↓Skip to main content

🏢 NVIDIA

QueST: Self-Supervised Skill Abstractions for Learning Continuous Control
·2477 words·12 mins· loading · loading
AI Applications Robotics 🏢 NVIDIA
QueST: A novel self-supervised skill abstraction architecture for continuous robot control, achieves state-of-the-art performance on multitask and few-shot learning benchmarks by learning flexible, tr…
Memorize What Matters: Emergent Scene Decomposition from Multitraverse
·3662 words·18 mins· loading · loading
3D Vision 🏢 NVIDIA
3D Gaussian Mapping (3DGM) achieves self-supervised camera-only 3D scene decomposition by leveraging multi-traverse driving data, memorizing permanent structures while filtering out transient objects.
MaskLLM: Learnable Semi-Structured Sparsity for Large Language Models
·2759 words·13 mins· loading · loading
Large Language Models 🏢 NVIDIA
MaskLLM learns efficient semi-structured sparsity in LLMs via end-to-end training, achieving significant speedup and memory reduction without sacrificing performance.
Guiding a Diffusion Model with a Bad Version of Itself
·1887 words·9 mins· loading · loading
Image Generation 🏢 NVIDIA
Boost image quality in diffusion models without reducing variation using Autoguidance: guide a high-quality model with a less-trained version of itself!
Diffusion-Reward Adversarial Imitation Learning
·2028 words·10 mins· loading · loading
Machine Learning Reinforcement Learning 🏢 NVIDIA
Diffusion-Reward Adversarial Imitation Learning (DRAIL) enhances Generative Adversarial Imitation Learning by integrating diffusion models, resulting in more stable and smoother reward functions for s…
Compact Language Models via Pruning and Knowledge Distillation
·4214 words·20 mins· loading · loading
AI Generated Natural Language Processing Large Language Models 🏢 NVIDIA
MINITRON: Efficiently creating smaller, high-performing LLMs via pruning & distillation, slashing training costs by up to 40x!
ChatQA: Surpassing GPT-4 on Conversational QA and RAG
·4802 words·23 mins· loading · loading
AI Generated Natural Language Processing Question Answering 🏢 NVIDIA
ChatQA, a new suite of models, outperforms GPT-4 in conversational QA and RAG by using a two-stage instruction tuning method and a cost-effective dense retriever.