# Journal Club in Deep Learning and Computer Vision (DLCV_JC)
Organisers: Kandakji Lynn, Mannering Harvey Minano Sofia, and Xochicale Miguel

## Aims
DLCV_JC aims to create a productive space for conversation on reseach computing including articles, blogs, papers with code, datasets and codebases.

To keep up to date with latest developments in deep learning and computer vision, and understand the gist behind the ideas that have revolutionised the field (e.g., transformers, stable-diffusion models).

## Topics
Focused on deep learning and computer vision, aiming to gain an in-depth understanding of well-established tools that we can use in RSDG projects (e.g., state-of-the-art object detectors, pose estimation models, image segmentation, synthetic image generation, etc.)
* Papers (Computer Vision, etc)
* Datasets (Medical Imaging, etc)
* Codebases
* Blogs
* Papers with code
* Reproducing papers with code
* Seminal papers and state-of-the-art papers
* Specifit topics like NERF, Transformers, Diffussion models, etc

## Format of the JC
* The idea is to prepare a paper to discuss + a coding tutorial to go along it
* We'd like to be quite flexible on the paper presentation format so we leave that up to the speaker.
* To start with we are thinking one speaker preenting one paper, but this could also be more flexible, for example:
* One paper each — coordinated: the coordinator is forced to choose a topic, screen several articles and make a selection of four references to be presented in hour session.
* One paper each frestyle: Four participants describe an article of their choice to be presented in hour session.
* Some ideas for the coding tutorial: it could be reproducing some of the paper's figures, a notebook based on a tutorial available online, or a proposed project to work on the next few weeks using the paper's repo.
* Potentially useful resources for the tutorial: PaperWithCode, existing projects in HuggingFace, Kaggle
* We are thinking of hosting the tutorials material in a GitHub repo
* Create an space to organise JC4RSE (e.g., https://www.youtube.com/watch?v=3v2MlfgkwUY, repo?)
* For now, we are filling in this [Spreadsheet for papers + possible tutorials, and speakers, etc](https://liveuclac-my.sharepoint.com/:x:/g/personal/ccaemxo_ucl_ac_uk/Eb45VIEdVnFKhrKeyKOAEBIBip0HSNdSq8EAvZv-oO9n3w?e=3WQF1d)
* Going forward we would like to have a voting system: This might be a good way to find common areas of interest https://www.youtube.com/watch?v=3v2MlfgkwUY.

## Guidelines for presenters
* Select a paper to present for 30 to 40 minutes using slides.
* We prefer the papers to be presented with replicated code and data accesiblity but also open to only present papers.

## Dates and times
* Journal club session on the second Thursday of the month at 3pm
* Potentially starting mid Jan 2023
* Drop-in sessions on the ~~three following Fridays~~ following Friday, same time (3pm)
* mainly to go through code together
* The last Friday before the next journal club session is reserved for planning and preparing the next session

## Roles and speaker list
We will rotate between the following roles
- Presenter: in charge of preparing paper presentation and tutorial
- Comms person: in charge of communication, announcement messages and replying to emails. Also drafts bio of speaker+abstract (& sends it for approval), and the poster.
- Chair: in charge of presenting the speaker and introducing the paper session.
- Backup person: Anyone aditional to support us in case of missing chair or representive of comms.

See [Spreadsheet for papers, speakers, etc](https://liveuclac-my.sharepoint.com/:x:/g/personal/ccaemxo_ucl_ac_uk/Eb45VIEdVnFKhrKeyKOAEBIBip0HSNdSq8EAvZv-oO9n3w?e=3WQF1d)

| Date | Speaker | Comms | Chair | Backup | Notes |
| --- | --- | --- | --- | --- | --- |
| Jan2023 | Harvey | Miguel | Sofía | Lynn | - |
| 02-March 2023 | Sofía | Harvey | Lynn | Miguel | Lynn--subs by Miguel |
| 31-May 2023 | Miguel | Sofia | Harvey | Lynn | Sofia also making central repo |
| 27-July 2023 | Harvey | Miguel | Sofia | Lynn | - |
| | | | | | |

## Invitation messages

### Typical timeline for messages
Options...
- ~~2 weeks before~~
- at least 1 week before, max 2 weeks
- ~~one day before~~
- on the day

