AI Congress London 2018 Day 1

AI Congress (not @jack_septic_eye – I feel I may be in a very small subset of AI professionals who get that…)

London is a hive of AI activity. The UK is positioning itself as a leader in AI technology and you can barely walk around London without passing an AI company or meetup or training course1. If I didn’t actually have a day job, I could fill my time with AI conferences without actually doing much more than my daily commute. That said I am quite picky about the ones I go to. I’d never been to the AI Congress before and liked the diverse set of speakers and topics.  I was lucky that the team at Logikk had invited me as their guest for the two days. So how did it stack up? Well, day 1 was at a much higher level than some of the other conferences I’ve been to, with a lot of implementation and enterprise discussions and far fewer talks on the technical implementations. If you’re senior then these conferences are for you. If you want someone to talk about their latest paper on arxiv then there are far more technical events that will suit you better.

One of the biggest problems I had was that there were three separate tracks and only one of me, so if I didn’t make notes on a particular talk then hopefully the slides will be available after the event at some point. I missed some of the high profile talks, in preference of other speakers, on purpose as I’d already heard those speakers at other events. Continue reading AI Congress London 2018 Day 1

AI better than humans at reading?

News that AI had beat humans in reading spawned a lot of articles.

I’ve taken longer than I normally would to respond to some recent news stories about AI “outperforming humans” in reading comprehension “for the first time”.  Partly because I can’t help the wave of annoyance that fills me when I see articles so obviously designed to instil panic and/or awe in the reader without any detail, but also because I feel it’s important to do some primary research before refuting anything1.  The initial story broke that an AI created by Alibaba had met2 the human threshold in the Stanford Question Answering Dataset (SQuAD) followed closely by Microsoft outperforming Alibaba and exceeding the human score (slightly). Always a safe bet for sensationalism, mainstream media pounced on the results to announce millions of jobs are at risk….  So what’s really going on? Continue reading AI better than humans at reading?

Fooling AI and transparency – testing and honesty is critical

The effect of digitally changing an image on a classifier. Can you tell the difference between the pictures? Image from Brendel et al 2017.

If you follow my posts on AI (here and on other sites) then you’ll know that I’m a big believer on ensuring that AI models are thoroughly tested and that their accuracy, precision and recall are clearly identified.  Indeed, my submission to the Science and Technology select committee earlier this year highlighted this need, even though the algorithms themselves may never be transparent.  It was not a surprise in the slightest that a paper has been released on tricking “black box” commercial AI into misclassification with minimal effort. Continue reading Fooling AI and transparency – testing and honesty is critical

Review: Oxenfree

Your Oxenfree blue-haired alter-ego, Alex

It’s not often I get chance to play computer games anymore. The precious few hours I get to myself most days are usually filled with trying to catch up on my OU study1. So when I do choose a game, it has to be something immersive, where I make a difference. I loved Fable when it first came out and having finished Fable 3 recently (well behind the rest of the community I know!) I was looking for something new. A game I could play for ten minutes or an hour, something that I could be absorbed into and something where it didn’t matter if I didn’t play it for a couple of weeks. My wonderful husband bought Oxenfree for me on the XBox (also available on Steam) and I was hooked from the beginning. Continue reading Review: Oxenfree

Cozmo – a good present?

Cozmo – image from Anki

One of the toys that’s been advertised heavily in the UK this year for Christmas has been Cozmo with it’s “Big Brain, Bigger Personality” strapline. I got one last year and it was a great present. Let’s get this out there. Cozmo is relatively expensive. For about £1501 there are a lot of other things you might prefer to buy for a child (or an adult) for what is, on the surface “just a toy”. If you treat it as such then maybe it’s not the right thing for you, but viewing Cozmo as a simple toy is far less than he deserves. He is a lot of fun to play with, and the more you play with him, the more he begins to do. Continue reading Cozmo – a good present?

Credit where it’s due in AI – capsule networks

In the past few weeks my social feeds have been littered with articles citing “Hinton’s latest breakthrough” in AI: capsule networks.  Like most people in the field, I make sure I read up on what’s new, and I’m yet to see the paper‘s first author Sara Sabour, get credit for her work in all of the tertiary reviews.

A somewhat irreverent take on author list from PHD Comics, although this is broadly true 😉

For those who aren’t in academia, there is a distinct order to the names on published papers either by contribution or alphabetically.  For contribution, the first author is the one who actually did the research, the last author is the person who runs the lab/department and any other names are listed in order of contribution.  Occasionally you will see notes that authors contributed equally.  Some subject or countries list names alphabetically, but this is not the case for this paper published on arxiv. Continue reading Credit where it’s due in AI – capsule networks

When did AI not being as good as humans be a news item?

StarCraft 2 is a big thing in e-sports – can AI live up to the human players?

I get very tired of the clickbaity journalism hyping up minor advances in AI, making news stories out of nothing or even the ones for those in the industry. You know the type: “Facebook AI had to be shut down”, “Google creates self learning AI”.

I demystify a lot of these when I’m asked about them – technology should be accessible and understandable and I deplore the tendency of those seeking to get article hits by over-egging with misleading headlines. What amused me over the weekend was that an AI not beating a human was a news story where the AI was “trounced”. Continue reading When did AI not being as good as humans be a news item?

Being a Panellist at a Tech Talk

Being a panellist – expect to be recorded

Following on from the meetup talk I gave on building your own personal brand, I’ve been asked a few questions about speaking at events and being a panellist.  I felt it would be a good idea to write a few posts on this for reference.  This one is particularly about panel sessions. Continue reading Being a Panellist at a Tech Talk

The Oxbridge Myth

Photo of a typical oxbridge college
My College, St Peter’s Oxford. Photo from Wikimedia

There’s been a lot in the news recently about how Oxford and Cambridge are failing young people who are educated by the state system and how they perpetuate the elitist machine that runs the UK.  As ever, I’m frustrated with the polarisation of the argument perpetuated by the media, which boils down to “it’s fine” and “everything is broken”, which no room for focussing on the nuances of the problems.  As a state school student who went to Oxford1, I thought I’d weigh in with my own experiences and where I believe the issues are. Continue reading The Oxbridge Myth

M337: Group theory becomes relevant

My M337 books are starting to look a little dog-eared after a few weeks of commuting!

One of the things I’ve enjoyed least about my OU Maths journey so far has been group theory.  I ploughed through whole swathes of M208 applying the techniques and not really seeing the relevance1  I found group theory and the proofs related to it tedious.  Mainly because I was proving something that was “obvious”.  However, I’ve always had a healthy acceptance of partial learnings – knowing that if I was being taught a technique then there was a reason for it. Two years later and that reason finally hit me. Continue reading M337: Group theory becomes relevant