Vicente Ordóñez Profile
Vicente Ordóñez

@bluevincent

Followers
4K
Following
2K
Media
85
Statuses
972

Please refer to @vislang

Joined April 2008
Don't wanna be here? Send us removal request.
@bluevincent
Vicente Ordóñez
3 years
RT @AmazonScience: Learn more about @alexa99's new story creation experience, which receives customer prompts and uses AI to generate short….
0
3
0
@bluevincent
Vicente Ordóñez
3 years
RT @RiceEngineering: With the introduction of SimVQA at #CVPR2022, @RiceCompSci PhD student Paola Cascante-Bonilla and the VisLang Lab at @….
0
4
0
@bluevincent
Vicente Ordóñez
3 years
I am wondering what will be the academic papers at this year #CVPR2022 that show promising results that will have astounding results as a matter of scaling them up in #CVPR2025 — I will report here — Looking forward.
0
0
8
@bluevincent
Vicente Ordóñez
3 years
When we came up with this paper most work was focused on GANs — an even earlier auto regressive model for text to image synthesis was also at #CVPR2018 — full 4 years earlier.
1
0
3
@bluevincent
Vicente Ordóñez
3 years
Our paper was selected as Oral and Best Paper Finalist at #CVPR2019 — it would be A LOT more competitive these days! .
1
0
2
@bluevincent
Vicente Ordóñez
3 years
Now a new model by Google scaling this up further.
@lmthang
Thang Luong
3 years
The recipe is simple: better image tokenizer (ViT-VQGAN) + scaling the visual component (that generates image tokens from text tokens). The beauty of this: all components are Transformer & any techniques that work for large language models are readily applicable! (2/n)
Tweet media one
1
0
4
@bluevincent
Vicente Ordóñez
3 years
Great explanation here for how DALL-E mini:mega works which in turns is how the original DALL-E worked. // Our Text2Scene model assumed visual tokens are actual object crops from the training data from a fixed (not learned) dictionary.
Tweet media one
1
0
3
@bluevincent
Vicente Ordóñez
3 years
Three years ago we showed seq-to-seq models could be used for text-to-image synthesis. In some way I’m glad to see new models are emerging in 2021/2022 where this idea is scaled up and combined with visual dictionary learning (ie. VQ-VAE, VQ-GAN, dGAN).
@bluevincent
Vicente Ordóñez
6 years
Look Mom No GANs: Text-to-image Synthesis as Machine Translation // Not the actual title of our paper but maybe should be. To be presented next week at #cvpr2019 in Long Beach, California.
1
1
22
@bluevincent
Vicente Ordóñez
3 years
In today's afternoon session @pcascanteb will be presenting our work in the Vision and Language poster track: Poster 242b #CVPR2022 -- I will be around to network/talk to people while she presents -- Join us!.
@pcascanteb
Paola Cascante-Bonilla
3 years
Thank you @ak92501 for having us! .This is my first #CVPR and also part of my first internship in industry @MITIBMLab, @IBMResearch!.Thanks for the great advice @HuiWu_, @RogerioFeris and my Professor @bluevincent, and the amazing work of undergrad student @DiegoWang51! 1/4.
0
2
8
@bluevincent
Vicente Ordóñez
3 years
Rice-UVA Vision Meetup at #CVPR2022
Tweet media one
0
1
38
@bluevincent
Vicente Ordóñez
3 years
Never met him but followed his work since I was student — pre-deep learning he was most well known for his work on detecting salient objects and haze removal. Rest in peace.
@YiMaTweets
Yi Ma
3 years
I was shocked to know that Dr. Jian Sun, my former colleague of the MSRA Visual Computing Group, has passed away. We will miss him dearly. May his soul rest in peace.
0
0
6
@bluevincent
Vicente Ordóñez
3 years
RT @pcascanteb: Thank you @ak92501 for having us! .This is my first #CVPR and also part of my first internship in industry @MITIBMLab, @IBM….
0
6
0
@bluevincent
Vicente Ordóñez
3 years
I was also wondering if Duolingo really works? I am still not sure if it’s the most efficient way but with persistence and for some languages it takes one very far. Add tutoring and media consumption in the target language and a high level of proficiency can certainly be achieved.
1
0
4
@bluevincent
Vicente Ordóñez
3 years
About four years ago I decided that I needed to remind myself of the uncomfortable stage of being clumsy in a foreign language. By now, I am fully reminded of how it feels. I added a new language to my skills and now toward a fourth.
Tweet media one
2
0
28
@bluevincent
Vicente Ordóñez
3 years
RT @VMFragoso: Are you a Hispanic/Latinx student in the US? . Microsoft is offering an exciting scholarship for you!. Check it out here👉 ht….
0
25
0
@bluevincent
Vicente Ordóñez
3 years
These are the moments that make academia worth it for me. Hearing about success stories from former students. Follow the work of a startup using computer vision to help decorate your house
0
0
19
@bluevincent
Vicente Ordóñez
3 years
Last week my Twitter feed was all about AI/consciousness and this week is all about tenure. But no, I will not comment and surrender to the pressure of our recommendation and engagement AI engine overlords. Nice try.
0
0
9
@bluevincent
Vicente Ordóñez
3 years
RT @marktenenholtz: When you learn new concepts in machine learning, try to understand them on an intuitive level. This blog from Distill….
0
187
0
@bluevincent
Vicente Ordóñez
4 years
RT @RiceEngineering: 📢 FACULTY OPENINGS: #RiceEngineering seeks outstanding candidates to help carry out our mission of groundbreaking rese….
0
3
0