Wednesday, November 5, 2014

ISWC2014 Trip Report

A few highlights from five intensive days at the International Semantic Web Conference (ISWC2014) in lovely Riva del Garda. See also my previous blog post Preparing for ISWC2014 and my live blog from all five days using Storify.

ISWC2014 Storify

Strong industry presence

ISWC is a research focused conference. However, this year it had a strong industry prescence with a full day Industry track, Semantic Developer workshop and many of the Lighning Talks came from industry. It was great to meet Business Analysts and Information Architects from large companies such as Roche, Genentech and NXP Semiconductors and also from small companies such as the Danish StatGroup.
  • All five Information Architects in the Data Standards Office at Roche / Genentech attended all five days to learn more about latest in semantic web research, especially traceability and provenance. Frederik Malfait, working for Roche and FDA/Phuse, described their RDF implementations of clinical trial data standards is the basis for a model driven architecture enabling computable protocols, component based authoring and automation of setting up clinical trial databases and generating submission datasets.
  • Marc Andersen, one of the two founders of StatGroup, presented the experience of the Pharmaceutical Users Software Exchange (PhUSE) developing a semantic representation of statistical results based on RDF and OWL. Providing clinical trial results as linked data will facilitate traceability, data sharing and integration, data mining and meta-analysis benefiting industry, regulatory authorities and the general public.
  • A business analyst described how NXP Semiconductor is making use of Semantic Web technology such as RDF and SPARQL to manage a product taxonomy for marketing purposes that forms the key navigation of the NXP website. 

Hot topics: Developer friendly, Linked Data Fragments, Provenance and Semantics for Sensors

  • The Semantic Developer Workshop and the conference program included many examples of RDF and SPARQL support in traditional programming languages, such as Java, Perl, C# and Javascript, as well as in data science languages, such as Python and R. The Semantic Developer of the Year, Kjetil Kernsmo, from Oslo University, presented RDF/Linked Data for Perl. JSON-LD was refereed to as the the developer-friendly serialization of RDF.
  • Many of the presentations described how they applied the Provenance standard from W3C for "information about entities, activities, and people involved in producing a piece of data or thing, which can be used to form assessments about its quality, reliability or trustworthiness." One example was how the standards had been used event based traceability in pharmaceutical supply chains via automated generation of linked pedigrees.
  • Semantics for Sensors for every-thing from smart building diagnostic,  traceability  in pharmaceuticals supply chain, and traffic diagnosis to predicting frost in vineyards on Tasmania.
  • "Everyone" talked about the work presented on the best awarded poster: Linked Data Fragments "so light-weight that even a Raspberry Pi can publish DBpedia (Wikipedia structured content) with high availability" 

Best workshop paper award

It was very nice to present our joint EHR4CR, Open PHACTS, SALUS and W3C HCLS paper. It got a best paper award in the pre-conference workshop: Context, Interpretation and Meaning for the Semantic Web.

Other ISWC2014 reports

Tuesday, October 28, 2014

RDF as a Universal Healthcare Exchange Language

Here's a short post about a nice webinar serie: Yosemite Manifesto proposing RDF as a Universal Healthcare Exchange Language. It is provided by  Semantic Technology & Business (@semanticweb).

Here are a couple of tweets I posted during Part 1 (video and slides) with David Booth.

The Yosemite manifesto has been critized. I recommend a "very civil discussion, in the face of clear disagreement" between David Booth, Thomas Beale (@wolands_cat) and Dean Allemang (@WorkingOntology): RDF for universal health data exchange? Correcting some basic misconceptions…

I look forward to Part 2, Friday 7 November evening (8pm CET), when Conor Dowling, Caregraph will talk about: "Lab tests and results have many dimensions from substances measured to timing to the condition of a patient. This presentation will show how RDF is the best medium to fully capture this highly nuanced data."

Tuesday, October 14, 2014

Preparing for ISWC2014

Next week I’ll have the great pleasure to attend my first Int. Semantic Web Conference (ISWC). I've been  fascinated by the power of the semantic web stack of standards for many years. Standards all based on the RDF model to represent and link data, as well as schemas, models and terminologies. I heard Tim Berners-Lee talk about the Semantic Web for the first time at the WWW8 conference back in 1999 in Toronto, Canada.

13th International Semantic Web Conference, ISWC 2014,
will take place in Riva del Garda, Terentino, Italy

At ISWC2014 I’ll present a follow-up paper to the one I presented in early September at the Medical Informatics Europe conference (see my earlier blog post: Preparing for MIE2014). It is a joint paper with colleagus from IMI EHR4CR, Open PHACTS, FP7 SALUS, W3C HCLS. Now with more details on the use of nanopubs and linksets in A Justification-based Semantic Framework for Representing, Evaluating and Utilizing Terminology Mappings. It will be discussed on Sunday in a pre-conference workshop organized by Alasdair Gray (@gray_alasdair), Paul Groth (@pgroth) et al. Workshop on Context, Interpretation and Meaning

I'm also also looking forward to participate in the Semantic Statistics workshop to learn more about things like the Data Cube for statistical data. This is a highly relevant topic for FDA/PhUSE and CDISC as representing clinical trial analysis results data as RDF Data Cubes is a topic at the ongoing PhUSE conference (presentation from UCB) and at the upcoming CDISC Interchange (presentation from DIcore Group, LLC, SAS Data Submission Consulting Services).

