IFT 6760B A26 - Consolidation of GFlowNet theory

name: gflownets-consolidation-20260921
class: title, middle

## GFlowNets: Sampling as sequential decision making
### IFT 6760B A26

#### .gray224[September 21st - Session 6]
### .gray224[Consolidation of GFlowNet theory]

.smaller[.footer[
Slides: [alexhernandezgarcia.com/teaching/gflownets26/slides/{{ name }}](https://alexhernandezgarcia.com/teaching/gflownets26/slides/{{ name }})
]]

.center[
<a href="http://www.umontreal.ca/"><img src="../../../assets/images/slides/logos/udem-white.png" alt="UdeM" style="height: 6em"></a>
]

Alex Hernández-García (he/il/él)

.footer[[alexhernandezgarcia.com](https://alexhernandezgarcia.com/) | [alejandro.hernandez.garcia@umontreal.ca](mailto:alejandro.hernandez.garcia@umontreal.ca)] | [alexhergar.bsky.social](https://bsky.app/profile/alexhergar.bsky.social) [![:scale 1em](/assets/images/slides/misc/bluesky.png)](https://bsky.app/profile/alexhergar.bsky.social)<br>

---

## Objectives of this session

- Expand and consolidatate the fundamentals of GFlowNets:
    - Formal definitions
    - Main properties
    - Additional loss functions

The goal is that at the end of the session:
- You will be able to explain the fundamentals of GFlowNets using well grounded theory.
- You will be able to connect the main properties of GFlowNets to the theory.
- You will know of additional loss functions valid to train GFlowNets.

---

## Literature

.references[
Yoshua Bengio, Salem Lahlou, Tristan Deleu, Edward J. Hu, Mo Tiwari, Emmanuel Bengio. [GFlowNet Foundations](https://arxiv.org/abs/2111.09266). JMLR, 2023.
]

.center[![:scale 70%](/assets/images/teaching/gflownets/gfn-intro/paper_foundations.png)]

---

## Literature

.references[
Tristan Deleu. [Generative Flow Networks: Theory and Applications to Structure Learning](https://arxiv.org/abs/2501.05498). PhD thesis, Université de Montréal, 2025.
]

.center[![:scale 80%](/assets/images/teaching/gflownets/gfn-intro/thesis_tristan.png)]

---

## Literature

.references[
Nikolay Malkin, Moksh Jain, Emmanuel Bengio, Chen Sun, Yoshua Bengio. [Trajectory balance: Improved credit assignment in GFlowNets](https://arxiv.org/abs/2201.13259). NeurIPS, 2022.
]

.center[![:scale 70%](/assets/images/teaching/gflownets/gfn-intro/paper_tb.png)]

---

name: title
class: title, middle
count: false

## GFlowNets: Sampling as sequential decision making
### IFT 6760B A26

#### .gray224[September 21st - Session 6]
### .gray224[Consolidation of GFlowNet theory]

.center[
<a href="http://www.umontreal.ca/"><img src="../../../assets/images/slides/logos/udem-white.png" alt="Université de Montréal" style="height: 6em"></a>
]

Alex Hernández-García (he/il/él)