Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!


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! ×

Submission + - Recurrent Net learns to play 'Neural Slime Volleyball' in Javascript (otoro.net)

Giant Robot writes: Who can resist a cute RNN-powered time waster?

It's not quite Deepmind's Atari AI yet, but this demo shows how powerful even simple recurrent neural networks can learn, entirely through self play, to master ball control, and yet have an air of sentience above traditional game AIs, all within a single browser session. Try the game, and see if you can beat them.

A nice illustrative piece combining processing and Stanford-built convnet.js for visualisation and interactivity — I would actually like to see more AI research presented this way.

Submission + - Neural Modularity Helps Organisms Evolve to Learn New Skills without Forgetting (youtube.com)

An anonymous reader writes: A long-standing goal in artificial intelligence (AI) is creating computational brain models (neural networks) that learn what to do in new situations. An obstacle is that agents typically learn new skills only by losing previously acquired skills. Here we test whether such forgetting is reduced by evolving modular neural networks, meaning networks with many distinct subgroups of neurons. Modularity intuitively should help because learning can be selectively turned on only in the module learning the new task. We confirm this hypothesis: modular networks have higher overall performance because they learn new skills faster while retaining old skills more. Our results suggest that one benefit of modularity in natural animal brains may be allowing learning without forgetting.

Comment China's internet will become a smaller intranet (Score 2) 128

As most websites are no longer self contained, but require numerous dependencies to other websites for data, content, analytics and js libraries, China's gated internet will become more isolated from the rest of the world.

Perhaps Hong Kong may face similar issues with regards to net access and online freedom in the near future? There has been talks about that recently.

Maybe web developers will need to write a "China mode" for front end sites, in addition to "Desktop" or "Mobile" mode that will only use old school 1990's style HTML look and feel. Bring back the frames :)

Comment In Japan (Score 3, Interesting) 331

In addition to very strict gun laws (pretty much the only guys with hunting licenses got them > 50 years ago), there are other laws which are a lot more strict compared to other countries.

For example, if a gaijin resident is caught with light marijuana -> Jail time or deportation. Drinking and driving, even one beer, will cause one to lose his job in a country that prides itself of life long employment.

Comment Q Learning (Score 3, Informative) 93

The methodology deepmind used for training the game player is based on a classical reinforcement learning algorithm called Q Learning (http://en.wikipedia.org/wiki/Q-learning), developed in the late 1980's. This approach of maximizing expected future rewards for the agent to select an action in a current state has some parallels with studies of how the basal ganglia region of our brain conduct reward learning (basal ganglia).

What has been done is to approximate the reward function Q (which originally used a look up table) by a more general function to approach larger problems with much larger (or infinite) number of states. The approach here was to use a function which can fit large amounts of data, in this case a multi layered neural network (with convnet layers to preprocess the raw image input first to identify features) to attempt to learn the game.

This has actually been done a while ago, by Tesauro (now at IBM research) who used the same approach to create a Q Learning agent to play Back Gammon at an advanced level.

The reason why this is new is because in recent years we can employ cheap GPU's to learn exponentially more quicker than conventional cpu's and can construct much larger and deeper networks to learn from more complicated systems. Also many new 'tricks' have been developed to optimize learning in recent years (sigmoid functions replaced by simplified rect linear function, and dropout, etc), so we are going to see better and more amazing uses for this relatively old technology.

Submission + - Google programmer declines $500K offer from startup as he is making $3mm already (businessinsider.com)

Giant Robot writes: In the past we have heard of stories or facebook, twitter, or apple poaching talent from google. Now with all the cash in the bank, google seems to be able to keep who it really wants, and outbid any rivals. In addition to the lure of money, the company has its most capable employees work on important, ambitious projects. For example Andy Rubin who used to lead the Android team now probably runs the group developing robotics technology, rather than leaving to start a new venture.


Rather than posting this story for envy, I'm wondering how typical are salaries > $500k for hired technical staff (as opposed to business stakeholders and founders of companies)? One can imagine someone getting rich from being one of the founders or early employees of a start up, or hired on expensive terms as an executive to manage a business division, lead an large engineering team or project, or world-class researchers solving a very difficult problem, but do we hear often many cases where people are offered salaries 4-5x above the norm as a software engineer or programmer?

Comment God of Gamblers (1989) (Score 1) 320

This is old stuff. Chow Yun-Fat and his cronies have used this technology (in glasses and contact lenses form) to cheat poker over 20 years ago at a private casino event on a yacht near the shores of Hong Kong to get revenge on the man who nearly destroyed him!


Slashdot Top Deals

Faith may be defined briefly as an illogical belief in the occurence of the improbable. - H. L. Mencken