Eleni Koutsomitopoulou - Ελένη Κουτσομητοπούλου
  • Home
  • Class material
    • Course Syllabus
    • Course Calendar
    • Sample Lecture
    • Sample Quiz
    • Additional Course Syllabi >
      • Corpus Linguistics
      • Machine Learning
      • Statistical NLP
  • Blog
  • Contact

How (not) to interview people

4/24/2016

0 Comments

 
So, that was a pretty intense interview at one of the most successful start-up's in Cambridge. Two interviewers and one victim. A barrage of questions that all had one thing in common: they were very little related to the description of the job I had applied for. I was confident based on my past work experience and skills that I could excel in the job, greatly contribute to the company goals, and learn new things in the process, so my main focus during the interview was to give my interviewers the reasons behind this confidence.
However, it seemed I was being "punished" for some bad past experience they had, or they were uncharacteristically unreasonable, condescending and rude, no doubt taking "revenge" over some past unlucky investment on a not-so-worthy employee. Cannot be explained otherwise.
It was the only interview where I felt uncomfortable with a feeling almost like the interviewers "were out to get me", rather than find common ground and see if we are a good fit. 
Because when you interview an experienced professional, it's more about the fit with the company culture and the personalities of the hiring managers and the team, and less about testing their knowledge and certainly much less about denigrating their experience, skills and qualifications. 
My good manners prevented me from responding with a f*ck off in several points during the interview. They wanted to know why my PhD took 5 instead of 4 years, as if they haven't lived in the real world where people actually hold full-time jobs in parallel with completing a demanding PhD. They wanted to know how I would define X, even though X has nothing to do with the tech I would use on the job (and they admitted that when I asked). They were sarcastic about the patent I was involved in co-inventing, quickly dismissing it as of no importance. It was as if their goal was to actually shoot down my CV rather than understand where I am coming from.
It was good luck that they behaved this way so early on, because when I left that room, there was absolutely no doubt in my mind that I would be miserable working with them.
I just wish people were a bit smarter, and minded their manners, especially when we're talking about people in the positions of CEO and Head of Engineering. Because it's smart to leave the room with a good impression about you as you represent your company and your company's overall culture. It is smart to be well-mannered and considerate to the person in front of you. Because even when an experienced professional does not get to work for you, they will still have formed an opinion about you and the company you represent as a result of this encounter.
You see, an interview is not only for the interviewer, but also for the interviewee to judge the opposite party.  
0 Comments

Probability vs. Statistics: a basic mis-conception

3/20/2013

0 Comments

 
With the popularity of statistical NLP methods and probabilistic models of natural language I have found the distinction below is paramount in the heads of some linguists.

Probability is a theoretical branch of mathematics dealing with the prediction of the likelihood of *future* events, and therefore is useful for the evaluation of the consequences of mathematical definitions.

Statistics involves the analysis of the frequency of *past* events, it is data-driven and therefore an applied branch of mathematics and is useful for the analysis of events based on observation of cumulative data about them.

Some people (linguists) further confuse in the above discussion the distinction between data and algorithms in modern NLP. This is a whole other topic
also related to the Norvig-Chomsky debate I mentioned in a previous blog post. In a later post I will explain some further common misconceptions. 


0 Comments

What is Cognitive Science?

3/15/2013

0 Comments

 
Recently I came across information about a French MSc program in Cognitive Science, which educates computational linguists.
It is not just a European phenomenon. Many schools is the States also like to baptize their Computational Linguistics programs as programs in "Cognitive Science".
Makes me wonder: what is the exact meaning of "cognition" and what is "cognition" about? Since when did a computational science become synonymous to a cognitive science? Are we confusing a computational system (and subsequent methods) with a cognitive system or -come to think- with human cognition?
Implied under the term 'cognition' is of course the modifier "human". 9 out of 10 times when we speak of cognition, we refer to "human cognition" . Why? Simply because it is human to reason and have intelligence, which constitutes "cognition" and "cognitive capabilities".

