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:You are not alone. (Score 1) 354

One example of a negative reaction to a suggestion I was surprised by had to do with Diablo III. They have 2-3 month "seasons" where you can create a new character and you get special rewards if you achieve certain things with that character in the timeframe of the season. The timing of the seasons was such though that they tended to end just as students would get out for winter or summer breaks. One person made the suggestion of shifting the seasons so that students could have more of an opportunity to participate. This led to a giant flamewar where people abused these "dumb kids" and "lazy students" for dare wanting to play Diablo III on their winter break. This wouldn't require any code changes, just an adjustment of when Blizzard hit the "start" and "stop" button for the season. People will get angry about anything.

Comment Re:And how many training-test cycles did they do? (Score 4, Informative) 47

This is not accepted practice in the machine learning field. A lot of people split the training set into training and validation , and "test" on validation. When they have their parameters set, then they perform a final test on the test set. For some datasets, they might not even be able to directly access the test set answers, they might have to go through a 3rd party server with limitations on how often they can submit.

Comment Re:Dear Nintendo (Score 2) 104

Other legal alternatives:

Play them on the Virtual Console for Wii, 3DS, and Wii U (and probably Switch eventually). That is, most Nintendo consoles in the past 11 years.
Play them in other repackaged forms on Nintendo consoles that have been coming out since the SuperNES (e.g. Super Mario All-Stars).
Play them on the original hardware that you can find in garage sales or eBay.

Comment Classics (Score 4, Informative) 259

I'm trying to read some classic Western literature to see what thought processes led to current Western culture. Currently I'm reading the Tragedies of Aeschylus (Agamemnon specifically). Encyclopedia Brittanica put together a list of the books they thought were most influential throughout Western history: . My goal is to work my way through all of them, eventually. There's a good variety: literature, philosophy, history, theology, math, and science.

Comment Re:Doesn't anyone pay as they go anymore? (Score 3, Interesting) 364

At least for a PhD, most people don't pay their own tuition, and even earn a small stipend for the work they do. They either get fellowships, graduate research assistant jobs, or other funding from the professor they do research for. It's not a lot of money, but it's at least "income" rather than "goes out."

Comment Re:"developed an artificial intelligence(AI) progr (Score 1) 153

In addition, it's such an ignorant devaluation of AI's incredible achievements. The field of AI, a term everyone except Slashdot naysayers have agreed upon, has performed incredible feats of image and language analysis (among other things) that the average software engineer stringing together API calls could never have come up with.

Honestly, if you told someone to just "write a program" to recognize whether an image has a cat in it or not, what do you think they would come up with? Probably 20,000 lines of messy code that only handles a handful of cases. The insights of learning parameters from data, optimization techniques, and convolutional neural networks were found by groping in the dark for 50 years. For 50 years people couldn't just "write a program" to do that simple task. For a long time people couldn't even write programs that recognized the 10 handwritten numerical digits with competitive accuracy to a human.

Oh my gosh, why does everyone make a big deal about "making cars?" It's called "engineering a solution," and we've been doing it for millenia! Can we just call "cars" what they really are, "engineered solutions?"

Comment Re:Defies the purpose of competition (Score 1) 127

When people refer to "democratizing AI," they probably mean:
  • 1. That the software tools (like Tensorflow or Microsoft's own CNTK are free and open source
  • 2. That it can be used on commodity PCs (with GPUs)
  • 3. That the education is free (there are many, many free online resources like MOOCs that teach machine learning principles)
  • 4. Even datasets used to train deep neural networks are free (Imagenet, Pascal VOC 2012, MS COCO, Youtube 8M).
  • 5. Even the latest academic scholarship is quickly published publically to ArXiV while sitting in the traditional academic publishing pipeline.

Machine learning is still heavily an academic discipline, but it's never been easier for a layman or business to use and benefit from the technologies.

Comment Re:But the median college-educated.... (Score 1) 495

Umm, or GP is exactly correct. When I was in middle school there was a poster outside the guidance office at school that specifically says that high cost is not a reason not to attend a college. What's important is that I choose the college "right for me." Cost doesn't matter because the average college graduate makes more than the average high school graduate. That statistic, while true, has been tortured to death to justify any college expense.

Fortunately my parents and I were a bit more practical when I went to college, but I know friends who got "the right degree" (i.e. engineering) who have good jobs who will be paying off their debt for 15-20 years.

Comment Re:Automated Post (Score 1) 409

I'm guessing that the "half" of work they quote are tasks that are basically pattern recognition exercises or are straightforward application of rules. It's work that at some point someone will find it economically feasible to automate (or economically feasible to sell an automation solution).

Even work that isn't "practical" to automate now is being picked at by AI and robotics research wherever it can be. For example, robots that can learn by example and can work in close proximity to people: https://www.technologyreview.c... This would find a nice savings in between 100% trainable human labor (expensive) vs 100% inflexible automated robot process (also expensive).

Slashdot Top Deals

Disk crisis, please clean up!