On 7/24/23, Paula .* wrote
They (Pytorch et Tensorflow authors) generally would not discuss the tensor operations used to implement the NNs.
adawan919@gmail.com Tue, Jul 25, 2023 at 3:27 PM:
Yes, indeed, it is possible for a string (or an expression or a lexical item... etc.) to refer to different things based on different contexts ...
Rodolfo Delmonte via Corporacorpora@list.elra.info Tue, Jul 25, 2023 at 4:01 PM
In fact tensors should capture ideally both paradigmatic and synthagmatic properties of a word in a sentence given the fact that they are usually made up of matrices, that is at least couples of vectors where the rows are represented by ...
At the risk of being considered a "purist", an "elitist" and if I am following your comments (making sense of them somewhat hopefully), tensors and matrices are definitely more than a visual table-like arrangement; which is also the case of vectors, vector operations, vector space, points in a space, distance between two points (as a zero order invariant tensor) ...
Take a bunch of texts and first show to me how do you define "space", then "vector", ... in a thoroughgoing "character-by-character" way. For example, how could you then use vector addition parallelograms to explain paraphrasing and go about summarizations in a corpus ...
Such concepts have been very profitably cultured for millennia by generations after generations of Mathematicians and empirical scientists to very precisely land rovers on the moon and to coordinate the work of the robots they use to make transistors.
If those Pytorch et Tensorflow yahoos (behaving more like politicians and magicians than true to matters tech monkeys) would not even show what they mean how are you so sure about what you mean when you speak of "tensors", "vectors", ...
How does the concept of vector in a space translates to whatever you mean by "vectors" in a text bank/corpus. What would be its magnitude and direction? How would you calculate a dot product between two vectors? ...
Here is a very basic introduction to what a dot product and a tensor mean:
// __ Tensors for Beginners 9: The Metric Tensor