Daniel’s Blog

I’ve started using the following setting for my university report writings, and I highly recommend it!:

• For finding articles and books: Google Books, Google Scholar and CiteseerX
• Zotero, a firefox plugin for Bibliography Management. Which includes the ability to sponge bibliographic data from all of the above and many of the article repositories on the web (and Amazon.co.uk books category ).
• BibTeX, which is a bibliographic format that can be exported from Zotero.
• LaTeX, which is a document markup language which is reasonably similar (although in some ways simpler, in other ways more complex than) HTML. LaTeX can also work with BibTeX, by using unique identifiers, LaTeX has the ability to give a code, pull data out of the BibTeX file and provide precise styling for the document that you are referencing.

Absolutely amazing. I highly suggest using all of the above. I also use something called Kile which is an “Integrated LaTeX Environment” for Linux (it was made specifically for KDE, but I use it through Gnome).

I’ve dabbled with LaTeX in the past, but it just didn’t really work for me until recently. So please, if you do write articles, then try out LaTeX if you haven’t already!

UPDATE - Friday 20th February 2009

I’ve given up on Kile, as it doesn’t seem to work on my Laptop. I’m now using Texmaker. Which is cross-platform, and seems a lot more responsive than Kile.

I was wondering about the Presidents in the United States of America, and wondered what their religions were. So I zipped up some SPARQL queries to run against DBpedia, they aren’t perfect (i.e. they don’t capture all of the presidents due to mislabeling and me not being too bothered about the answers), but it provides some answers.

The Democrats:

SELECT ?PresidentName ?ReligionName
WHERE
{
?president
dbpprop:party <https://dbpedia.org/resource/Democratic_Party_%28United_States%29>;
dbpprop:order dbpedia:President_of_the_United_States;
<https://dbpedia.org/ontology/religion> ?Religion;
rdfs:label ?PresidentName.
FILTER (lang(?PresidentName) = "" || langMatches(lang(?PresidentName), "en")).
?Religion rdfs:label ?ReligionName.
FILTER (lang(?ReligionName) = "" || langMatches(lang(?ReligionName), "en"))
} ORDER BY ?ReligionName

Which the answer is viewable “Democratic Presidents Religions”

The Republicans

SELECT ?PresidentName ?ReligionName
WHERE
{
?president
dbpprop:party <https://dbpedia.org/resource/Republican_Party_%28United_States%29>;
dbpprop:order dbpedia:President_of_the_United_States;
<https://dbpedia.org/ontology/religion> ?Religion;
rdfs:label ?PresidentName.
FILTER (lang(?PresidentName) = "" || langMatches(lang(?PresidentName), "en")).
?Religion rdfs:label ?ReligionName.
FILTER (lang(?ReligionName) = "" || langMatches(lang(?ReligionName), "en"))
} ORDER BY ?Religion

Which the answer is available directly here: “Republican Presidents religions”

Please note

• It doesn’t return all presidents of the parties
• It does use the DBPedia SPARQL engine (powered by OpenLink Virtuoso)
• It does use some shortcuts (i.e. not defining PREFIXs and the FROM clause), as the DBPedia SPARQL engine provides these nice shortcuts.

It’s absolutely fantastic that more and more people are taking up Semantic Web solutions and providing RDF.

Guys and Girls it’s time for the next step!

This is something that I’ve been thinking about for a while. People, although they are taking up the SemWeb, they don’t seem to understand the full power of it’s expressiveness. This isn’t a providing problem, this is a usage problem…. Yes it’s fantastic that we can interlink objects now, but what does it mean to be an object? What can we learn implicitly and explicitly from this highly-structured and highly-distributed web of data objects? and finally, what can we do with what we learn?

These questions are the real heart and real soul of the Semantic Web, and an area in which us Semantic Web advocates have really mentioned much before.

And please note, this shouldn’t *just* be an academic thing. Systems which embrace this, and which are in the wild, would greatly benefit. It’s breaking down the walled-garden and combining the intelligent systems which have previously been in websites like amazon.co.uk and google, and exposing those methods to the real semantic data.

