Forgot your password?
typodupeerror

Comment: Add noise to fix (Score 1) 230

by SpinyNorman (#47100729) Attached to: The Flaw Lurking In Every Deep Neural Net

If the misclassification only occurs on rare inputs then any random perturbation of that input is highly likely to be classified correctly.

The fix therefore (likely what occurs in the brain) is to add noise and average the results. Any misclassified nearby input will be swamped by the greater number of correctly classified ones.

Comment: Re:Que Oversaturation in 3...2...1... (Score 1) 126

by SpinyNorman (#47082319) Attached to: US Wireless Carriers Shifting To Voice Over LTE

This will actually help!

First, voice doesn't use that much data. For example, Viber (a popular VOIP app) uses 0.5MB/min which would be about 0.5GB for 1000min.

More importantly, once every one is transitioned off 3G onto 4G/LTE (i.e. VOIP over LTE) the carriers can repurpose the 3G spectrum for 4G and thereby gain more 4G/LTE capacity.

Comment: Re:Don't understand it. (Score 1) 198

by SpinyNorman (#46959781) Attached to: Apple Reportedly Buying Beats Electronics For $3.2 Billion

The deal doesn't make sense to me, but presumably it would involve Dr Dre and Jimmy Iovine being contracted to stay for some minimum amount of time, which brings a lot of clout (esp. Iovine) in the music biz.

The $3.2B price if true seems insane though. Between 2012 and 2013 Beats bought out HTC's 50% ownership for a total of $415M (25% in 2012 for $150M, 25% in 2013 for $265). So, if half the company is worth $415M, the whole thing should be worth closer to $430M, not $3.2B!

Comment: Re:The more simple you make it the less complex it (Score 1) 876

by SpinyNorman (#46195787) Attached to: Ask Slashdot: Why Are We Still Writing Text-Based Code?

In practice, I believe that the present text-based programming paradigm artificially restricts programming to a much simpler logical structure compared to those commonly accepted and used by EEs. For example, I used to say "structured programming" is essentially restricting your flow chart to what can be drawn in two dimensions with no crossing lines. That's not strictly true, but it is close. Since the late 1970s, I've remarked that software is the only engineering discipline that still depends on prose designs.

You appear to be thinking about a very limited subset of software where the essence is captured by the "two dimensional" control flow.

As Fred Brooks famously wrote: "Show me your [code] and conceal your [data structures], and I shall continue to be mystified. Show me your [data structures], and I won't usually need your [code]; it'll be obvious.''

Nowadays he probably would have updated that pithy formulation to include mention of your threading model as well as data structures.

If you start trying to visualize the dynamic behavior of complex synchronization-heavy multi-threaded programs or ones with significant non-trivial shared data structures, then I can assure you there'll be plenty of crossed lines!

The time when most programs could be described by flowcharts was probably 40 years ago. We've moved on a bit since then!

Comment: We've mostly got the tools we need ... (Score 1) 876

by SpinyNorman (#46195469) Attached to: Ask Slashdot: Why Are We Still Writing Text-Based Code?

I think part of what your missing is based on your own self-described lack of experience.. that you can write simple programs but get bogged down writing more complex stuff. Professional programmers don't really have this problem (or at least the experienced ones don't - there is a learning curve as in any field).

The main "trick" to designing/writing complex programs is to be able to think at many different levels of abstraction and therefore to "divide and conquor" the complexity. At each level of your program (think of it as a layered onion from the highest level on the outside down to the low level stuff of simple programs on the inside) you're going to be implementing one level of complexity/capability by using software components that are essentially only ONE level lower in capability than the level you're at... ditto for the next lower level, etc, all the way down. Designed this way, it's no harder to write the highest levels of the program than it is to write the lowest levels that you are familiar with.

Note though that the software components you're using at any level of your design are going to be domain-specific components that you've designed yourself to make the job easy - they are not going to be 100% off-the-shelf components (other than company internal re-use), other than at the bottom most layer of the design. It's having the right components that makes the job of implementing the next layer up easy (like the idea of the adjacent possible - without having "adjacent" components the corresponding "adjacent possible" is NOT possible, or at least is way more difficult). So, the real issue is not whether one is using a visual vs text-based method of composition but rather in having (creating) the right components at each level.

It's also worth noting that since programmers are the ones with the skills to create programming tools that we therefore necessarily have pretty much the tools we need. Good programmers are lazy (strive for minimalism), and arn't going to fight the same battles every day if they can build a tool to make their lives easier. Of course there's always a bleeding edge of new technology where the tools havn't yet matured (e.g. now that clock speeds are topping out and parallelism is replacing it, there's more need for better tools to deal with parallelism), but basically we DO have the tools we need.

Comment: Re: "Driving like a fool" (Score 1) 666

by SpinyNorman (#45306543) Attached to: Atlanta Man Shatters Coast-to-Coast Driving Record, Averaging 98MPH

Presumably you're assuming a 6 chamber gun with one bullet and "therefore" a 1-in-6 chance of getting shot.

However... due to the weight of the bullet, when you spin the chamber prior to firing, the bullet will tend (due to gravity) to end up in one of the lower vs the highest (firing) positions, so the average chance of getting shot should actually be somewhat less than 1-in-6.

Comment: Re:scale (Score 1) 327

by SpinyNorman (#44404967) Attached to: Tim Cook May Not Know Why, But Samsung Is Winning in China

The story is clearly about the competition between Apple and Samsung, and fact is that Samsung now makes more profit ($5B last in most recent quarter - that's earnings, not revenue) from smart phones than Apple does.

Apple has mostly saturated the US market and to save the stcok price from collapsing needs to find other markets for growth... China was one big hope, but it appears it's not happening. That's certainly news.

Comment: Re:For how long...? (Score 1) 106

by SpinyNorman (#44404935) Attached to: News Worth Buying On Paper

And now someone like Henry Blodget is trying to say that newspapers need stuff that can't be found elsewhere to survive, which basically means to become the local gossiping outlet?

That's your conclusion, not his.

Investigative journalism is indeed unique content - the product of unique real people with the investigative skills to unearth and develop these stories. Watergate didn't hit the news via a syndicated Whitehouse press release.

You're also going to find more compelling content in media that is published less frequently, or in occasional rather than daily editorials. At web-speed you're just going to get a firehose of daily chatter.

FORTUNE'S FUN FACTS TO KNOW AND TELL: A black panther is really a leopard that has a solid black coat rather then a spotted one.

Working...