Monday, August 18, 2008

Two Complaints

Briefly, two problems that should already be solved:

  1. Text Messaging: Why are we still being charged for ~150 measly little ASCII characters? Why hasn't instant messaging on mobile devices entirely replaced SMS/MMS? IM is free and SMS/MMS costs ~$0.20/each or ~$20/month unlimited. In fact, its actually more expensive to send data via text message than the Hubble Space Telescope.

    Yes, I know there are currently IM apps on cell phones, but they're really no good and not well integrated into the device. But poor execution doesn’t damn a plan (or technology). Given the right implementation, this should be seamless and invisible, the only noticeable changes being the elimination of SMS charges on my bill and the ability to send short messages to any of my friends on any mobile or IM network. In fact, Apple could do this with my iPhone, keeping the same UI & everything, and I would be none the wiser, but of course AT&T would never allow them to eliminate such a valuable revenue stream.

  2. Social Media Status Updates: Why do I have to log into several different sites to ensure all my followers across the web can enjoy my witty status updates? Why can't there be one site that allows me to fire off an update via web or mobile, and my Facebook, Gmail, Twitter, LinkedIn, etc. statuses are immediately updated? Bonus points for periodically checking my pages to see if I've made a rogue "native" update and keeping the rest in sync.

Sunday, March 23, 2008

DOD's "Sentient" World Simulation

Apparently the US military is looking into developing a massive simulation of the entire geopolitical canvas of the Earth, with a node representing "every man, woman and child." Although I'm sure it involves lots of data and some serious computation, I sincerely doubt the predictions made by this system will be any better than that of a seasoned human analyst.

I suppose you could turn over strategic nuclear decisions to this thing and just wait for Judgment Day...

While articles of this type are often titillating, I know from first-hand experience that most far-out projects involving the military don't live up to the hype. For one thing, the military plays fast & loose with commonly used terminology. The word "sentient" actually means something; it shouldn't be used as a vacuous buzzword designed to woo congressional funding.

Maybe they should call it WOPR 2.0?

(Sometimes I just can't help myself with the cultural reference one-liners!)

Friday, March 21, 2008

Are you smarter than a four-year old?

RPI researchers have apparently created an avatar in Second Life driven by an AI equivalent to a four-year old human child. Sounds pretty interesting, and could be a good reason to check out Second Life, although I imagine it would be difficult to locate this particular character in the "game."
Currently, the team is grappling with computational tractability issues to do with the sorting of growing amounts of knowledge that is collected as a artificially intelligent character matures.

Massive parallelism perhaps?

A possible future application of this technology could be cost-effective solutions for pesky child-labor laws for TV game shows...

Wednesday, February 27, 2008


In another (yes, another) sensationalist attempt to attract readers and sell ads, we find an AFP article entitled Automated killer robots 'threat to humanity': expert.

Just some more fear-mongering from our wonderful media if you ask me.

So we can add psycho killer robots to the list of things to be worried about in this troublesome modern age. But I thought Hollywood already taught us that? In fact, the article even refers to The Terminator.

Tuesday, February 26, 2008

If It Ain't Broke, Don't Fix It

As mentioned earlier, my department recently held a very successful Computer Science Research Day. The keynote speaker was Dr. Guru Parulkar, Stanford researcher (and UD alum) working on a project called the "Clean Slate Design for the Internet." The goal of this work is to re-design the Internet with 'modern' features that will improve security and quality of service, as well as provide new classes of services.

I viewed his entire presentation as a direct attack against Network Neutrality. The basic premise was that the Internet is broken, it's getting worse, and it needs to be fixed. How do we fix it? Increased 'control & management' by ISPs and government. A central theme seemed to be finding new ways for ISPs to monetize traffic and develop 'value-added services.' In other words, getting me to pay for things I can get for free elsewhere. This will be accomplished by denying or degrading access to competitors in the name of 'security,' 'reliability,' and 'return on investment.'

My question to the speaker was, "Many people would argue that the freedom and success of the Internet is a result of what you refer to as the 'dumb' infrastructure. If we were to implement the changes you propose, how would Network Neutrality survive?" [paraphrase]

Dr. Parulkar's response was based on the claim that it wasn't 'fair' that Google made billions while Comcast's market capitalization remained stagnant. The last time I checked, these companies were in entirely different businesses, although Dr. Parulkar conveniently lumps them together as "service providers." Why in the world would I ever want to pay Comcast (or Verizon, or Time Warner, etc.) additional fees for poorly designed implementations of services I can get for free elsewhere? I pay my ISP for access to the Internet, and Google et al. pay ISPs for their net access, so why do the ISPs believe they are entitled to any additional remuneration for this connection?

The phrase "Network Neutrality" was not mentioned until I brought it up in my question. I believe the subject was intentionally avoided.

ISPs are a utility, just like the electric company. I just want to pay a competitive market price for unfettered Internet access with no strings attached. Is that really an unreasonable expectation? Unfortunately, many ISPs don't understand the business they are in, and apparently, neither do many researchers.

It's ironic that Stanford, original home to Google who has thrived due in large part to Network Neutrality, is leading the charge against it.

Apparently I'm not the only one who takes issue with the underlying purposes of this project.

Finding Myself

So I was Googling myself recently (which isn't as creepy as it sounds) and discovered that I was acknowledged in a paper written by a former instructor. The paper discusses a capstone project in my senior software engineering class (CS 482) at Iona College. My group built an online course evaluation system, similar to the bubble-sheets used by many students at the end of each semester.

I suppose it's nice to be acknowledged, but being listed as a co-author would have been even better... especially when I was applying to grad school!

