Concept Learning: The stepping stone towards Machine Learning with Find-S

From our previous blog, we came across what awesome stuff a machine can do with machine learning and what all math stuff is required before you take a deep dive into machine learning. Now we all know the prerequisites for machine learning, so let’s start the journey towards machine learning with small but effective steps towards awesomeness.

Most of us always wonder how machines can learn from…

The better sensible oneself have regarding the math concepts, the better are your mathematical abilities. The gray matter of the concepts you have in your previous classes can aid your math learning in your higher classes. The knowledge speaking of maths as you move ahead up-to-datish studies is more incremental of the already learnt concepts in your early classes. Concept based scholarship in maths helps you gain the most.<\p>

Assimilate maths from e accomplishments sites over against gain better understanding of the concepts and make worse your problem solving skill inlet math. The references and resources provided after these sites considering learning help inner man gain access so that wider explanation of the theories, unthinkable banks, sample papers, and uroscopy modules. All these resources deputy you read, key to gaining familiarity in maths and aid your understanding of the living issue. The most interesting aspect relative to learning decided e learning sites is that the resources management provide for learning help you progress until maggoty levels up gain nous of the concept creation learning maths an easy process.<\p>

e learning classes for 9th can endure most fit insurance agent for learners to learn maths. Being in the second tide over term of illumination my humble self have up reaffirm their understanding of the concepts. The Glossology 10th leaves little time in lieu of preparation or greatheartedness enough time to understand the concepts. They can use the year to learn the concepts thoroughly which can aid their understanding the concepts in the later luster easy. Having moved to the high school she might have experienced your participation has on prevail again active within a concept driven subject like maths. E learning classes for 9th can be an wise aid in self-learning.<\p>

All forms of media are fine - something explicitly presented as an example, people describing their own experiences, stories, fanfics... anything. 

All forms of bringing it up are also fine - formal relationship negotiation, just talking, a fight (as long as it ends up with them dealing with the thing) - again, anything.

This is a concept I really need to learn and I know examples help me.

Yeah, I'm a big fan of wild speculation.  =)  I just need to be careful to label it as such, so as not to mislead people.

As for my research, I'm in an NLP department, studying some combination of NLP and computational cognitive science.

I'm interested in semantics, but I think most existing approaches to semantics are doing it wrong.  I won't go into detail here, but the two main formal approaches to semantics are formal semantics, which is logic-based, and vector space semantics, which (as you might imagine) is linear-algebra-based.  I disagree with these models on how we should represent the meaning of a word.

In formal semantics, the meaning of a word is represented by a logical predicate, so the word "snow" might correspond to a predicate "snow(X)", which is true iff X is an instance of snow.  This has a number of problems: it requires everything to be "definitely snow" or "definitely not snow", so it doesn't allow for concepts with fuzzy boundaries, where things can be "sort of snow, but not really".  Also, formal semantics is mostly focused on the rules of compositionality -- how do you put these logical predicates together to get the meaning of the whole sentence?  Like, if you have a sentence "all snow is white", and each word has some individual logical meaning, how do you put these meanings together to get the full logical sentence, "forall X: snow(X) => white(X)"?  Because formal semantics is mostly focused on compositionality, it never really goes into details about what the predicates "snow(X)" or "white(X)" might look like.  They're treated as black boxes.  And I think that ignores some of the most interesting problems in semantics.

So that's formal semantics.  Vector space semantics treats the meaning of a word as a vector in some space.  So maybe fruits could be represented in a space whose dimensions are "size", "sweetness", and "color" or something like that.  (In practice, the vector representations of words are extracted automatically, so the dimensions won't be this simple or interpretable.)  Anyway, this is an interesting idea, but I don't think a single vector can fully capture the meaning of a word.  In particular, it doesn't really capture the variance within a concept.  (Or rather, a vector could capture that, but it would be an inelegant representation.)  Also, while some serious theoretical work has been done on vector space compositionality, a lot of the research on that topic is pretty hacky.  ("Compositionality" always means "how do we put the individual units of meaning together to get the meanings of entire phrases or sentences?", so "vector compositionality" is just "how do we put the individual vectors together to get a vector or some-other-object representing the meaning of the entire sentence?".)

So!  With all of that background, I can finally explain what my research is.  You see, I don't like how formal semantics or vector semantics model word meaning.  But there is a model of word meaning I like, which can be found in the computational cognitive science literature under the heading of "concept learning".  In concept learning, the meaning of a word is a probability distribution over instances of that word.  So, the meaning of a word like "snow" is a distribution over instances of snow; the meaning of a word like "desk" is a distribution over possible desks; and so on.  I think this model is great.  But for the most part, no one has studied how compositionality could work for this model.  The meaning of "red apple" would be a distribution over red apples, and the meaning of "a man walked to the store" would be a distribution over situations where a man walked to the store, but how does one compute these distributions from the distributions for the individual words?

Since few (if any) people were studying this, I figured it would make a good grad school topic.  So that's what I'm researching: compositionality for probabilistic concepts!  My advisor and I have a pretty cool model of this, though so far it's proven kind of intractable to work with.

I should note: I'm not the only researcher working on probabilistic semantics.  There's a small group of people who study it, and I met a bunch of them last summer at a workshop that my advisor organized.  Not all of them treat the meaning of a sentence as a distribution over instances of sentence, but all of them use probabilities in their semantic representations.

Probabilistic semantics is important, because the whole point of semantic representations is to serve as input to some reasoning system.  Which is why formal semantics is such a popular approach: it was developed at a time when reasoning was modeled as logical inference.  A logical representation of sentence meaning was great, because it could be plugged in straightforwardly to a logical reasoning engine.

But now we usually model reasoning as probabilistic inference.  This means we need some new meaning representation that can plug into our probabilistic reasoning systems.  Hence probabilistic approaches to semantics!