fitzwatermellow

Does anyone else see the similarity between designing Redis keys and game art asset filenames?

Being able to mine as much metadata from a file name and path is a plus. As is having a logical naming convention that can be queried from the string programmatically. Not to mention fast lookup and asset name generation merely from string concatenation...

This is a good problem to have, mfcecilia! You are on the path to making great games. Its stuff like this that gives you the confidence to say my solution is just as good as anyone else's and I am just going to do it and learn from my mistakes. Time pressure is actually your best ally. If you want your game to be playable by the end of the decade you will solve this challenge and level up to the next big one. Such as figuring out how to organize your assets in a single compressible pak that is laid out to support fast and efficient patching via binary diffs!

Good luck! And if you are still stuck check out the Unity and Unreal engine docs and forums for further tips on pipeline production. And of course, post a link when you build a playable alpha ;)

Expecting to hear more Vulkan details from Khronos at the Game Developer's Conference 2016 in March:

https://www.khronos.org/news/events/gdc-2016-khronos-session...

Thanks, cfcef! Implementation details are in the "Methods" section. Have started experimenting with small GPU ML cloud jobs and the costs do add up. Wanted to get a sense what a large job looked like and indeed, AlphaGo is gargantuan. 50 GPUs approx train time one month for the policy/value network. So, a Google R&D size budget would be a prerequisite ;)

Official Google Blog post also has some background info:

AlphaGo: using machine learning to master the ancient game of Go

https://googleblog.blogspot.com/2016/01/alphago-machine-lear...

Interesting that DeepMind was using Google Cloud for compute. I imagine that the MCTS expansion can become massive. Any chance DeepMind may publish some of the internals about how many instances were used, how computation was distributed, any packages or frameworks used, etc.

And congrats on achieving this impressive AI milestone!

From the F.A.Q:

Accelerated Mobile Pages are just like any other HTML page, but with a limited set of allowed technical functionality that is defined and governed by the open source AMP spec

And from the Spec:

Enable the AMP runtime to manage the loading of external resources, which may slow down the initial render or cause jank. Allow AMP authors to include functionality above and beyond standard HTML, while maintaining the security- and performance-minded requirement that no author-written JavaScript is executed

Appears we won't have access to Canvas2D, WebGL, WebRTC, Fetch and other HTML5 Web APIs. AMP's subset of declarative elements and caching will be great for longform-y "static" content. The kind of articles I find myself reading more of on Mobile Chrome during down times in transit. Thinking of Medium confessionals and NewYorker dispatches, etc.

But how do you implement something like ChartBeat? Snapshots of dynamic chart images and data pre-rendered on the server and pushed out to the cdn? Sort of eliminates the performance benefits of geospatial caching doesn't it...

I do take heart however in Tony Haile's pull quote from the ampproject.org homepage:

The mobile open web experience is terrible and some have suggested it's too slow to compete. We can choose to see the mobile open web as a relic of its time and flee to the warm embrace of closed platforms and apps. Or we can say that the open web means something important to the world and if it’s broken it’s our job to fix it. I think we should fix it.

Thanks for the replies! Indeed, Multiple goroutines can receive on a single global channel to create a simple worker pool in a fan-out configuration. Analogously a second fan-in channel can be used to merge the parallel computation. With the caveat that the channel should be closed properly to make sure all tasks are complete.

Inspired now to start a "go-saturate" library for measuring max net capacity...

Great summary of current state of the art with links to interesting projects: GVG-AI, VGDL...

Videos games are also essential for AI pedagogy. Creating Pac-Man agents in Stanford's AI class is a great example. Most players can barely get a "strawberry" but to see a trained agent mimicking human expert level play is eye-opening.

Quick reminder: Global Game Jam 2016 starts Jan. 29 and NYU is hosting its annual jam!

http://gamecenter.nyu.edu/event/global-game-jam-2016/

Sorry, but is this golang concurrent networking pattern correct:

    func main() {
        ch := make(chan string)
        for i := 0; i < MAX; i++ {
            go fetchCert(ch)
        }
        scanner := bufio.NewScanner(os.Stdin)
        for Scanner.Scan() {
            line := scanner.Text()
            ch <- line
        }
    }

All goroutines receive on the same channel! Instead a new goroutine should be launched for each net conn. One should be able to spawn 1000s (or 1Ms) of conns and avoid ulimits using buffered chans, waitgroups, timeouts, or counters...

Kudos to the Princeton team on yet another interesting find! I recently submitted the work on de-anonymizing programmers from binary signatures using machine learning that was also fascinating (not to mention a bit scary)

But I confess I'm perplexed as to how a canvas render can yield a consistent unique fingerprint? Even allowing for subtleties in font anti aliasing implementations amongst drivers, wouldn't two machines with identical gpu hardware and drivers be identical down to the last pixel? And what about browsers using software rendering?

My interest is purely academic of course ;)

Happened to catch rockstar astrophysicist Neil deGrasse Tyson on Charlie Rose the other night and the subject of asteroid deflection came up. For small candidates, vaporization would be ideal. But for larger ones, any attempt at destroying them may have unwanted consequences. Debris fields, partial cleaving, etc. Sending a team of human miners with demolition expertise is not in the cards. So the current vogue in asteroid deflection would be to send something heavy to intercept the asteroid and deflect it a degree or two using only gravitational pull! A very clean and elegant solution that would no doubt make a great simulation or video game ;)

Mission Concepts and Operations for Asteroid Mitigation Involving Multiple Gravity Tractors

http://www.ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/201...