For those with journal access, the paper can be found on
A software development project: a student-written assessment system.

Friday, February 22, 2008

CS Research Day 2008

So, I have to say that (despite the weather) the first annual Computer Science Research Day at UD was a success. The keynote speaker from Stanford provoked quite a reaction, at least from me (more on that soon). During the panel discussion on Privacy & Security regarding digitizing medical records, Emily made a good point by suggesting that the controversy could be avoided by giving individuals the ability to choose exactly what data would be made available.

I presented a poster about ICICLE (my first as a grad student), and ran some demos afterwards to anybody who was interested. In addition to ICICLE, I also demonstrated my Answer Machine and summary program. I'm hoping to have a CGI version of the summarizer online soon for anyone to try out.

By the way, I'm thinking of calling it 'gist,' which could stand for either "Greenbacker's Instant Summary Tool" or "gist Is a Summary Tool," depending on how narcissistic I'm feeling at the moment...

Wednesday, February 20, 2008

What I cannot create, I do not understand

This excellent blog post draws parallels between Feynman's observations regarding the Challenger disaster and current trends in Software Engineering. Makes for a rather interesting read and is worth checking out. In particular, I appreciated the bit about unit testing and 'step by step' increases -- things I believe good programmers should constantly strive for in single-author programs... but which may become increasingly difficult/expensive in larger endeavors.

(via Slashdot)

Tuesday, February 19, 2008

Implications of a Closed Universe

Does infinity actually exist, as something more than just a concept? If the universe is a closed, finite system (which is currently unknown) then it cannot. I believe mathematical concepts are not real and valid unless they can be expressed in a physical system. So, if we cannot have an infinite number of any 'thing,' then infinity is an invalid concept. There are no 'Platonic ideal forms.' Just because we can imagine something does not grant existence to that thing. I can imagine lots of things, like unicorns and flying pigs, but they aren't real.

So, if it turns out that we live in a finite universe, and infinity doesn't exist, then time is finite too. Time will eventually end and existence will cease, in a 'Big Crunch' or similar event. Would this provide a solution to the Halting Problem then? If there is an end of time, then no program or algorithm will run forever, so for all X, HALT(X, X) = TRUE.

I should probably file this under Ramblings...

Monday, February 18, 2008

21 Years From Now

Ray Kurzweil recently said,
We will have both the hardware and the software to achieve human level artificial intelligence [...] by 2029.

Ever the optimist, Kurzweil apparently believes horsepower & libraries are the things holding back AI. I would argue that the tools are relatively inconsequential; ideas and techniques are what matter. The Wright Brothers combined decades-old equipment & materials in a radical new way to build the world's first airplane. My intuition is that the code expressing human-level AI will be elegantly efficient and small enough to run on current systems. Unfortunately, innovation doesn't easily lend itself to predictions and prognosticators.

By the way, 2029 is the year John Connor sent Kyle Reese back in time to stop the Terminator. Coincidence?

(via Slashdot)

Friday, February 8, 2008

The Future of Warfare

Wired recently covered Israel's plans to develop a powerful artificial intelligence to take over military command & control in the event of an emergency, leading to some nervous thoughts about Skynet.

This is where all global human conflict is heading. We've built machines to replace many jobs humans don't want to do. Thinking and fighting are the next logical steps.

Every modern military should be working towards this. Maybe they already are in secret...

Thursday, February 7, 2008

Baby Data Mining?

Researchers performing cognitive experiments on toddlers claim that children acquire language through a process similar to data mining. The toddlers were shown a series of object pairs while recordings of two words were played. During each iteration, one of the words corresponded to one of the objects. Over time, the children were able to disambiguate the references and determine which words corresponded to which objects.

While the scientists claim this language learning method to be akin to data mining, to me it seems less like a statistical searching process and more like rule-based pattern recognition and association building.

via Slashdot

Gas-Pumping Robot

First vacuuming the rug, then moving the lawn, now robots can pump your gas for you. If this operates as well as advertised, the designers must have overcome several significant vision and pattern recognition challenges to make it work.

I wonder if they'll allow these in New Jersey?

And when will it be upgraded to wash your windows and check your oil?

via Slashdot

Tuesday, January 22, 2008

Using PageRank in a Brain

Following up on my post about modeling & consciousness...

The use of a PageRank-like algorithm would explain some behaviors of the human brain. The most popular ideas or concepts require less processing time to recall, and the less popular take more. Likewise, the most popular pages show up first in a Google search, and you have to dig down deeper to find the less popular. The PageRank algorithm constantly updates; as new pages are indexed, the scores of the pages they link to are refreshed, just as the neural connections between frequently access bits of memory are strengthened in the human mind.

Other people have suggested that PageRank works like the human brain, and it makes a lot of sense to me. The amount of information collected by a brain (of any type) is vast, and requires some efficient method of determining what's important and what's not.

Outline of Conscious Modeling Framework

First post in a long time. I had trouble falling asleep last night, and this is the idea that was floating around in my mind:

  1. Scan input for modeling (this input could be a sentence, image, or even perfect knowledge of a closed system provided by some form of perception).

  2. Identify & label things (nouns) as "objects," actions (verbs) as "functions," etc.

  3. Use (& dynamically update) dictionary/knowledge base to fill out this working model by expanding characteristics & features of the component items to some reasonable limit.

  4. Cross-reference with Google-like search of memory using something akin to PageRank to generate connections, etc.

  5. Update PageRank scores with new data collected from current modeling.

  6. Commit new ideas into permanent model.

But can something as simple as this be robust enough to build a foundation of consciousness from?