Machine-Powered Medical Info: HealthBase Semantic Search
We've all seen how semantic technologies improve search results, but rarely do we see those results put to use in such a targeted way. Jens Tellefsen, VP of Marketing and Product Strategy at NetBase Solutions spoke to ReadWriteWeb about today's launch of healthBase - a medical search and discovery application. Using a variety of semantic indexing techniques, the company crawls the web's leading medical and health players including the Mayo Clinic, PubMed (US National Library of Medicine) WedMd, Medical News Today and Discovery Health. What makes this a truly unique technology is that rather than requiring any data manipulation from humans, Netbase's search results are completely automated.
Says Tellefsen, "Rather than using keywords or basic entities to search through billions of documents, NetBase can actually read and extract linguistic meaning from entire sentences and concepts." According to Tellefsen, healthBase can determine causal relationships, treatments and conditions and automatically aggregate that data into meaningful answers. Given the fact that more than 75% of the population seeks out online health information, a semantic tool with sentence-level understanding can potentially help dispel medical myths on a massive scale.
NetBase employs the same principals across a variety of enterprise tools, but healthBase is its first foray into consumer-facing products. While the company is used to powering corporate, federal and market research, healthBase allows NetBase to show off its content intelligence tool in a way that gives us insight into our selves and our bodies.
Because NetBase is not reliant on manual annotation or custom taxonomies, the system is also very scalable. It took roughly 2 days to produce all of the data in healthBase - a feat that would never be possible by a combination human and machine system.
"It's important for us to address real issues with semantic technologies outside of a lab," Says Tellefsen. To try healthBase visit healthbase.netbase.com
Share
Facebook
Comments
Subscribe to comments for this post OR Subscribe to comments for all ReadWriteWeb posts
There is no future in this.
This search engine absolutely sucks.
According to it, Jews and magnetic fields are two of the leading causes of AIDS
http://healthbase.netbase.com/#aids&Causes
Useless. Even if the semantic technology was working, which is isn't (comment 2), this overlooks a couple of fundamentals: 1 the problem with health information is locating quality by being selective with sources, which healthbase doesn't do, and 2 health data probably has more well structured metadata than any other area, so why not use it? Netbase have made a mistake by releasing Healthbase, rather than encourage use it will make people run a mile.
Come on, it cannot be that hard. Other semantic search engines lıke Wolfram Alpha and Yauba seem to understand what AIDS is quite easily
Wolfram: http://www.wolframalpha.com/input/?i=aids
Yauba: http://www.yauba.com/?vid=l8510317539I1251926714&q=aids&target=all
According to Yauba:
Acquired immune deficiency syndrome or acquired immunodeficiency syndrome (AIDS) is a disease of the human immune system caused by the human immunodeficiency virus (HIV).
This transmission can involve anal, vaginal or oral sex, blood transfusion, contaminated hypodermic needles, exchange between mother and baby during pregnancy, childbirth, breastfeeding or other exposure to one of the above bodily fluids.
According to Wolfram:
AIDS and HIV
World mortality totals:
number of deaths 2.917 million deaths per year
cause of death probability 1 in 20 » 5.1%
rate of death 47 deaths per 100000 persons per year
DALY 84.46 million life years lost per year
As covered in yesterday's ReadWriteWeb article, we launched a microsite - healthbase.netbase.com - with the intention of publically demonstrating a new kind of semantic search technology that actually reads web content and delivers more relevant answers to health-related queries. healthBase is built on our Content Intelligence Platform that has been deployed successfully in different domains by Fortune 1000 companies, global publishers, and the federal government over the last few years. A ready-for-primetime consumer search engine it is not.
It is a powerful and automated technology that, when applied to something as messy as the Web, will produce some amazing results, but also sometimes some strange, funny and irrelevant ones. As noted in the comment section here, our first release of healthBase surfaced a few embarrassing and offensive bugs. These were far in the minority of results but enough to keep us up late improving the site. We sincerely regret and apologize in particular for any offense caused.
We've learned a lot in the last 24 hours and are fully committed to do better in providing an effective and accurate demonstration of our technology. We appreciate the feedback so please keep telling us what you think.
Thanks,
Jens Tellefsen, VP of marketing and product strategy & The Netbase Team
Sometimes it may be useful.
There is a very strong technology behind this, but some distinct and fixable problems led to the problematic results that have been widely cited. Those problems, and their solutions, here:
http://nlpconfidential.blogspot.com/2009/09/medicine-for-healthbase.html
Great work! Very amusing. I wonder if they found Semantic Web in PubMed or Mayo Clinic…
I did more mundane things like find pros and cons of resveratrol or aloe vera, which gives decent answers. HealthBase seems a bit like PowerSet for the medical vertical – a good idea. Probably more people will find it immediately useful.