Paramanand Chandramouli

[vaishnavi gandikota @kvgandikota]

Happening now in #CVPR 2024 poster session 6, poster number 170, text guided explorable super-resolution

[vaishnavi gandikota @kvgandikota]

I am at #CVPR in Seattle this week to present our paper on Text-guided Explorable Image Super-resolution! Joint work with @ParamChandram . Please stop by our poster at the final Poster Session on Friday, June 21. Looking forward to discussing and building connections!

[Michael Bronstein @mmbronstein]

Impressive

[AI Conference DL Countdown @DlCountdown]

MICCAI'24 abs: 5 days.
ECCV'24 abs: 12 days.
RLC'24: 13 days.
IROS'24: 13 days.
ECCV'24 paper: 19 days.
MICCAI'24 paper: 19 days.
MM'24 abs: 50 days.
BMVC'24 abs: 69 days.
MM'24 paper: 54 days.
ECAI'24 abs: 61 days.
ECAI'24 paper: 67 days.
BMVC'24 paper: 76 days.

[Jonas Geiping @jonasgeiping]

We're also looking to hire PhD students interested in machine learning! You can find more information about joining my group on my webpage: https://t.co/sZ4uqcs7I5

[AI Conference DL Countdown @DlCountdown]

CVPR'24 abs: 19 days.
CVPR'24 paper: 26 days.
ECCV'24 abs: 137 days.
ECCV'24 paper: 144 days.

[Paramanand Chandramouli @ParamChandram]

Reposting our work on exploring solutions to super-resolution. We make use of open-sourced version of DALLE-2, Karlo-unCLIP by @diffusers to achieve text guided super-resolution. Check out our initial results at
https://t.co/xbkL4CoaGT
More coming soon @_akhaliq

[vaishnavi gandikota @kvgandikota]

How can we explore solutions to super-resolution through text? In our recent work we attempt to address this problem using pretrained text to image diffusion models. Joint work with @ParamChandram accepted at ICML workshop on AI&HCI https://t.co/wPJ9vsXr5Q

[Paramanand Chandramouli @ParamChandram]

@_akhaliq @huggingface @ICCVConference It was great talking to you.

[Shashank Agnihotri @shashankska]

In this work: On the unreasonable vulnerability of transformers for image restoration and an easy fix, we explore ViT-based models for restoring real-world images and expose the vulnerability of such models to adversarial perturbations. Additionally, we propose an easy fix. #AROW

[Shashank Agnihotri @shashankska]

Very excited about the @ICCVConference
Looking forward to meeting everyone on Monday, 02 October at the #AROW workshop on Day 1 of ICCV2023, Paris.
At #AROW I will be presenting a poster on our recent work done with Kanchana, @JuliaGrabinski, @ParamChandram, and @margret_keuper

[Vincent Sitzmann @vincesitzmann]

Introducing “Diffusion with Forward Models”, 𝗮 𝗺𝗼𝗱𝗲𝗹 𝘁𝗵𝗮𝘁 𝗰𝗮𝗻 𝗴𝗲𝗻𝗲𝗿𝗮𝘁𝗲 𝗱𝗶𝘃𝗲𝗿𝘀𝗲, 𝗿𝗲𝗮𝗹 𝟯𝗗 𝘀𝗰𝗲𝗻𝗲𝘀 𝗳𝗿𝗼𝗺 𝗮 𝘀𝗶𝗻𝗴𝗹𝗲 ���𝗺𝗮𝗴𝗲, 𝘁𝗿𝗮𝗶𝗻𝗲𝗱 𝘄𝗶𝘁𝗵 𝗶𝗺𝗮𝗴𝗲𝘀 𝘄/𝗼 𝗮𝗻𝘆 𝟯𝗗 𝗱𝗮𝘁𝗮!

https://t.co/8CbXGME46Q

1/n

[vaishnavi gandikota @kvgandikota]

How can we explore solutions to super-resolution through text? In our recent work we attempt to address this problem using pretrained text to image diffusion models. Joint work with @ParamChandram accepted at ICML workshop on AI&HCI https://t.co/wPJ9vsXr5Q

[Walter Hugo Lopez Pinaya 🍍 @Warvito]

Discover the latest breakthroughs in medical generative AI with our package (https://t.co/AKUe0nEG3O) and explore the code for the experiments at our GitHub repository: https://t.co/3YkStY7t3y. 🔬🔍 #AIresearch #GenerativeAI #OpenSource

[Hyungjin Chung @hyungjin_chung]

Great work by many collaborators🎉🎉 if interested in generative models + medical imaging, definitely check it out!

[Volodymyr Kuleshov 🇺🇦 @volokuleshov]

Diffusion models produce great samples, but they lack a semantically meaningful latent space like in a VAE or a GAN. We augment diffusion with low-dimensional latents that can be used for image manipulation, interpolation, controlled generation. #ICML2023

[Prof. Anima Anandkumar @AnimaAnandkumar]

Please include our genome-scale language model with @arvindr_ @argonne which is open access. Trained on 110 million genomes of viruses and bacteria and able to predict new variants of concern for #SARSCoV2 and other viruses and bacteria. https://t.co/sGgLTHmtVF

[Li Junnan @LiJunnan0409]

🔥BLIP-Diffusion🔥 is now open-sourced! We release code and models for zero-shot subject-driven image generation and stylization. Here are some Pikachus I generated by transferring the styles from other animals. Can you guess what animals are they?
Github: https://t.co/47vkSuifJR

[Sander Dieleman @sedielem]

New blog post: perspectives on diffusion, or how diffusion models are autoencoders, deep latent variable models, score function predictors, reverse SDE solvers, flow-based models, RNNs, and autoregressive models, all at once!

https://t.co/sBG8Waa9Ql

[Kyle Sargent @KyleSargentAI]

Accepted to ICCV2023!

[Matt Deitke @mattdeitke]

Introducing Objaverse-XL, an open dataset of over 10 million 3D objects!

With it, we train Zero123-XL, a foundation model for 3D, observing incredible 3D generalization abilities: ����👇

📝 Paper: https://t.co/2oNakoka7v