The key to all of this is improvement of society. If your project improves society (or helps the environment in some way), then your project should be excellent. Whether or not it makes money!

Hope this all makes sense…. feel free to ask questions, or give your opinion.

Daniel

Rumours have it, that there will be another Bristol Knowledge Unconference.

The last one was a success… a good turn out, some interesting talks and discussions. So it’s only right to do another one.

This one coming up doesn’t have a date yet, or a location. It will, unlike the previous, have a “theme” which I’ll be trying to get speakers for…. and of course, it’ll be in Bristol somewhere.

So… I need some helpers. Please do get in contact with me if you can:

• help me find a location
• help do some general organisational/administrative assistance
• offer to do a talk about knowledge, set to a specific them
• give some money to pay for food and drink

I will strive to make this Unconference completely free for attendees, which does mean that we require the location to be free (or sponsored) and the food to be supplied (or paid for) by some organisation(s)/people.

So please do get in contact if you can help in any way by sending me an email ( danieljohnlewis at gmail dot com ). Else, I’ll keep you all up to date through my blog.

Thank you,

Daniel Lewis

Here are some predictions for Computing & IT in 2009:

• Hardware
  • More and more Cloud Computing services (and improvements)
  • Personal Supercomputer Hardware (such as NVidia Tesla)
• Software
  • Advancement of ease-of-use in Operating Systems, particularly Linux… which will see more of a take up this year than ever before.
  • Browsers (particularly Mozilla-based and Opera) will start to contain more and more Semantic Web based features.
• Web
  • More people and organisations providing Semantic Web services. Leaders in this field will help greatly, such as the BBC, BT and the UK Government.
  • The original developers and advocates of the Semantic Web, will start to move away from it a little in an attempt to improve the semantics of the Semantic Web. Areas such as Fuzzy Logic, Bayesian Probability, Advanced Description Logic, Intelligent Agents and Artificial Neural Networks will be bridged with the Semantic Web by researchers.
  • Web development will become more and more like desktop application development.
  • Web design will become more focused on providing the data rather than providing pictures, fusing with Information Architecture.

BUT! More importantly, things may seem to slow down in terms of development within the computing field. This won’t be entirely the case, we’ll see more efficiency improvements than feature additions… and these developments will probably help the efficiency of those companies using that software/hardware.

Although I’ve mentioned all of this. Please note that I truly feel that we should move away from a materialistic view to a more societal and/or spiritual point of view. This means that although new things may be released in the future, we should recognise them, but not be tempted by them if they are not necessary updates.

Right, I’ve been doing a lot with Genetic Algorithms (I’m actually implementing some using the Ruby Programming Language). However, the masters unit which covers G.A.s is quite theoretical, and aimed at mathematically improving the evolution process. So, I’m sending this message out:

Wanted: Links (e.g. Hyperlinks or Articles/Documentation/Papers) of Real World Applications which implement Genetic Algorithms. Particularly if they are engineering based rather than biological modelling based, but I’ll look at either. Source code (or even pseudocode) of the GA would be a huge bonus but not essential.

Please comment on this post or send me an email, I’d love to hear about it!

Cheers,

Daniel

Well, well, well… you may have noticed that I haven’t made a blog post since the one about the diary.

Well, I’ve just started the Masters (MSc in Machine Learning and Data Mining) at the University of Bristol. I had an intro week last week, and started attending lectures and other events yesterday.

I am really really enjoying it, the only one thing that I am a bit worried about is that it’s going to be taking up a lot more time than I realised. Which may affect me earning some money, unless I work on things which are similar enough to the stuff that I am working on at University. We shall see, it is only the first week, and things shall probably calm down a little.

I am particularly enjoying the biological aspect of the course (e.g. Genetic Algorithms/Programming and Computational Genomics and Bioinformatics), and I’m looking forward to touching the psychological aspects also.

This is actually going to sound quite weird coming from someone who titles himself “Technology Evangelist”, but I really don’t like using technology… computers (hardware and software) are so awkward to use, mobile phones never have enough signal and it’s all so reliant on electricity. The one thing that I truly feel about the whole subject of Machine Learning and Data Mining is that it is supposed to make things easier! In order for technology to become really useful we have to look towards some things that we have outside of technology:

• High/Deep Semantics and Art
• Biology, Psychology and other Sciences
• Ubiquity/Omnipresence
• Humanities and Society

Maybe the seven liberal arts can also help out:

1. Grammar
2. Rhetoric
3. Logic
4. Geometry
5. Arithmetic
6. Music
7. Astronomy

I do not believe that we can continue with technology in it’s current state. As for the Semantic Web, I do believe it is a step in the right direction, but it is only a small piece of the future puzzle.

For the first time in my life I have started using a paper-based diary (two days to a page, academic year diary for 2008 - 2009).

It is so liberating!!!! For the first time for a while I actually feel like I’m organised! It’s great, I’m able to record all my forthcoming events:

• University events
• Random geeky meetings in Bristol that I’d like to attend
• Hospital appointments
• Some other important events which are coming up
• Work stuff

It’s fantastic!

You see, this is the interesting thing… no matter how far we advance technology, it’s still easier to use a paper and pen for organising yourself. Maybe I should get one of those InfoScan pens and hook it up to ODS.

As one of my further education college tutors used to say back in 2002:

Organisation is the key to success

Yeah, you heard me right:

Linked Data is more important than the Large Hadron Collider

My points:

• I feel that the Large Hadron Collider is a bit of a waste of time, money and a big waste of energy (think of all the carbon emissions!).
  • Why do we need to know the very very smallest parts of things?
  • Why do we need to know what happened microseconds after the big bang?
  • Why not just analyse what happened if the Higgs Boson was and wasn’t found (every possible aspect)?
  • Face it: It’s not going to cure Cancer, or prevent HIV
  • Face it: we’re still going to be here whether or not the Large Hadron Collider was successful
• Linked Data on the other hand is incredibly important
  • Everything in real life (conceptual/abstract and objective/physical) is interconnected with everything else
  • We have a lot of data about everything, but it’s not interconnected
  • Linked Data allows for the interconnectedness of data, and therefore true computation modelling of everything. Which then allows for a real and useful insight into scientific (and artistic and historic) data!
  • Just imagine a world where you can easily browse through the history of the atom, and then delve into the science found on the atom, and then go deeper into the subatomic level, and then browse back out into the historic realm, finding out about experiments that happened and whether it had any impact on society.
  • Thanks to Linked Data:
    • a lot of problems can be solved before they arise
    • new areas of research can be formed
    • “recreating the wheel” can be prevented
    • a really interesting browsing experience can be achieved

So there we go. That is why Linked Data is more important than the Large Hadron Collider….

Sorry to the physics-geeks, but maybe you can gain something useful from Linked Data if you aren’t already. I know that some Chemistry and Physics people already have some Semantic Web / Linked Data research & development going on, and hey Tim Berners-Lee did start the Web project at CERN.

I fear that a lot of people within science and even computer science forget to look at things with Holistic eyes, as a lot of science and mathematics is incredibly Reductionist.

Now for an incredibly Formal Blog Post. For equality I thought that I should open this up to as many people as possible, I’ve been pondering about making this blog post for a while, and I have even discussed the project with a few of you. Receiving my introduction pack and registration guide this morning has prompted me to write this post.

Intro

Many of you know that, I will be starting postgraduate study at the University of Bristol from the end of this September (2008). This will be a one-year full-time MSc course with project, the subject area is “Machine Learning and Data Mining“.

I would like to open up my project idea to an organisation in return for sponsorship of the course. I only ask for the cost of the tuition (£3950 GBP, which is currently roughly $7711 USD / €4982 Euros / $8032 Canadian Dollars / $8428 Australian Dollars / 323385 Indian Rupees / 830725 Japanese Yen / 52861 Chinese Yuan / 8102 Swiss Francs). In return the sponsor would have some say in the project itself, in addition to logos and names on all relevant websites and in documentation (negotiable).

I would (potentially) be interested in taking the project further after the project has been completed, this could be in the form of a business idea or another academic project. This is also negotiable.

My Project Proposal

I do already have an idea for a project, and have even written a project proposal which has been approved by the department. However, the content of that project proposal is negotiable and I could even change it completely depending on my interests/skills, the criteria for approval and also the sponsors interests.

The idea I already have will be an entirely new way of turning Free-Text into Linked Data, with the specific use-cases being very humanities (including anthropology and religion) based. But as said before, this is negotiable.

Course Information

The postgraduate course is officially titled “MSc in Advanced Computing (Machine Learning and Data Mining)”, it is a taught course but contains a full academic project. Topics within the course include Logic Programming (using Prolog), Bayesian methods, Natural Language Processing, Genetic Algorithms/programming, Neural Networks, Fuzzy Logic, Reinforcement Learning and some Bioinformatics content.

The official programming languages used at the University of Bristol include: C, Java, Haskell, Prolog, Progol… etc

Systems/Frameworks include: Weka, MatLab

A little about me

If you have just stumbled across this blog post, then you’ll be wondering about who I am. So a brief biographical summary:

I currently work as a Technology Evangelist for OpenLink Software dealing with Semantic Web/Linked Data technologies in addition to data portability, social web and database technologies. In May 2007 I completed my undergraduate degree with a 2:1 honours and a Double A grade for my final year project titled “Adding Semantics to Social Web Tagging Systems”, the undergraduate degree was a “BSc(Hons) in Intelligent Systems and Software Engineering” at Oxford Brookes University. My computer science interests are in the “Evolution of the Web” (so Semantic Web/Linked Data, the Social Web, Web 3.0, WUPnP and the Intelligent Web), Open Source Software and Open Data, Programming Languages and finally Intelligent Systems (mainly Logic Programming, Machine Learning, Data Mining, Agent Technology and Knowledge-Bases). I have experience in many programming/query languages, but some of my favourites include: Ruby, Java, Haskell, Prolog and SPARQL. My interests outside of computer science include: psychology, philosophy, music, religion and esoterica. I am a “Liberal Catholic Anglican” Christian and active/open-to/passionate-for Interfaith/Interreligious dialogue. I am a full member of the British Computer Society (BCS) and the Association of Computing Machinery (ACM). There is more about me on my simple homesite, my blog and my linkedin account, other information (including a CV) can be provided on request.

Important Dates

• Autumn Term: Monday 29th September 2008 to Friday 12th December 2008
• Spring Term: Friday 9th January 2009 to Friday 20th March 2009
• Summer Term: Monday 20th April 2009 to Friday 19th June 2009

I must have details about sponsorship as soon as possible, the final date for setting sponsorship up is Friday 12th September 2008…. and I really don’t want to leave it until then.

Useful (Official) Links

• Machine Learning and Data Mining Unit Descriptions
• University of Bristol and the University of Bristol’s Computer Science Department

If you are interested

If you are interested then please do let me know. I’m available by email or jabber… we can converse by phone or skype… and we can even meet up in/near the wonderful city of Bristol (UK).

• Email: danieljohnlewis [-at-] gmail [-dot-] com
• Jabber/GTalk: [email protected]
• Skype: daniel.lewis
• Phone: +447834355516 (UK specific: 07834355516)
• Current Location: Clifton, Bristol, England, UK (it’s in the South-West, about a 2 hour train journey from London)
• Current Time Zone: British Summer Time (GMT+1)

I will announce

I will announce on my blog when I receive funding. So if you don’t hear, then please do assume that I have not found any yet.

Many many thanks for reading, and I hope to hear from interested people soon.

Cheers,

Daniel