Send a calendar invite to [e-mail] and [e-mail] for the two sessions with Teams link

### Setting teams invite for consecutive days
Use microsoft teams to set recurrence to next day at the same time!

### Webinar link
Guidelines to set-up a webinar in Teams [here](https://learn.microsoft.com/en-us/microsoftteams/quick-start-meetings-live-events#webinars)

### Channels to post in
Announcements should go on the following slack channels too:
- https://ucl-programming-hub.slack.com/channels/dl-cv-journal-club
- https://ucl-programming-hub.slack.com/channels/general
- https://ucl-arc.slack.com/channels/arc (Ilektra mentioned the research programming hub is not regularly checked by everyone)
- UCL Centre for AI: https://uclcentreforai.slack.com/archives/CMDSF4YEB

Feedback from David Perez-Suarez:
> That link is only for today's event... I'd suggest to either invite all ARC (as done with the tech socials) when each event is confirmed, create a new calendar under one of our shared inboxes and share the calendar within ARC so people could activate and see it next to them and add themselves to each event when needed, or create a MS Teams team (or channel under ARC) that people can join and then see the events planned there.
https://ucl-arc.slack.com/archives/C067ELXT9/p1673610843226809?thread_ts=1673605318.951179&cid=C067ELXT9

### Template

Dear All

We are delighted to invite you the Journal Club for RSEs, aiming to create a productive space for conversation on reseach computing including articles, blogs, papers with code, datasets and codebases.

Title: ?
Abstract: ? 
Short-bio speaker: ?

Harvey Mannering will present paper and live coding of:
"Denoising diffusion probabilistic models by Ho, Jain and Abbeel. Advances in Neural Information Processing Systems 33 (2020): 6840-6851. https://arxiv.org/abs/2006.11239 

Time and date

Journal club is going to be in person at [ADD/BOOK SPACE] and also online [TODO: ADD LINK]. 

Thanks,
Kandakji Lynn, Mannering Harvey  Minano Sofia, and Xochicale Miguel 

## Chairing event
Hi everyone!
Welcome to the first session of the deep-learning and computer vision journal club.

**[intro to JC]**
My name is Sofia, I'm a Research software developer at ARC and together with Harvey, Miguel and Lynn we are organising this monthly journal club, with the goal of keeping up to date with the latest developments in computer vision and deep learning. Our main aims are
- to gain a good understanding of how fundamental tools in computer vision and deep learning work,
- and identify in which scenarios they are best applied [...]

**[structure]**
We wanted to include a coding component in the journal club so we planned the following structure.
- The second Thursday of the month, we will have a 1h session focused on a specific paper, which we will discuss for about 20-30min and then for the remaining we will work on a coding tutorial related to the topi
- Then the following day at the same time, we plan to have a drop-in coding session. We were thinking we can use this time to code together, explore further the tutorials, ask questions or try out other tutorials or datasets. This will take 1h too

**[slack]**
We have started a **slack channel** in the RHub for the journal club, to post announcements, reminders, to chat about the methods we would like to discuss, further reading material etc.

**[today]**
These are the main admin bits, and now for the session of today we have Harvey Mannering presenting a paper and a tutorial on diffusion models. Harvey Mannering is a research software developer at Advanced Research Computing Centre here at UCL, and before joining this team he completed his masters in computer graphics, vision and imaging. His research interests primarily are around AI in the context of computer vision and deep learning.