Tuesday evening I’ve been asked by James Malone (@jamesmalone) from EBI to sit on a panel at the European Ontology Network (EUON) Town Hall meeting together with Mark Musen, Stanford Center for Biomedical Informatics Research (I’m star strucked ;-) (Started earlier this year at the 1st EUON workshop in Amsterdam.)

I’m looking forward to meet interesting people from the semantic web community, and also newcombers to the community from organizations such as WHO, Roche, and CDISC. Magnus Wallberg from Uppsala Monitoring Centre WHO, working on a API and  RDF project for Global ICSR statistics. Also from the FDA/PhUSE Semantic Technology project there will be presentations form Landen Bain, CDISC, and Frederik Malfait, Roche, will present at the Industry Track on Wednesday.

I’ll try to keep my Storify-ISWC2014 updated during the week with interesting tweets, links and notes. And I know the Twitter-tag #iswc2014 feed will be lively as I’ve followed several earlier ISWC conferences on a distance.  

Monday, August 25, 2014

Preparing for MIE2014

After a fantastic warm and sunny summer here in Sweden it's time for me to get prepared for the European Medical Informatics Conference - MIE2014, Istanbul, 31 Aug. to 3 Sept.

Our joint paper co-authored by members across the IMI EHR4CR, Open PHACTS, SALUS projects and W3C HCLS community describing "A Framework for Evaluating and Utilizing Medical Terminology Mappings" has been accepted. And I have got the opportunity to present it in the main conference on the 2th September. 

For me the paper started from some great discussions at ICBO (Int. Conference Biomedical Ontologies) in Montreal last year with Trish Whetzel (@TrishWhetzel) and Jim McCusker (@jpmccu) on the topic: "mappings are not sufficient - need the justifications for the mappings". We started to talk about using so called Nanopublications to capture the justification for the mapping for users to make better use of for example the mappings provided via the NCBO Bioportal.

When I came back from the ICBO conference I wrote a blog post outlining some more ideas on using Nanopublications and/or Linksets, both stemming from the IMI Open PHACTS project. Some nice comments and sharing of my blog post: Justifications of Mappings encourage me to work more on these ideas. My colleague in the EHR4CR project, Sajjad Hussain (+Sajjad Hussain), pointed me to a very interesting blog post: SALUS project on Terminology Mappings. After some great discussion over a lunch at the SWAT4LS conference in Edinburgh with Hong Sun, from SALUS, Charlie Mead and Eric Prud'hommeaux, from W3C HCLS, Alasdair Gray (@gray_alasdair) from Open PHACTS, and many more, Sajjad and I started to outline a paper decribing a framework combining solutions and ideas on evaluating and utilizing terminology mappings.

Beside presenting this paper I look forward to participate in an MIE2014 tutorial and workshop:
  • Tutorial on the IEEE 11073 Standards for Personal Health Devices (Wikipedia: ISO/IEEE_11073). This is a standard I have been looking into earlier. It nicely combines my interest in clinical trials and health care data standards together with my previous industrial PhD studies in Mobile Informatics (see the slides presenting my PhLic thesis from 2001: Mobile Newsmaking).
  • Workshop on Interoperability Challenges for enabling secondary use of Electronic Health Records — ICEH 2014 In this workshop I look foward to meet and talk with many including the great metadata and ontology experts Gokce Laleci Ertukmen and Anil Pacaci (@aasinaci), Software Research, Development and Consultancy, Turkey.

I hope to be able to use my Twitter (@kerfors) feed to share interesting things I learn about in the conference, and from the historic city of Istanbul. And gather tweets, links, photos from each day using Storify. In the same way as I have done from eralier conferences. 

So, have a look at my MIE2014 Storify for daily updates 31 Aug. to 3 Sept.

Friday, June 13, 2014

openFDA a Game Changer?

I’ve been fascinated by innovative people in the FDA organization since I had the pleasure to meet Dr Norman Stockbridge, the father of FDA’s Janus datawarehouse model, F2F back in 2005 in Washington, DC. 

So when I saw some early notes about an openFDA initiative in June 2013 and early 2014 I posted a couple of tweets.

In April I wrote a short blog post about openFDA. And, when I saw how the new Chief Health Informatics Officer at FDA, Taha Kass-Hout (@DrTaha_FDA) started to count down on Twitter a couple of weeks ago I got really excited. It was nice to follow the #hdpalooz feed on Twitter from the health care data event in early June when openFDA was launched.

