# Causality: some notions, grammars, perspectives and links with GDS/cadCAD

###### tags: `Presentations`

*Danilo Lessa Bernardineli (BlockScience), May 2023*

---

## Agenda

- Causality: what it is, what shape it takes, the form which can take and the importance of it.
- Expressing Causality Graphically: DAGs, CLDs, Discrete Time DAGs and linking with cadCAD Legacy & 1.0.
- Perspectives on Causality and Simulations: discussions

---

## Causality

"Causality (also called causation, or cause and effect) is influence by which one event, process, state, or object (a cause) contributes to the production of another event, process, state, or object (an effect)" [1]

----

### Causality in different contexts

- In science: RCT and controlled experiments. Causal Inference.
- In law: proximate cause, cause-in-fact
- In econometrics: Granger Causality, VAR
- In decision theory / psychology: Causal Decision theory
- In engineering: causal, acausal and anticausal systems.

----

### Some causal relationships

- Type
- Unidirectional: A causes B
- Bidirectional: A causes B and B causes A
- Structure properties
- Delayed: A causes B, and after an delay, B causes C
- Mediation: A causes B and C. B causes C.
- Non-relationships
- Confounder: A causes B and C. As a result, B and C are correlated.
- Collider: A and B causes C. As a result, spurious correlations between A and B can be generated if C is taken into account

----

### Why causality is important

3 levels on the ladder of causality reasoning:

1. Association - Is A related to B?
2. Interventing - Will A be related to B if it is modified?
3. Counterfactuals - What if A was related to B and C instead?

Statistical methods answers L1 type of questions, with some progress on L2 questions. L3 is on the human expert domain.

Causal methods provides an form to reason, model and test L2 and L3 questions and to avoid bias on L1 ones.

Governance and decision-making questions tends to deal with L3 and L2 questions.

Reference: [4]

---

## Expressing Causality Graphically

----

### Putting all together: Causal DAGs


----

### Making it mutually causing as a whole: Causal Loop Diagrams.

Filecoin's economy CLD [3]


----

The Societal Costs and Benefits of Commuter Bicycling: Simulating the Effects of Specific Policies Using System Dynamics Modeling [2]


----

### Discrete Time and Causality


----

### Legacy cadCAD and Causality

Legacy cadCAD SUBs can be understood as an imposed base structure on top of an Causal Loop Diagram.


----

### cadCAD 1.0 and Causality



---

## Perspectives on Causality and Simulations

----

### GDS & cadCAD 1.0 as an grammar for multi-scale causal systems

.

----

### Coupling causal unobservables with associated observables


---

## References

[1] - https://en.wikipedia.org/wiki/Causality
[2] - Macmillan, Alexandra & Connor, Jennie & Witten, Karen & Kearns, Robin & Rees, David & Woodward, Alistair. (2014). The Societal Costs and Benefits of Commuter Bicycling: Simulating the Effects of Specific Policies Using System Dynamics Modeling. Environmental health perspectives. 122. 10.1289/ehp.1307250.
[3] - Danilo Lessa Bernardineli Michael Zargham and Jamsheed Shorish (2023). Reviewing the FIP-0056 and CDM Debate on Filecoin. Retrieved at 17 May 2023 at https://medium.com/block-science/reviewing-the-fip-0056-and-cdm-debate-on-filecoin-6a6af0ed4b78.
[4] - Reference: Pearl, J & Mackenzie, D. The Book of Why
[5] - Facure, M. Causal Inference for the Brave and True. Retrieved at 17 May 2023 at https://matheusfacure.github.io/python-causality-handbook/landing-page.html