Explore tweets tagged as #Sparse
🚨Paper Alert 🚨. ➡️Paper Title: Diffuman4D: 4D Consistent Human View Synthesis from Sparse-View Videos with Spatio-Temporal Diffusion Models. 🌟Few pointers from the paper. 🎯This paper addresses the challenge of high-fidelity view synthesis of humans with sparse-view videos as
0
0
0
We will be presenting Self-Distilled Sparse Drafters at #ICML this afternoon at the Efficient Systems for Foundation Models workshop. Come chat with us about acclerating speculative decoding! 🚀. @CerebrasSystems @UCalgaryML
0
4
8
6. Mistral - The Efficient Rebel. Small but mighty. Mistral 7B beats much larger models using:. → Grouped-query attention.→ Sliding windows.→ Sparse MoE in Mixtral. Open, fast, and crushing benchmarks with fewer parameters.
2
0
9
Announcing our EXAIT@ICML workshop paper: CURATE!. Have a difficult target task distribution with sparse rewards that you want to train an RL agent to complete? 🤔. We tackle this problem using our curriculum learning algorithm, CURATE. 🎓. Link: 1/6
1
6
19
🚨 What happens under the hood when you fine-tune or unlearn LLMs?. We introduce MNEME—a sparse model diffing method that predicts & explains unintended side effects (like emergent toxicity or forgotten knowledge) without needing fine-tuning data. @ActInterp #ICML July 19 Poster
1
1
6
Diffuman4D: 4D Consistent Human View Synthesis from Sparse-View Videos with Spatio-Temporal Diffusion Models. Contributions:.• We introduce Diffuman4D, a novel diffusion model that generates spatio-temporally consistent and high-resolution (1024p) human videos from sparse-view
5
55
460
A comprehensive and nicely delivered tutorial by @macavaney et al. on learned sparse retrieval methods and their applications across various tasks. Slides and other materials are available via the link shown in the last image. #SIGIR2025
1
3
21
Checkout our new paper: Video-RTS 🎥.A data-efficient RL method for complex video reasoning tasks. 🔹 Pure RL w/ output-based rewards. 🔹 Novel sparse-to-dense Test-Time Scaling (TTS) to expand input frames via self-consistency. 💥 96.4% less training data!.More in the thread👇.
🚨Introducing Video-RTS: Resource-Efficient RL for Video Reasoning with Adaptive Video TTS! . While RL-based video reasoning with LLMs has advanced, the reliance on large-scale SFT with extensive video data and long CoT annotations remains a major bottleneck. Video-RTS tackles
0
3
8
5. Gemini - The Multimodal Beast. The evolution of PaLM. Gemini understands text, images, code, and audio all in one model. → Massive context windows.→ Sparse expert layers.→ Enterprise-ready and phone-capable. This is Google's bet on the AI future.
2
0
9
🚨 We just released 🎞️MoVieS — a feed-forward model that reconstructs 4D scenes in ⚡️1 second. My favorite part: It learns dense (pixel-wise) sharp 3D world movements from novel view rendering + sparse point tracking supervision 🤯🎯. Check it out 👉
2
32
160
Content will likely be sparse the next couple of days. My latest attempts to tackle insomnia have failed miserably, and I'll likely get even less sleep for a few days.
2
0
7
#Garlasco ma quindi non sono mai state analizzate le mutandine sporche sparse per casa? 4 erano nella busta sul divano. Due sulla scrivania di Chiara, due sul lavandino del bagno, due sul bordo vasca. Mai state analizzate? Assurdo, assurdo. La camicia rossa? Mai analizzata.
9
0
14
5. Tibetan Cavalry Armor (18th–19th c.). Leather, iron, and wool made for high-altitude warfare. Sparse but effective, it mixed mobility with spiritual symbolism.
1
9
287
An MU-led project to rescue African climate data has taken a major step forward in its mission to recover vital historical weather data from data-sparse regions of the continent.
0
1
1
🌌🛰️🔭Want to explore universal visual features? Check out our interactive demo of concepts learned from our #ICML2025 paper "Universal Sparse Autoencoders: Interpretable Cross-Model Concept Alignment". Come see our poster at 4pm on Tuesday in East Exhibition hall A-B, E-1208!
🌌🛰️Wanna know which features are universal vs unique in your models and how to find them? Excited to share our preprint: "Universal Sparse Autoencoders: Interpretable Cross-Model Concept Alignment"! . (1/9)
3
46
230
5. Gemini - The Multimodal Beast. The evolution of PaLM. Gemini understands text, images, code, and audio all in one model. → Massive context windows.→ Sparse expert layers.→ Enterprise-ready and phone-capable. This is Google's bet on the AI future.
1
0
8
Causal Language Control in Multilingual Transformers via Sparse Feature Steering.
0
0
2