And, also to see services that directly were picking up the first openFDA API and launced services and apps to search the 3.4 million adverse events, such as Research AE

For a brilliant intro to what sits behind the first openFDA API I recommend Alex Howard's (@digiphile) excellent article: openFDA launches open data platform for consumer protection openFDA launches open data platform for consumer protection.
"Instead of contracting with a huge systems integrator.. FDA worked with a tiny data science startup.. to harmonize the data, create a cutting-edge website, and write and release open source code for a data publishing platform for it [on GitHub]"

I think this will be a game changer for how we think about open data, open source and open communities in industry. And yes, I do think we will soon will see much more Open, and Linked Data from FDA, and hopefully also from EMA and across industry.

Kudos to the devlopers behind all of this great work,
e.g. Sean Herron (@seanherron) and  Brian Norris (@Geek_Nurse)

Wednesday, April 2, 2014


It's exiciting to see how the FDA (Food and Drug Administration) now starts to make some nice buzz about their new project called openFDA:  A research project to provide open APIs, raw data downloads, documentation and examples, and a developer community for an important collection of FDA public datasets.

Excellent blog post from Dr. Taha Kass-Hout (@DrTaha_FDA), Chief Health Informatics Officer of FDA. He writes: "Our initial pilot project will cover a number of datasets from various areas within FDA, defined into three broad focus areas: Adverse Events, Product Recalls, and Product Labeling."

I do hope that the idea of not only open, but also linked data, will be part of this effort. For a quick intro to Why Linked Data? check out this nice video explaining the utility of linked data and how its being used by the UK's Ordnance Survey.

I don't have the full context to all of this, but I may think there are some excellent opportunties for Dr Kass-Hout and his team to leverage linked data intitative such as these:

Thursday, February 27, 2014

Why I am so obsessed with this Semantic Web thing

In an earlier blog post I reflected on the fact that it is now 25 years since the web was born. I had the opportunity to bring web technology into a large organisation. Many colleagues asked Why are you so obsessed by this "Web thing"? (remember that this was the time when a Swedish minister said that "Internet är bara en fluga").

So, now in 2014 many ask me Why are you so obsessed with this "Semantic Web thing"?.

I had a good chance to reflect on this question when I was asked to be one of the keynote speaker at a very nice conference: SWAT4LS, Semantic Web Applications and Tools for Life Science, in Edinburgh. 

I was also interviewed together with other speakers by the eCancer organisation in relation  to the EURECA (Enabling information re-Use by linking clinical REsearch and Care) project, Always scary to see, and hear yourself, but I think I managed to convey some of my thoughts. And it is really nice to watch the interviews with Frank van Harmelen,Eric Prud'hommeaux, Robert Stevens and David Kerr.

However, I think the one that best expressed the answer to the question was Charlie Mead. Charlie has been around in a long time in the standard world, working with HL7 for health care data and CDISC for clinical research data. Charlie is now a co-chair of the W3C interest group for semantic web for health care and life sciences (HCLS). I recommend this 7 minutes interview with Charlie. Below I have transcribed the last part of it as I think Charlie well express the reasons for Why I'm so obsessed by this "Semantic Web thing".

Charlie Mead
W3C HCLS semantic web interest group
"The thing that is really astonishing about the semantic web, the tools and technologies, really solve all of the core problems that we struggled with for a very long time. 
And they solve them in a very elegant way, which almost by magic, that live on top of the Internet that we now works and have brought tremendous value. 
And I think the real barrier to adopt these technologies is that is if more people understood what they can do I think the change curve will come faster and the resistance would melt more quickly."

Kudos to Scott Marshall, W3C and EURECA project, (@mscottmarshall)
for arranging the interviews and to the eCancer TV team.

Thursday, February 6, 2014

Why are you so obsessed with this Semantic Web thing

A lot of nice buzz today in sociala media when Tim Berners-Lee discusses the future of the web in the March issue of Wired UK. The web turns 25 years in March.

It reminded me of what collegues asked me almost 20 years ago: Why are you so obsessed with this "Web thing"??

Thanks to some great people in the Volvo business and data organisations I was exposed to "this web thing" and it made me change direction in my professional carrier. From a fancy job as Account Mananger to leading a small network of people get the Volvo Web Wave moving.

Today, 2014, my collegues ask me: Why are you so obsessed with this "Semantic Web thing"?

Recently I, together with other speakers at the SWAT4LS (Application and Tools in Semantic Web for Health Care and Life Sciences) conference, had the opportunity to reflect on the main difference the semantic web can make for patients, health care and clinical research professionals in video interviews by for the EURECA project. Stay tuned for these via my Twitter (@kerfors) feed and in a coming blog post.