Video Summary - EMAp Summer Course - TDA w PH - Lesson #1 Topological Spaces

Main Ideas, Concepts, and Lessons

Course introduction and goals
- The speaker introduces himself as Rafael and welcomes participants to the first class of Topological Data Analysis (TDA).
- The course aims to develop a topological viewpoint on geometry and data:
  - In topology, objects are considered up to deformation (you can twist and bend).
  - Key consequence: many shapes that look different geometrically can be considered equivalent topologically (examples mentioned: circle ~ square, donut ~ nug).
- TDA motivation/impact
  - TDA is presented as a young theory (~20 years old) with applications in medicine, chemistry, and image analysis.
- The speaker emphasizes that participants may need to interrupt for questions because language communication may be difficult (French accent / Brazilian accents, and the speaker not speaking Portuguese).
Course structure and schedule
- The course is planned over three weeks, each with increasing topic breadth:
  - Week 1: General topology
    - Learn topological spaces, homeomorphisms, homotopies, and simplicial complexes.
    - The first lesson today focuses on topological spaces, described as “abstract/opaque” for beginners.
  - Week 2: Homology
    - Algebraic topology concepts.
  - Week 3: Persistent homology
    - Presented as a main tool of TDA.
- Coding and tutorials
  - Implementation will be in Python.
  - For persistent homology, they will use the GUDHI library (mentioned as “goody” in subtitles).
  - They will use Jupyter notebooks and NetworkX.
  - Participants are instructed to download required libraries before tutorials.
Homework and interaction rules
- Participants:
  - May interrupt anytime during the lesson for clarification/repetition/questions.
  - Must not skip homework, as the course is intense.
  - Will have daily homework for the next lesson(s).
  - Should send homework by email for correction.
- First homework assignment (intro stage)
  - Send an email answering:
    - Whether they know topology
    - Whether they know how to code in Python
    - Any remarks/questions

Detailed Methodology / Instruction-Style Items

How to run/prepare for Python tutorials
- Before the first tutorial at the end of the week:
  - Download required libraries:
    - GUDHI
    - Jupyter
    - NetworkX
  - Ensure the provided notebooks can be executed.
Class support process
- If participants have questions:
  - Use the chat on the laptop.
  - If the question is not seen immediately, open mic and ask verbally.

Topological Spaces: Formal Definition (Core Lesson)

A topological space is a pair ((X, \mathcal{T})) where:

(X) is any set.
(\mathcal{T}) is a set of subsets of (X) (called open sets) such that:

Empty set and whole space included
- (\emptyset \in \mathcal{T})
- (X \in \mathcal{T})
Arbitrary unions closed under openness
- If ({O_\alpha}\subseteq\mathcal{T}), then (\bigcup_\alpha O_\alpha \in \mathcal{T})
Finite intersections closed under openness
- If (O_1,\dots,O_n \in \mathcal{T}) (finite list), then (\bigcap_{i=1}^n O_i \in \mathcal{T})

Closed Sets Derived from Open Sets

A subset (C\subseteq X) is closed iff its complement is open:

Properties shown:

(\emptyset) and (X) are both closed.
Arbitrary intersections of closed sets are closed.
Finite unions of closed sets are closed.

Demonstrations rely on complement logic using open-set axioms.

Constructing a Topology on (\mathbb{R}^n): Euclidean (Metric) Topology

Use the Euclidean norm / distance:
- Distance between points (x,y\in\mathbb{R}^n) is (|x-y|).
Define open balls: [ B(x,r)={y\in\mathbb{R}^n : |x-y|<r} ]
Key geometric inclusion fact (proved using triangle inequality):
- If (y\in B(x,r)), then [ B\bigl(y,\, r-|x-y|\bigr)\subseteq B(x,r) ]
Define open sets in (\mathbb{R}^n) via balls:
- A set (A\subseteq\mathbb{R}^n) is open if:
  - for every (x\in A), there exists some radius (r>0) such that (B(x,r)\subseteq A).
Verify the Euclidean open-set collection is a valid topology:
- (\emptyset) and (\mathbb{R}^n) are open.
- Arbitrary unions of open sets are open.
- Finite intersections of open sets are open (uses minimum of finitely many radii).

Examples About Openness/Closedness in (\mathbb{R})

Open interval ((0,1)): open
Closed interval ([0,1]):
- not open; but closed (complement is open)
A set “open on one end” (from subtitles’ second example):
- concluded to be not open and not closed (its complement is not open, especially at endpoints)
Singletons ({x}):
- not open but closed
Rational numbers (\mathbb{Q}):
- claimed (as an exercise/meditation) to be neither open nor closed

Subspace Topology (Topology on Subsets)

Given a topological space ((X,\mathcal{T})) and a subset (Y\subseteq X):

[ \mathcal{T}_Y={Y\cap O : O\in \mathcal{T}} ]

(i.e., intersections of (Y) with open sets of (X)).

The speaker outlines checking the topology axioms via set operations (union/intersection/complement reasoning).

Examples of subsets mentioned (as contexts for applying subspace topology):

Circle (as a subset of (\mathbb{R}^2))
Sphere
Cube/square
- defined via max coordinate absolute value (\le 1)
Closed balls vs open balls
Standard simplex (\Delta)
- highlighted as important later when simplicial complexes are introduced

Continuity of Maps (Defined Using Topology)

Setup:
- Domain: ((X,\mathcal{T}))
- Codomain: ((Y,\mathcal{U}))
- A function (f:X\to Y)
Topological definition of continuity
- (f) is continuous if for every open set (O\in\mathcal{U}), [ f^{-1}(O)\in\mathcal{T} ] (the preimage of open sets is open)
Equivalent characterization mentioned:
- (f) is continuous ⇔ preimage of closed sets is closed
Example of non-continuity:
- A step function-like example (value 0 for negative inputs, 1 for nonnegative inputs) is shown to be not continuous by finding a closed set in the codomain (containing the value 1) whose preimage is not closed in the domain.
Properties:
- Composition of continuous maps is continuous:
  - If (f:X\to Y) and (g:Y\to Z) are continuous, then (g\circ f:X\to Z) is continuous.
Connection to metric intuition:
- The speaker states that this topological notion matches the standard metric/epsilon-delta continuity between Euclidean spaces:
  - for every (x) and every (\varepsilon), there exists (\eta) such that if (y) is within (\eta) of (x), then (f(y)) is within (\varepsilon) of (f(x)).

Speakers / Sources Featured (As Identifiable)

Raphael / Rafael — main instructor/speaker (TDA course lecturer)
Cesar — organizer/introducer mentioned
Participants/students (unnamed) — multiple questions/comments in chat
- Guillermo — student commenter
- Lucas — student commenter
Frederick Chazal — referenced as part of a “best team in the world” / associated with TDA
Archer (Allen) Hatcher — referenced for the book “Algebraic Topology” (free online resource)
GUDHI — library used for persistent homology
NetworkX — Python library used for coding tutorials
Jupyter Notebook — tool used for tutorials

EMAp Summer Course - TDA w PH - Lesson #1 Topological Spaces

Key takeaways