Computational systems tend to get inspired by or try to emulate human cognition (i.e. the cognitive power of the human brain) but they are far from being cognitive systems. In that respect, studying Computational Linguistics is not the same like studying human cognition unless there is a focus on the brain sciences and neuroscience and the biological (neural) substrate of natural language.
0 Comments

A delayed mention of Aaron Swartz

1/23/2013

0 Comments

 
Aaron Swartz was only 26 when his dead body was found in his New York apartment on Friday January 11th 2013.
The video below is a a face-to-face interview that nicely portraits what Aaron Swartz  stood about.
It's hard to add something new to the world tribute about Swartz.
Perhaps you remember that back in 2011 he was charged for his downloading about 4 million academic journal articles via JSTOR, a service that provides students and researchers access to a limited number of articles free of charge.
Or, his role in preventing SOPA. Because of his obvious leadership in activism for internet freedom, many think that his hanging was not exactly accidental. It would be a sad sad world indeed if the rumors were true.
RIP Aaron.

0 Comments

Facebook's new graph search 

1/16/2013

0 Comments

 
It was high time Facebook would improve on its searching capability.
Searching within Facebook and in your network of friends and friends of friends was an overdue feature.
So far, Facebook search has meant searching in the rather narrow field of people names within Facebook but outside a specific user's network of friends. Web search results (delivered by Bing) were also available often completely unrelated to the targeted Facebook search. Most FB users when searching within FB, clearly don't need to search on the web or they would simply switch  to google or some other general-purpose search engine.
With FB's new "graph search", the search will be specific to the search user's network of friends (as far as the information has been shared with this specific user or with the public) and it won't only be about people's names.  Indexed are also places, photos and any other clearly categorized information people put on their FB pages.
Still in beta for US users.
Update [March 15th]: Graph Search is rolling out very slowly. Will come back with examples when I have access to it. In the meantime, enjoy this as an intro to the tech behind it.

0 Comments

Norvig vs. Chomsky on the future of computational linguistics

1/16/2013

0 Comments

 
Well, yeah, most people think this is about "the future of AI" but the specific branch of AI that deals with natural language processing is called Computational Linguistics.
When Norvig and Chomsky's voices met at last year's MIT's Brains, Minds, and Machines symposium we had a chance to finally witness out in the open a debate that has been brewing for the last decade or so in the field.
Norvig, the head of research at Google, represents the modern CL, which emphasizes what Chomsky called "a method of doing linguistic analysis without involving language and the linguistic science" (my paraphrasis).  Machine learning and statistical analyses of language are two well-known manifestations of such method.
The debate was inevitable.
Google is notorious for (successfully) doing natural language processing without special regard for natural language.
And Chomsky is notorious for (successfully) doing linguistics without special regard for real-life data.
So, who's right and who's wrong? Can truth be found on one side of the table? Is there a possibility of finding the truth in between the two extremes? And if so, what exactly is the truth about the "right way" of processing natural language?

0 Comments

testing prezi 

12/22/2012

0 Comments

 
Online presentations made simple and fun. 
The result is a visually appealing "prezi" that gets the message across. 
Testing Prezi
0 Comments

Nouns, verbs, semantic network, and the delusion of permanence

9/3/2012

0 Comments

 
In Tibetan zen philosophy  "delusion of permanence" is a central condition of the human mind and an obvious obstacle to being happy and enjoying life. I read all about it in the booklet "How To Be Happy" by the Lama Riponche. 
It's true that the world around us is composed of nouns and verbs. Nouns are objects, entities. Verbs are relationships. The Sun and the Moon are nouns. The eclipse of the Sun by the Moon is a relationship. Nouns are permanent, verbs are transient. 
The delusion of our perception is that this noun permanence dictates a verb invariance. In reality, there is no such thing as verb invariance. All relationships are dynamic, i.e. change in time. 
Attaching to "permanent" objects encourages attachment to impermanent relationships. 
Nouns are binary: they either exist or not. 
Relationships are part of a continuum: they exist in various degrees. 
That's what the network is about. The nodes are nouns, the arcs are verbs. The arcs are weighted ever-dynamic relationships. The nodes set the scene. 
Relationships can be nominalized when "fossilized" in the network.
0 Comments

    What is this space for

     A computational linguist's pond. Sometimes informational, some others instructional, or just a vent about the state of the field nowadays. 

    Archives

    April 2016
    March 2013
    January 2013
    December 2012
    September 2012

    Categories

    All

    RSS Feed

Powered by Create your own unique website with customizable templates.