From the F.A.Q:

Q: What kinds of things can I use Magic+ for?

Anything you want. As long as it’s not illegal. Seriously.

"I need you guys to unscramble this omelette by 5pm..."

Kidding, and violating laws of nature aside, am rooting for the Magic+ team. Good luck!

Watch Ikiru. Be enlightened.

http://www.hulu.com/watch/215850

A more interesting line of inquiry is why so many of those blessed with the highest instincts for creativity have so little empathy towards those around them. As if they are so in thrall with their own genius they have little capacity for the smallest kindnesses. Or are so focused on manifesting the perfect visions contained in their minds, any quotidian concerns seem far beneath them and an utter waste of time. Consider Tolstoy and his wife Sofya, or Picasso and his muses. Einstein even referred to a "glass pane" separating himself from the norms of society.

Apart from Crispr, Deep Learning, Autonomous drone swarms, and Mars missions, an important development in human society seems to be the dawning of an era in which "crime" becomes impossible.

Accuracy in determining our genetic and digital fingerprints, combined with ubiquitous surveillance and predictive machine learning may make going "on the lam" prohibitively difficult. In concert with LE's newfound prowess would be a subtle but pervasive mass media campaign that targets pre-criminals with cautionary tales of the apprehended. One hopes that progress in rehabilitation and societies capacity for forgiveness matches these developments apace.

ps happy 2016 HN!

Great stuff from MSR! This particular "synthesizer" translates simple queries such as "generate md5 hash" into C# API calls. Wondering if there are any analagous projects targeting NLP code snippet generation for bash shell scripts?

Congrats! Massive step into the era of reusable rockets.

Direct link to paper:

Combining Model-Based Policy Search with Online Model Learning for Control of Physical Humanoids

http://www.eecs.berkeley.edu/~igor.mordatch/darwin/paper.pdf

Chapter 14 is twenty-four carat gold. Would love to see optimize.com as a dedicated site with Rosetta Code-style ports of optimization recipes across languages, platforms ;)

Quick link to grab all five manuals:

http://www.agner.org/optimize/optimization_manuals.zip

An impending arms race between quants on Wall Street could also provide a colossal impetus for driving QC innovation:

Quantum Computers Entice Wall Street Vowing Higher Returns:

http://www.bloomberg.com/news/articles/2015-12-09/quantum-su...

Companion article to original paper:

Paradox at the heart of mathematics makes physics problem unanswerable: Gödel’s incompleteness theorems are connected to unsolvable calculations in quantum physics.

http://www.nature.com/news/paradox-at-the-heart-of-mathemati...

Toby Cubitt, a "quantum-information theorist" at UCL should win this years relevant username contest ;)

Here's the paper:

Better Computer Go Player with Neural Network and Long-term Prediction by Yuandong Tian, Yan Zhu

http://arxiv.org/abs/1511.06410

"Adding MCTS to darkforest creates a much stronger player: with only 1000 rollouts, darkforest+MCTS beats pure darkforest 90% of the time; with 5000 rollouts, our best model plus MCTS beats Pachi with 10,000 rollouts 95.5% of the time."

Noticed that as well. Still writes with such reverence for the power mathematics has in unveiling the mysteries of nature. Particularly enjoyed the moniker "distinguished differential geometer" to describe Calabi! I guess its only a matter of time before "wavelet nets" are re-discovered and resurrected as the "new innovation" in machine learning ;)

Wonderful site! Includes a comprehensive open-access encyclopedia of photonics and links to suppliers.

As to the actual question of storing light, article admits the impracticality of any sort of "box o' mirrors" approach. Although truthfully it would make a fun experiment. I think nature provides a much more elegant solution in the form of crystals that can absorb and re-emit photons when excited, of which modern tech has made ample use thereof.

The name of the program is "LIQUi|>" and is an acronym for "Language-Integrated Quantum Operations". Direct link to MSR page http://research.microsoft.com/en-us/projects/liquid/

HN formatting cuts off trailing ">" char in title for obvi reasons ;)

I was doing some research for a small demo. As a proof of concept, implementing a small ConvNet using WebGL shaders. But in the course of my search I stumbled on this very broad and interesting patent granted to Microsoft Corp.

Processing machine learning techniques using a graphics processing unit

http://www.google.com/patents/US7548892

Pretty amazing when you consider they wrote the application a decade ago! There is no arguing we are experiencing a golden age and an embarrassment of riches provided by ML toolkits and GPU cloud capabilities. But further inspection of patents in the space brought up this gem, in which Emotient is attempting to patent the crowdsourcing of training data...

Collection of machine learning training data for expression recognition

https://www.google.com/patents/US20150186712

This AWS re:Invent 2015 talk by Dave Hahn is also really illuminating: A Day in the Life of a Netflix Engineer (Using 37% of the Internet)

https://www.youtube.com/watch?v=-mL3zT1iIKw

Interesting to me that they use Atlassian Confluence for team stuff. Thinking of testing it out as well...

Great job, guys! The Docker of Neural Compute..

Here's a link to the original announcement:

Announcing CGT: http://joschu.github.io/

"Compelling stories. Intense multiplayer. Endless replayability. Qualities that made StarCraft, Warcraft III, and Diablo II the titans of their day," Blizzard said. "Evolving operating systems, hardware, and online services have made them more difficult to be experienced by their loyal followers or reaching a new generation. We're restoring them to glory..."

Wondering if the best strategy wouldn't just be porting to Unity?

OTOH, some of the best algorithmicists I know hail from former Soviet-bloc repubs ;)