So whenever you are ready Harvey, feel free to get started

## Invitations

#### 05 Jan 2023 @10am ~~04 Jan 2013~~

>
> Hi all,
>
> Happy new year! :confetti_ball:
>
> We are delighted to invite to a new Journal Club for RSEs, aiming to create a productive space for conversation around research computing areas including articles, blogs, papers with code, datasets and codebases. We are starting with topics on deep learning and computer vision including the coding component.
>
> * Title: Denoising diffusion probabilistic models (DDPM).
> * Summary: Diffusion models have recently achieved state-of-the-art results on image synthesis. In this talk, Harvey Mannering will discuss and present a tutorial with live coding examples of:
"Denoising diffusion probabilistic models by Ho, Jain and Abbeel. Advances in Neural Information Processing Systems 33 (2020): 6840-6851. Arxiv: https://arxiv.org/abs/2006.11239
> * Speaker: Harvey Mannering is a research software developer at the Advanced Research Computing Centre in UCL having just completed his masters in computer graphics, vision and imaging. His research interests primarily revolve around AI, in particular computer vision and deep learning.
> * Date and time: Thursday the 12th January at 15:00
> * Microsfot Teams meeting. Join on your computer, mobile app or room device. Click the following link to join the meeting: https://teams.microsoft.com/l/meetup-join/19%3ameeting_OTA3MjMwMTEtZDNlMS00NDc3LTkwNTAtMzdkZGZjNTZhMGQ5%40thread.v2/0?context=%7b%22Tid%22%3a%221faf88fe-a998-4c5b-93c9-210a11d9a5c2%22%2c%22Oid%22%3a%2271503059-6d17-49a0-87c9-2dbd5b9e2fed%22%7d
>
> Looking fordward to meeting you all!
>
> Best wishes,
> @Lynn Kandakji, @Harvey Mannering @Sofia Miñano & @Miguel Xochicale
> PS. Please reach us if you are interested to learn more (schedules, organisation, next steps, feedback, etc).

### 12 Jan 2023 @9ish-am: Reminder

> Hi all,
>
> Hope you are having a great week, we are looking forward to meeting you online today at 15:00 for the first talk of the new Journal Club for RSEs. Harvey Mannering is presenting “Denoising diffusion probabilistic models (DDPM)”. See further details below.
>
> Journal Club for Research Software Engineers is aiming to create a productive space for conversation around research computing areas including articles, blogs, papers with code, datasets and codebases. We are starting with topics on deep learning and computer vision that include the coding component.
>
> * Title: Denoising diffusion probabilistic models (DDPM).
> * Summary: Diffusion models have recently achieved state-of-the-art results on image synthesis. In this talk, Harvey Mannering will discuss and present a tutorial with live coding examples of:
"Denoising diffusion probabilistic models by Ho, Jain and Abbeel. Advances in Neural Information Processing Systems 33 (2020): 6840-6851. Arxiv: https://arxiv.org/abs/2006.11239
> * Speaker: Harvey Mannering is a research software developer at the Advanced Research Computing Centre in UCL having just completed his masters in computer graphics, vision and imaging. His research interests primarily revolve around AI, in particular computer vision and deep learning.
> * Date and time: Thursday the 12th January at 15:00
> * Microsfot Teams meeting. Join on your computer, mobile app or room device. Click the following link to join the meeting: https://teams.microsoft.com/l/meetup-join/19%3ameeting_OTA3MjMwMTEtZDNlMS00NDc3LTkwNTAtMzdkZGZjNTZhMGQ5%40thread.v2/0?context=%7b%22Tid%22%3a%221faf88fe-a998-4c5b-93c9-210a11d9a5c2%22%2c%22Oid%22%3a%2271503059-6d17-49a0-87c9-2dbd5b9e2fed%22%7d
>
> Best wishes,
> @Lynn Kandakji, @Harvey Mannering @Sofia Miñano & @Miguel Xochicale
> PS. Please reach us if you are interested to learn more (schedules, organisation, next steps, feedback, etc).

### Fri 13 Jan 08:51:42 GMT 2023: Invite to the drop-in seasson

> Hi everyone,
>
>Thanks for joining yesterday to the first talk of the new Journal Club for RSEs which also will follow up with one drop-in session for a deeper dive into the coding component of such papers.
>
>So, today at 15:00 to 16:00, we are having a drop in sessions with Harvey Mannering,
he will go through this tutorial of diffusion-models-class from HuggingFace: https://github.com/huggingface/diffusion-models-class
>
>Please register to the event using the following link which will
add the event to you teams calendar:
https://events.teams.microsoft.com/event/70e8330c-f255-4d1f-a2b4-30e34edb113e@1faf88fe-a998-4c5b-93c9-210a11d9a5c2
>
> Looking fordward to meeting you all!
>
>
> Best wishes,
> @Lynn Kandakji, @Harvey Mannering @Sofia Miñano & @Miguel Xochicale
> PS. Do not hesitate to reach out, if you are interested to learn more the JC4RSE (schedules, organisation, next steps, feedback, etc).

### Harvey: 27 and 28 July 2023

#### Invitation details
> Subject: Deep Learning and Computer Vision Journal Club > “Introduction to Text-to-image models” by Harvey Mannering
> [e-mail]
> CC:[e-mail]; [e-mail]; [e-mail]; [e-mail]

#### Content of the message for slack and invitation

:mega: Deep Learning and Computer Vision Journal Club (dlcv-jc)

Hi all,

We are delighted to extend our invitation to the talk on "Introduction to Text-to-image models" by Harvey Mannering, which will be hosted online on Thursday 27th July 2023, at 15:00 GMT, followed by coding season on Friday 28th July 2023, 15:00 GMT. Additionally, we are proud to introduce our new [GitHub organisation](https://github.com/dlcv-journal-club), where you can access additional information, code, data and more related to the event. Please mark your calendar for this exciting talk!
 
Title: Introduction to text-to-image models.
Summary: Text-to-image models are neural networks that take as input a text prompt and output an image. This groundbreaking technology has ushered in a wave of exciting new applications, spanning image editing, AI-generated art, story boarding, and data augmentation. In this talk, Harvey will delve into the technical details of these models (e.g., Latent diffusion models, ControlNET, ODISE, and DreamBooth), explore their real-world applications, and examine the ethical questions they raise.
Speaker: Harvey Mannering is a research software developer at the Advanced Research Computing Centre in UCL having just completed his masters in computer graphics, vision and imaging.  His research interests primarily revolve around AI, in particular computer vision and deep learning.
Date and time: 
Talk > Thursday 27th July 2023, 15:00 GMT.
Coding session > Friday 28th July 2023, 15:00 GMT.
Microsoft Teams meeting. Join on your computer, mobile app or room device. Click the following link to join the meeting: 
https://teams.microsoft.com/l/meetup-join/19%3ameeting_ZTQ2ZGJkMTEtMjVjYS00ZDM0LTkyMDctYTkyMTBjMjI1NzQw%40thread.v2/0?context=%7b%22Tid%22%3a%221faf88fe-a998-4c5b-93c9-210a11d9a5c2%22%2c%22Oid%22%3a%2271503059-6d17-49a0-87c9-2dbd5b9e2fed%22%7d

Looking forward to seeing you all there!

Best wishes,
@Lynn Kandakji, @Harvey Mannering @Sofia Miñano & @Miguel Xochicale
PS. Please reach us if you are interested to learn more (schedules, organisation, next steps, feedback, etc).

* Reminder

Hi #arc #general
This is a kind reminder about the talk “Introduction to Text-to-image models” by Harvey Mannering at the Deep Learning and Computer Vision Journal Club. Super looking forward to seeing you there!
Best wishes,
@Lynn Kandakji, 
@Harvey
 
@Sofía
 & 
@Miguel Xochicale

PS. Please reach us if you are interested to learn more (schedules, organisation, next steps, feedback, etc).

* BLURS

* [We belive this platform will not only enhance your research experience but also expand your knowledge on using the best software eingieering practices.] 
* [occasion, and we look forward to you participating in this enriching experience!]

## Notes on past sessions
### Sofia's session on nerf
#### Intro session
Sofia Miñano - Thu 2 Mar 15:00:39 GMT 2023
**Deep Learing & Computer Vision Journal Club**
**Title**: Capturing 3D scenes with Neural Raudiance Fields.
**Bio**: Sofía is a research software developer in the Collaborations teams at the Advanced Research Computing Centre (ARC). Before joining ARC, she completed her PhD at Oxford, for which she investigated computational approaches to model vision in birds of prey. Prior to that, she obtained her bachelors and masters in aerospace engineering. She has experience with computer vision and deep learning, particularly on their applications to animal conservation and experimental neuroscience.
**Messages**:
Coding session Friday 10 of March at 14h00

**Some numbers**
* 7 participants
* Asked questions?
* Occlusion problems with sparce dataset.
* What kind sampling and how error impact error accumilation
* How sparce is defined and the overlap of images (each point should be view twice)
* Can these be combined with simulations?
* Shared link https://climatenerf.github.io/
* Single camera with a moving object. > NeRF on moving objects.

**Potential questions from MX**
* ~~Does NeRF have any kind of validation?~~
* How the quality of NeRF models are evaluted?: "Generative models made use of Fréchet inception distance (FID) or Kernel inception distance(KID)"
* Smaller network 5Mb? How many parameters?

#### Coding session
Fri 10 Mar 14:00:32 GMT 2023
Joined persons:
2 organisers, 1 speaker
3 persons

* Slides: https://docs.google.com/presentation/d/1RNiAhRzMHrr6JfGN78SnTtAIKxdvOADF5EppPM6WA4E/edit#slide=id.g218bd9d589f_0_0
* Code:
https://colab.research.google.com/github/keras-team/keras-io/blob/master/examples/vision/ipynb/nerf.ipynb#scrollTo=f85eMbOTZSFM
* You need to be looged into google to see slides.

---
### Miguel's session on medical imaging

Repo [here](https://github.com/mxochicale/medisynth)

#### Planning
- presentation: Wed 31st May?
- tutorial/demo using Myriad: Fri 2nd June --- Harvey&Miguel have some guidelines
- include example of stopping and restarting training from checkpoint
- also visualisation of training: w&b or tensorboard (Sofia can help maybe?)
- if time: include example of using the synthetic dataset to train a classifier (this could be used in a future tutorial)
- practice presentation: 19th May
- central repo ready by then (at least a draft)

#### Intro session
20 attendees (incl 3 of the organisers)
We started [recording](https://liveuclac-my.sharepoint.com/:v:/g/personal/ucabhgm_ucl_ac_uk/EQ_uaF_-hX1Ns5nO8qVdDOMB751wYbFhICJoxu3m0IEwmw)
Prep Qs
- GANs mode collapse
- domain change
- tradeoffs diagram: how is quality defined there?
- examples of AI-based devices
- is this a suggested approach based on FDA requirements for devices? is this in place now or is it a proposal?
- how many individuals? frames as in, are actually videos collected?
- have you evaluated the effect of augmentation on performance?
- how much variability is there in the scoring across experts?
- why diffusion + superresolution? was this to increase the 'quality'?
- slide 29 compare sample synth images: what are the FID scores? how are middle and right columns diff?
- so the transformed-based performed worse in FID, do you have an intuition of why that may be? was that expected?
- what is the max spread / max FID score in the dataset? FID score between diff machines or practitioners?
- why didnt you consider VAE models?

Qs asked
- detecting abnormalities
- so the transformed-based performed worse in FID, do you have an intuition of why that may be? was that expected?
- what is the max spread / max FID score in the dataset? FID score between diff machines or practitioners?

Feedback
- Sofia: fantastic structure, and I really liked the TOC

#### Coding session
6 attendees (incl 3 of the organisers)

Feedback
- add making a new terminal before running scp
- git clone repo from the start? and add change of directories when scp?
- change to use wget + link to zenodo (rather than scp)
- can we have one copy of the data in Myriad that is shared and we can all access?
- add ref for interactive/non-interactive shell: https://www.geeksforgeeks.org/shell-scripting-interactive-and-non-interactive-shell/
- can we git clone the repo in the myriad node?
- and can we share a conda environment?
- add instructions for uncomment line in train script sh (or directly uncomment it)
- we need to load the modules in the compute node (only? or as well?)
- link to docs with the meaning of qstat state letters
- https://www.rc.ucl.ac.uk/docs/howto/
- add how to get help from Myriad?

---
### Harvey's session on `intro to text-to-image models` Thu 27 Jul 15:00 BST 2023

#### Teams invite got this response
Accepted 8, Tentative 7, Declined 28, Didn't respond 2

#### Talk
* 12 attendees (incl 3 of the organisers)

##### Notes
* MX idea: We can play some kind of background music, and timer saying we starting in 5 minutes. Example: https://www.youtube.com/watch?v=6nvsGTGnbIQ
*

##### Questions and answers
1. Cloud/local inference of models
2. size of images in latent space
3. SM: where does stable diffusion come in ODISE? Answer: as a feature extractor
5. MX:
* Do you have any thoughts on how to address content replication (e.g., more datasets, computing power, better models, etc?)
* Others > language drift, overfitting, preservation loss > https://dataconomy.com/2022/09/26/google-dreambooth-ai-stable-diffusion/

