Researchers from Google DeepMind, including Dhruva Tirumala and @BarnesMJ, have announced the acceptance of their papers at ICLR 2024. The papers focus on topics such as lifelong learning for RL, world-scale models for IRL, and a SandBox for fine-grained alignment data at scale. InfoBatch, another accepted work, achieves 20%~40% speedups in tasks like image classification and semantic segmentation. Additionally, advancements in embodied AI have been recognized with five submissions accepted and two spotlights awarded. The researchers have also released data, code, and models, and provided links to ArXiv, a blog, and code repositories.
Check out our #ICLR2024 Accepted Papers. Congratulations to all of our authors! https://t.co/ktxNWIxghc
Excited to share that we have 5 submissions on embodied AI accepted at #ICLR2024 with two spotlights! Through these works, we improve the versatility, generalizability, effectiveness, and training efficiency of embodied AI models. Check them out in the thread below:
InfoBatch is accepted as Oral to ICLR'24! ๐ฅ InfoBatch prunes data on the fly and speedups 20%~40% on img classification, semantic segmentation, MAE, Diffusion, LLM instruction tunning.๐งต3 ArXiv: https://t.co/3RKELMonnd Blog: https://t.co/tbLZMP7y8W Code: https://t.co/C1tznhNMnb https://t.co/Vyz2OHhQzH
Now Stable Alignment is accepted to #ICLR2024! We developed a SandBox to obtain fine-grained alignment data at scale, and used simple contrastive learning to train models! We have released data/code/models. Please try it if you are interested! https://t.co/w9yoRD4zhO
Great experience with #ICLR2024! Pure joy of working with our team @GoogleDeepMind leading to two accepted papers! Fantastic work by Dhruva Tirumala & @BarnesMJ! Something for #RL (lifelong learning) and for #IRL (world-scale models!!); both heavily data-centric! Thread ๐งต๐