When browsing through posters in the visualizer, it is currently hard to know whether you've checked something out already or not. In particular, when trying to decide which posters to click, it's not easy to "mark as seen" to indicate that a given dot was acknowledged, but that I decided I'm not interested in it.

This could be done in a variety of ways. Off the top of my head, one option would be to "right click" to mark as read, which will change the color of the dot to indicate it was been "read"

Another unrelated feature that will be nice to have is a grid underneath the visualization, which would alleviate the problem of not knowing which dots I've already hovered over and which ones need my attention.

In speakers.html, the image of Profs. Yoshua Benjio and Yann LeCun is missing due to missing image source in speakers.yml. Also, their affiliations are interchanged.

Possible fix:

    - speaker: Profs. Yann LeCun and Yoshua Benjio
      image: static/lecun+bengio.png
      title: Reflections from the Turing Award Winners
      institution: Facebook AI + NYU / MILA

Rendered page:
The uniformity of the page is lost though.

The image used (from Voices of ICLR2020: Announcing our Keynote Speakers Medium Blog)

If this is fine, I can submit a pr.

Clean up the style of the homepage and add some content.

We have customized recs for each author. Need a non-creepy way to integrate into the static site.

We should have documentation for each of the files used for the data storage on the website.

Group together the keywords into categories.

Currently these links are broken on paper page if they are not there.

For the PCs, we need to fill in the FAQs

Hi,

there are no keywords for the poster https://iclr.cc/virtual_2020/poster_B1lDoJSYDH.html

The poster cannot be found via keyword search.
Can the keywords be added for the public version of the website?

Best,
Benjamin

Shouldn't there be an "index.html" template?

Once there are available zoom links we need to wire them up to the poster pages and the main talk page.

Once times are established make it easy to browse by timeline.

The generation of ICS events with icalendar

I'm not sure why icalendar doesn't do that for dtstart and dtend, but does it for dtstamp, and maybe there's a good reason for it. If all else failed, one could hackily append the Z, but I'm sure there's a principled thing to do...

Check mobile for all pages / integrate better with bootstrap CSS.

Let users star the posters that they want to see to come back to them.

We should include bios for the speakers and area specific times for their talks.

We need a button at the top of the sponsor page (and perhaps in other locations) that allows people to get to the sponsor booth site on 6connex. The simplest solution is a button at the top of the sponsor page that say 'Visit Sponsor Hall' that then opens a new tab/window.

Document process for authors to change their papers images.

Hi @srush Amazing work with ICLR2020. I enjoyed the conference a lot. :)

I was looking into the visualizations and the code for generating the embeddings using the paper abstracts. I am unable to load the pre-trained model you shared here When I try to run the cache_paper_embeddings.py file, I face the following error:

Loading model
Traceback (most recent call last):
  File "cache_paper_embeddings.py", line 31, in <module>
    model, epoch = load_model(None, "scratch/similarity-model.pt")
  File "/Users/gollum/Desktop/Work/iclr.github.io/recommendations/models.py", line 20, in load_model
    model = torch.load(load_file, map_location=lambda storage, loc: storage)
  File "/Users/gollum/anaconda3/lib/python3.7/site-packages/torch/serialization.py", line 529, in load
    return _legacy_load(opened_file, map_location, pickle_module, **pickle_load_args)
  File "/Users/gollum/anaconda3/lib/python3.7/site-packages/torch/serialization.py", line 692, in _legacy_load
    magic_number = pickle_module.load(f, **pickle_load_args)
_pickle.UnpicklingError: invalid load key, 'v'.

This is imo problem with the saved model keys. If possible can you share steps to run the code for training this model?

Thanks!

Page with the PCs / Workshop chairs etc.