#### Coding session on 28 of July
8 attendees (incl 3 of the organisers)

##### Feedback
* SM:
* Nice to have myriad and colab options!
* Also nice to go thru the main bits of the paper first
* MX:
* Agree with SM
* Just thought that during the talk, speaker can strongly suggest to have his setup ready to be used for the code seasson (vpn, myriad, google-colabs). Also paperspace and codespace might work as well.
* Encourage our community to contribute to our work via repo and follow up with any questions they migth have
* Feel I do not know how to use myridad. Just learn `Eqw: there was an error in this jobscript. This will not run.` > https://www.rc.ucl.ac.uk/docs/howto/#job-states

---
## Feedback on sessions
### Notes from Thu 12 Jan 15:36:02 GMT 2023
MX
* Adding screenshots of the paper in the notebook
* Roles: Presenter, communication, chair, back-up person.
* "We will go to this tutorial"

SM:
* share link to colab/share jupyter notebook beforehand
* the paper and the notebook can be attached as files to the webinar invite I think
* try to give a broader overview of the topic (eg. give a bit of context for generative models before going into diffusion models, etc....). it might be a bit tricky with time tho.

----------------
## Meetings
### Tue 20 Dec 12:24:04 GMT 2022
* Lynn is happy to select a topic and be considered as an backup

* Format of the session

* Half paper presentation, half coding, Fridays for drop-in sessions to work on the coding tutorial together
* Remotely or in-person; aiming for 2nd Thursday of the month
* Could be a demo of the papers!
* Journal club session
* 20 minutes paper
* 10 minutes Q/A
* 30 minutes tutorial (can include some time on Q/A and discussion too).
* The tutorial could be just introduced during the journal club (then we can go into more detail in the drop-in sessions)
* How is it going to run the tutorial? bringing their own laptop using colab notebooks!

* Harvey presenting https://arxiv.org/abs/2006.11239 with tutorial/live coding on 12 of Jan 2023 (after lunch)
* Look for tutorials to avoid starting from scratch
* Why these ideas work well? When would we use this method?
* Share slides for internal feedback if desired

* SM: Potentially create two separate slack groups, for organisers and community?

* Advertise JC on 4thJan2023, 9Jan2023

* Next Meeting 11Jan2023at12:00
* trial run of the session
* in prep: everyone has read the paper, sent suggestions of tutorials to Harvey

### Wed 11 Jan 12:24:47 GMT 2023
* 20-mins paper, 20-mins code, 10-mins QA
* MX Send messages for remainer of the JC
* MX Reply emails of speakers
* Journal club dropping seesion: 13Jan2023;15-
* Sofia will present on 9-Feb-2023
*

### Fri 3 Feb 14:43:39 GMT 2023
* Friday 17 Feb 2023 or push it to 03 March
* SM shares https://docs.nerf.studio/en/latest/quickstart/custom_dataset.html > https://docs.nerf.studio/en/latest/quickstart/custom_dataset.html#kiri-engine-capture > https://colab.research.google.com/github/nerfstudio-project/nerfstudio/blob/main/colab/demo.ipynb#scrollTo=WGt8ukG6Htg3
* SM to decide which day and time
* MX to reach out David-PS for suggetsions on club
*

### Tue 20th June: prep meeting for next session

- Harvey proposes fine tuning stable diffusion for next session
- he's gonna be looking at text to image papers
- tutorial on DreamBooth; if time: loading a pretrained model in Myriad
- we can try the broad concept and narrow down at the end with the tutorial
- 24th - 28th July week? Thu & Fri?
- roles: Miguel is comms and Sofia is chair
- practice run 21st July and do the repo---and send invite
- send message

- Repo
- Miguel's repo from last session
- maybe an organisation instead and we have a repo for each session
- we can have a template for each repo with standard readmes etc

- Generally,
- it'd be nice to get feedback from the audience
- reach to Lynn and ask her if she'd like to present sth

- Attention and transformers
- context-RCNN has a simple attention module
- we can do this session next all together?

