On the Go: New Start
I’m in DC this week in support of an exciting new project. And in totally unrelated news, here’s a picture of me with a fighter jet on my head.
On the Go: 3 Things I Think
Three things that don’t yet require a full post… all after the jump.
Tags: on-the-go, question answeringCode Names and Academic Snobbery
Somebody at Yahoo! has a fondness for not-so-secret code names.
AllThingsD’s Kara Swisher reported yesterday that Y! officials like to use the names American universities to refer to their competitors, partners, or suitors. Alliteration’s a must: Microsoft becomes M.I.T., AOL becomes Amherst, and Google becomes Grand Valley State Georgetown.
In fact, Y! even has a code name for itself: Yale.
I get it: who wouldn’t want to align themselves with these prestigious institutions of higher learning? But as Kara points out, the fit isn’t exactly perfect. Google and Georgetown are both very good at what they do — and their names start with ‘G’. The similarities mostly end there.
But why universities? It’s not a particularly productive set-up for a code. Sure, AOL is Amherst, but where does that leave Amazon or Ask? Austin Peay? Auburn? Abilene Christian? And don’t get me started on the Bs: Bing, Baidu, Blogger, and (e)Bay will all have to slug it out for Brown, Bard, Biola, and Boise State.
Frankly, we should feel for YouTube: with Yale off the board, they’re down to Yeshiva, two York Colleges (one in Nebraska, one in Pennsylvania), Young Harris, and Youngstown State. I bet they’ll go with the Penguins.
What would be more apropos? Harry Potter characters? (Google’s such a Hermione.)
How about car companies? While there are lots of satisfying matches:
- Google-GM
- Amazon-Audi
- Ask-Acura
- Bing-BMW (Buick? Bentley?)
- Facebook-Ford
- Twitter-Toyota
- Microsoft-Mercedes
… it’d leave Y! out in the cold with Yugo. It’s Yale, then! Boola-boola!
Test: Posting from Blackberry for WordPress (Beta)
Just a quickie test of the Blackberry for WordPress app — here’s a picture of Dax McCarty, central midfielder for FC Dallas. Taken 8-01-09 at Pizza Hut Park in Frisco, TX.
For Later: Heilmeier Catechism
As someone who’s constantly evaluating a lot of early-stage research projects, I’ve recently become a fan of the famous (or is it infamous?) Heilmeier Catechism. First attributed to George Heilmeier (wiki) (a former Deputy Director of the Department of Defense, (D)ARPA director, White House Fellow, VP (later CTO) of Texas Instruments, CEO of Bellcore, and Chairman Emeritus of SAIC), these questions were designed to gauge whether a research project was significantly mature to warrant R&D investment.
They look easy, but they’re deceptively hard to answer.
- What is the problem, why is it hard? (What are you trying to do?)
-
How is it solved today? (Why couldn’t we do this before?)
-
What is the new technical idea; why can we succeed now? (What’s the breakthrough that makes you think you could do what people couldn’t before?)
-
What is the impact if successful? (Who cares?)
-
How will the program be organized?
-
How will intermediate results be generated?
-
How will you measure progress?
-
What will it cost?
Better Living Through Semantic Parsing
Unless you’re a recovering linguist, English major, or Latin teacher, chances are that you view grammar as a particularly nasty form of torture, one that was effectively wielded by such language-mavens as Torquemada, leader of the Spanish Inquisition, and Ms. Grizzwald, your 6th grade English teacher.
For most of us, the intimate couplings of nouns, verbs, and adjectives just aren’t that sexy. And why should we care? We get along fine without knowing the difference between definite and indefinite articles. Our ability to chat with friends, order pizza, or, heck, woo women (!) isn’t enhanced by any knowledge of gerunds and other assorted verbals. And in most cases (notwithstanding guys like Faulkner and Joyce), we’re not any more (or less) able to understand text, just because we know how all the little twiddly bits go together.
But semantic apps aren’t like us. Read more after the jump…
Avoiding Search Overload
Like you, we’ve heard a lot this summer about the challenges facing America: the financial crisis, healthcare reform, and worst of all: search overload.
Well, here at Swingly HQ, we’ve been doing our part. We’ve been trying to find new ways to figure out what kinds of information are most relevant to a particular search topic.
While relevance modeling isn’t exactly new, it’s becoming an increasingly important problem for semantic search applications. Information Extraction apps are rapidly increasing the amount of factual information that’s available from the Internet. That’s good. Unfortunately, instead of being buried under mountains of irrelevant information, we’re now being overwhelmed with gigabytes of factual information which may (or may not) be exactly what we’re looking for. That’s bad.
So, what’s a new semantic search app to do? Full details after the jump.
Tags: relevance, search, semantic search, SwinglyBack in the Lab
- out of the office: montara beach, ca
After what felt like a month of traveling, I’m finally back in the lab here in Dallas in August. Yippee. While I’m not excited about the DFW late summer weather heat, I’m definitely excited about some new breakthroughs that the Swingly & LCC dev teams have made recently.
10 Things to Know About Swingly
This post originally appeared on April 30, 2009.
Greetings to everyone who read about Swingly in Marshall Kirkpatrick’s excellent ReadWriteWeb article.
10 things you should know about Swingly appear after the jump…
Three NLP Techs to Watch in Next-Gen Semantic Search Apps
This post originally appeared on April 15, 2009.
Arun Radhakrishnan over at Search Engine Journal set the Twitter-sphere all a-buzz on Monday with his profile of 9 new semantic search engines that are looking to change the face of search. (For the record, the 9 included Hakia, Kosmix, Cognition, Lexxe, Exalead, Factbites, Swoogle, Sensebot, and Powerset. Helpful commenters also included pointers to Evri, DuckDuckGo, and Semote as well.)
Given that excellent introduction, I thought I’d contribute a short post on a couple of the natural language processing (NLP) technologies that are making these new kinds of search applications possible. (This is by no means a comprehensive list — nor is it an endorsement or repudiation of any particular approach to semantic search. )
So, without any further ado, here are 3 techs to watch: