Tuesday, September 28, 2010

Web science: A new frontier

so this two day meeting at royal society brought together  a range of different people across the whole web social ecosystem, videos and tweet feeds here:

so yesterday and today i have been at

you can watch the talks online too i believe... but dont necessarily
watch the barabasi one, too boring he covered his 99 and 2000 papers
on power laws etc

there was some engineering talks (ipv6 and etc) and semantic web which
I didn't pay enough attention, so I focus on social networks ones)

overall, it was great, seems like people are now thinking what we
thought with Mia this time last year, how do we use social network data to
- Show quoted text -
answer long standing questions in social networks and how is the web
affecting the .. Kleinberg picked up exactly on that.. how do we see
what threshold (number of friends doing an action) is required for
people to adopt an idea (join a group on live journal, edit
wikipedia), so this is exactly what we started to touch on on the
twitter data but didn't actually get to follow up on.. I think it will
be amazing to test this in different contexts, start tweeting about
iphone, start spreading URLs and etc, there is definitely room for
exploring on that

there was also a relatively ok talk about social recommendation from
anne-marie kermarrec , it turns out that the large or small number of
usrs can both be bad for recommendation so they focus on distribution
of interests of a person with different size of user groups , some
interesting discussions of positive and negative sides of this
approach on twitter http://twitter.com/#search?q=%23RSwebsci

the notion of collective intelligence and decision making seemed to
also be very strong, but no one knows how to quantify it... on twitter
this would be amazing (maybe better in slashdot?!) but obviously Levy
would insist on that a bit too much! but he has this idea of having
semantic URL (USL = Universal Semantic Locator ) to enable extraction
of collective intelligence etc (philosophy!) but a neat idea

then manuel castells on socialnets in the net: media has given bad
image to web as they just report on bad/negative news (their nature),
but he thinks we are more social today since we have more practice for
sociability online and offline, he believes the internet enables people to
pursue happiness easier hence strong presence of less powerful
people... so inernet use has led to more autonomy, sociability and
happiness (based on their survey)..

Ramesh Jain gave  a decent talk on twitter and events, and why the
real time web is so important as it captures the events in life.. so
camera publishes to flickr/tweeter directly with location info.. the
location can also help with event detection and speed of propagation
of events, so you can for example detect activity rates of say swine
flu or demonstrations.. so a kind of heat plot if you can imagine...

finally tim berners-lee talking about social cloud and storage
(dbpedia for example) .. he mentions a big issue with social networks
being so "walled" so sharing things across is difficult. so better to
have webpages and URLs??!! also issues with privacy and control of
data (for example download all data).. overall not sure what he
wanted.. overall interesting though

Wednesday, September 1, 2010

Productivity at work

I enjoyed reading this http://matt.might.net/articles/productivity-tips-hints-hacks-tricks-for-grad-students-academics/

I think productivity is severly affected by trying to do too many projects, just as bad as by focusing only one a single task for a period longer than half a day.. a bit of writing or analysis while writing code for another simulation or measurement always helps, and even reviewing can be joyful when interlaced properly with related work.. anyway, have a read and see what you think.. it even mentions "Optimize transaction costs" !