expensive since they're taxed 250% as luxury items as imports. A "cheap" Hyundai would cost $18K-$20K. But the roads in Nepal are filled with middle class families buying cars for status.

If someone wants fuel efficiency, they'll buy a model of Hero Honda Splendor or Bajaj XCD. Both of them give over 60-70 km/l. But even then, people don't buy low powered bikes that save fuel.

Nowadays, in Kathmandu, people seem to prefer 180cc to 220cc even though they will never use the power. Personally, I have a Bajaj Pulsar 150cc. Based upon my own weight, road conditions, hills and so forth, I figured 150cc is plenty. I rarely go over 55 kmph since it's hard to find a straight stretch much less clear roads. During the day, it's usually 20-30 stop and go traffic and occasionally 40. While most of India doesn't have the hills of Kathmandu, it does have the congested traffic.

Comment Re:It's cheap compared to India... (Score 1) 1259

In India, you can get 10% interest from certain banks. What can you get in the US? What are the mortgage rates in India and how does it compare with the rates for education loans? So many important factors you've left out in your comparison.

And this doesn't began to cover issues like normal salaries and cost of living and how that relates to whether education is affordable or the job market competition afterward and ability to pay off loans. It's nice to have simple comparisons or explanations. But often, they're biased, grossly inaccurate or a small glimpse which is misleading.

Comment Re:That's what you get (Score 2, Insightful) 124

I'm pretty sure Take Two is doing just fine, even considering this. Also, nothing says "edgy" like "we got sued for something that wasn't even part of our game, because we're just that badass."

Seriously, I'm sure the execs are laughing all the way to the bank. And the shareholders, well WTF do you expect if you invest in a company that makes games where you can get your money back after fornicating with a professional via vehicular homicide?

Slashdot Top Deals

Anyone can do any amount of work provided it isn't the work he is supposed to be doing at the moment. -- Robert Benchley