Sunday May 14th

5/14/17 - De-stressing and some Notes


  • Just finished my Queens lab (I updated to include a solution in my last post).

  • Man, that was stressful! I have one more assignment to do, and then a board on which I must comment/ provide a solution and then I’m all caught up (I’ll probably get more on Tuesday, so..yeah.)

In the Meantime

Notes from Ray Puzzio’s “Homotopy Type Theory” lecture. I believe they’re from the first chapter of the HoTT book. His presentation was part of the LispNYC group’s meeting.

I also follow this blog, which is neat, too :)

I really enjoyed this presentation. I watched all the way from the beginning, to the end. Really, really enjoyable, and I learned a lot. I also found his explanations to be incredibly clear. The Predicate Logic course that I’m taking also went hand in hand, which was helpful, because it did touch on \Sigma and \Pi vs \Forall and \ThereExists.

Some of my writing is in LaTeX notation.

I’m putting them here so that later on, I can look back, amend, etc.

Homotopy Type Theory - Ray Puzzio

Two types of Type Theory

  • Simple types -> based on Church
  • Martin Luff -> Dependent Type Theory

Type family

  • morphism -> mappings
  • objects -> collections, types

same types a -> b then you can hook them together and compose them

  • groupoids

  • group -> has one object and is invertible
  • monoid -> a category with one object
  • monoid has one objects which means any two categories can be composed

object - bookkeeping device that tells us when we can multiply things together

structured sets: a canonical class of examples of categories are sets with structures. Morphisms are mappings which preserve the structure in question.

  • sets -> no structure, anything goes
  • graphs-> structure is edges and morphisms must map edges to edges
  • topological spaces -> have notion of neighbourhood,
  • morphisms must be continuous
  • algebraic systems - have operations which morphisms must respect

objects as graphs and morphisms to be mappings between graphs that is an example of a category

  • Forgetful map -> erase all the edges that give me vertices

  • Functor is a map between categories that preserves the structure of the category

Topological space -> concept of neighbourhood (far away vs near)

  • you want mappings that preserve that structure


Yoneda’s lemma shows that any category can be represented as mappings of its generalized elements which respect the composition structure of that category

Determination through Universals For any object with this property, there must be a unique map to this one

pullback fiber product

  • a set of all subsets = powerset

cartesian closed category elementary topos

Grothendieck -> Topoi

BHK Interpretation

  • Brouwer, Heyting, Kolmogorov


Shoenfinkel showed that all combinators (lambda terms with no built-ins or free variables) can be constructed by repealted application of the following two basic combinators K = \lambda_{y} \lambda_{x} x S = \lambda_{z} \lambda_{y} \lambda_{x} x(z(y(z))))

Look at: Unlambda Programming language


type is like an object -> generaal notion of a collection

Objects of Category C(T) are types of T Morphisms C(T) are pairs (x, tau) where x is a variable and tau is a term with only x as a free variable

Identity morphism is (x, x)

Composition of (x_{1}, \tau_{1}) with (x_{2}, \tau_{2}) is (x_{1}, \tau_{2}(x_{2})\tau_{1})

This defines a Cartesian closed category

Geometric picture

number theory topology

Grothendieck -> theory was about numbers as topological spaces

(this was a table: analogy of terms)

Logic -> Geometry (heading)

proposition -> space witness -> point and -> product or -> coproduct implication -> mapping space

In constructive math, instead of quantifiers (for all, there exists), which return truth values, use PI, Sigma as quantifiers instead. f: Pi_{x:A}P_{x}

Lawvere quantifiers


Kock-Wraith diagram

slice category -> given a category create a new one whose subsets are a slice of the old category

fibrations -> Geometry analagous to adjunctions

fibrations -> closely related to Sheaves

Path interpretation We can interpret equality types as paths.

Other (upcoming lectures)

This was at CUNY (is coming up), from one of the authors of the HoTT book:

Emily-> Higher Dimensional Category Theory paths-> equivalences into paths into category theory

Written on May 14, 2017