Thursday, December 31, 2015

OWN-PT Logo, yay!

Zach Sanchez provided a beautiful logo for OpenWordNet-PT.

The colours shows that we start from Brazil, but want to encompass Portugal and the other Portuguese speaking countries. The network says that we want to be part of the big resources in the Semantic Web, more specifically of the Linked Open Data networks.  The shape  of the net should remind you of a lambda, as our implementation was basically in Lisp and we value our functional programming origins. The dashed indicates that we're not complete and we hope people will help to make it better!

Yes, I think we covered all the bases.

This year wasn't as good as 2014 for the work on lexical resources. We actually only had three (make it two) papers accepted. I must be getting worse at writing stuff, since there were 8 (yay!) in 2014. Maybe it's the problem with focusing getting worse (very interesting but full of distractions blog post here). May be it's just the ups and downs of getting stuff done and then writing about it, I don't know. Must make sure I don't get disheartened by all the stuff that didn't get anywhere.

Monday, December 14, 2015

Can you make stress your friend?

SAY is a hidden gem of Stanford.
and Psychology a fascinating and baffling subject.

Maybe you should watch this.

As she says, it may save your life.

How to make stress your friend

(picture of traffic lights by Gary Smith)

Saturday, December 12, 2015

Language nerd

This week I gave a talk to the Nuance Sunnyvale Research Lab on the work on OpenWordNet-PT that I'm doing with friends in Brazil, Alexandre, Livy, Fabricio, e Claudia, amongst others.

I talked about work that we have been doing since Sept 2014, when Fabricio joined us, and that I had decided  needed to be written up for a workshop in the summer.  The official venue, the ACL-LDL workshop apparently worked well and Alexandre gave the talk in Beijing. Semantic Web and Linked Data (Linguistic or not) are much more his cup of tea than mine.

The talk was ok (any talk given is better than a talk still to  give), but the audience was very focused on a problem that I do not have solutions for, yet. Anyways, giving talks is also about getting things clearer on your mind. Not only what you want to do next and how to do it, but also, a bit surprisingly, the things you've done already and what/why the panning the way they did, actually means to you. The slides are in slideshare (Seeing is Correcting) and I really need to get my act together and list all the talks,  the papers, the drafts,  the code that I have been working on organized.

Wednesday, November 25, 2015

Listening to Friends' Talks

So I've been organizing the seminars at work for a while now. I believe the foundation of both innovation and invention is actually knowing what people are doing and relating it to what other people are doing. Being a middleman, a peddlar of knowledge, as it were. Not very highbrow at all.
But it's very easy to forget all these talks and who gave them.

Given that the new Nuance Research website provides a list of the talks, I've decided to repeat the list here to add my own comments, when I have them. The official listing is in the Nuance Research webpage.

My personal listing is attached, it has most recent external talks. The links will not work, but they shouldn't really. (I will modify them for the appropriate external links, at some stage)

Thursday, November 19, 2015

Ford Smart Mobility: PARC Forum, 06/25/2015



Several months back, I went to the PARC Forum on Ford, the automobile and mobility company as they want us to call them. Dr Ken Washington gave an excellent talk, that was very different from what I was expecting.

I went expecting to hear about mobility, as in services through your mobile phone, but it was all about how people will commute and go to places in the next decades. And how Ford is hoping to make sure that it is still relevant in those decades.

Washington (vice-president of Research and Advanced Engineering at the Ford Motor Company)
started on what he called "threats to freedom of mobility",  which were actually threats to Ford's business model:
1. global urbanization
2. global middle class growth
3. air quality and other health risks
4. changing customer's attitudes (to car ownership in particular, millenials and later gens don't care for cars)

He talked about the research in the newer lab they have in Palo Alto (3200 Hill View), where they started around a year ago with some experiments on gauging customers pain points and interests.

These are apparently now transitioning to products and he described 5 of them.
1. Go!drive: city driving on demand, where you pay by the minute, an electric vehicle with designated and guaranteed parking places. They're testing it in London.
2. peer-2-peer car sharing, I think the name of the product is easyCarClub, they're launching in Chicago/London/Washington/San Francisco, partnership with a SF start-up, but I didn't catch their name.
3. they have been collecting data on bikes and cyclists since 2012
4. out of the experiment in 3, they're launching eBike, which integrates electric bikes and cars/trucks. 
Three modalities: modeMe ebike, modePro ebike and  modeFlex ebike (which he says is not launched yet, PARC heard it first). ModePro and modeFlex are sold with a transit van Connect. all are electric bikes, where you plug in your cell phone and that navigates you, so there are cell phones involved too, but they're not central.
5. Finally they're also getting on the business of finding parking places in big urban centers via Parking spotter, which detects empy parking slots using sensors in your car and tells other people about them, using the cloud. they have several partnerships in the Valley already, but he did not mention who.

Summarizing: Ford smart mobility focus on flexible use and ownership, multimodal styles of transportation for urban solutions mobility; car sharing/fractional  ownership/ pay-as-you-go.

Because with 9 billion people in the planet in 2050, 70% living in the big cities, (by 2030 they're predicting 41 mega-cities with more than 10 million people), all needing to go the same places and wanting to park, they will need help.

Q&A addressed  expected issues: fuel price, mass transportation, different contexts (Europe vs US vs China). I did not managed to ask my question about the voice interaction of humans and cars, as there were too many other people asking questions. But it was nice to see old friends like Johan de Kleer, Scott Elrod and Teresa Lunt.

Google on Ex-Machina

Ok, mostly keeping a copy of their paper on neural nets for conversational agents, to read it later.

 Google tests new artificial-intelligence chatbot
Several weeks ago, Google released a research paper chronicling one of its latest forays into artificial intelligence.
Researchers at the company programmed an advanced type of “chatbot” that learns how to respond in conversations based on examples from a training set of dialogue.
And the bot doesn’t just answer by spitting out canned answers in response to certain words; it can form new answers from new questions.
This means Google's researchers could get a little creative with it, and they certainly did — they asked the bot everything from boring IT questions to inquiries about the meaning of life.
The responses were alternately impressive, amusing, and unnerving.
Here are some of the most poetic interactions from the paper:
Human: What is the purpose of life?
      Machine: To serve the greater good.
Human: What is the purpose of living?
      Machine: To live forever.

Tuesday, November 17, 2015

The Future of Artificial Intelligence

Or so they say.
Fourth in a collection of quick posts, not much about anything in particular. 
Just shooting the breeze, as they say too.

I have a good friend that works for the Institute for the Future, but I have no idea how come I got to this blog post by A. Trabulsi. Anyways, if you're in a hurry, just look at this picture, that tries to answer the question: Where is the money going?

It's not exactly how I'd classify it, but hey, I'm not in the business of predicting the future. They are.