- Miguel mentions image to text for diagnosing X-Rays
- he has a student working on this
- they got feedback re validation
- they have a prototype that could be turned into a notebook/tutorial
- the student, she will finish early July, so maybe August

- Miguel comments in2research project
- maybe after August
- they are trying to make a demo for NVIDIA (they have specific hardware for medical devices - AGX)
- still early stages atm but aiming for a demo by the end of the year

### Wed 12th July catchup
- We started the repo
- set a date for next session and a practice day
- we send the message next Friday

Next repo session
- make a template repo
- add links to MX repo and make a repo for SM

For me: ask Nik if she wants to present

### Fri 21st July
- We reviewed Harvey's presentation on text-to-image models
- Miguel will prep poster and circulate invite on Monday morning
- Shall we agree on a template for future presentations (UCL template?)

### Fri 27th Oct
- Miguel and Sofia (Harvey we miss you)
- - post for a new organiser? - maybe next time we meet (7 or 14th Nov) y mandar 3 dates
- estimated 1st-2nd week of Dec session on pose estimation (focus on animals)
- we can use this chance to make a template for the slides and add it to the repo
- [Miguel's guide to create slides in html](https://github.com/oss-for-surgtech/ucl-open-science-awards-2023/tree/main/slides)

- Other topics we could discuss in the future?
- mujoco?
- Actions
- send to Miguel 3 dates for next session (for prep, session and tutorial) 7th or 14th Nov
- post for a new organiser in the slack channel then?

## References
### Journals
* Journal of Vision: https://jov.arvojournals.org/; https://twitter.com/ARVOJOV
* The Journal of Open Source Software: https://joss.theoj.org/

### Blogs
* AK tweeter with latest AI/ML papers: https://mobile.twitter.com/_akhaliq
* TwoMinutePapers https://www.youtube.com/channel/UCbfYPyITQ-7l4upoX8nvctg

### Github sites for Journal clubs/seminars
* Github repositories of Journal clubs : https://github.com/inform-health-informatics/journal-club
* https://ucl-ellis.github.io/dm_csml_seminar_home/ https://github.com/ucl-ellis/ucl-ellis.github.io https://twitter.com/uclcsml https://ucl-ellis.github.io/dm_csml_seminars/2023-02-03-Ye/

### Other references
* http://blog.thegrandlocus.com/2022/03/journal-clubs-ranked-from-worst-to-best
* https://www.benty-fields.com/
* https://paperswithcode.com/
* https://danielskatzblog.wordpress.com/
* https://www.benty-fields.com/
* https://colinraffel.com/blog/role-playing-seminar.html
* Papers We Love: https://github.com/papers-we-love/papers-we-love; https://www.youtube.com/@PapersWeLove
* Yannic Kilcher: https://www.youtube.com/@YannicKilcher/videos
* https://www.rc.ucl.ac.uk/docs/Clusters/Myriad/

### Annoucements of seminars, journal clubs
* https://twitter.com/RadNetCRUKCoL/status/1454094472808177669?cxt=HHwWioC5mYHO_a0oAAAA

### Seminars and journal-clubs using github-pages
* https://github.com/MexicanSocietyUoB/seminars
* https://github.com/uf-repro/uf-repro.github.io
* https://github.com/melbintgen/statgen-journal-club
* https://github.com/cambridge-ceu/CEU-journal-club
* https://github.com/dlsun/ds-seminar
* https://github.com/mariabnd/journal_club

### Others:
* List of talks: https://github.com/girafe-ai/journal-club; https://github.com/mbbu/Journal_Club
* Reproducible presentation in html: https://github.com/sschmutz/computational_life_sciences_seminar_presentation
* Reproducible presentation in html: https://github.com/kelly-sovacool/bioinf603-fall2019
* Templates for conferences: https://github.com/mikepierce/conference-website-template

## Survey
* Document to [survey](https://liveuclac-my.sharepoint.com/:x:/g/personal/ccaemxo_ucl_ac_uk/EWs0hpAlRmRIkX28pkrNJeYBBI2CoSJcTjXnFw3mHR9U7w?e=ye2UJg) the interest of ARC community to join JC4RSEs. [Preview](https://forms.office.com/e/ZSXpeMxRr6) ----> SM: I would just clarify here we are focusing on DL + CV
* Slack post:
* to gauge interest and how broad we can go
* select between: DL, CV, Image Processing, all of them

## Blurs
* Create an space to organise JC4RSE (e.g., https://www.youtube.com/watch?v=3v2MlfgkwUY, repo?)
* ...