Explore · 每周精选
发现最受关注的研究论文,追踪研究趋势,订阅感兴趣的期刊与关键词。
Anton Razzhigaev, Arseniy Shakhmatov, Anastasia Maltseva +7 authors
Kandinsky1, a latent diffusion architecture, achieves high-quality text-to-image generation by integrating image prior models and modified MoVQ autoencoders, outperforming other open-source models.
21 篇论文 · 按点赞排序
Junsong Chen, Jincheng Yu, Chongjian Ge +8 authors
PIXART-$\alpha$ is a Transformer-based text-to-image diffusion model that achieves high-quality image synthesis with low training costs using advanced training strategies and efficient architecture.
Haotian Liu, Chunyuan Li, Yuheng Li +1 authors
Modifications to LLaVA using CLIP-ViT-L-336px and academic VQA data improve multimodal performance, achieving state-of-the-art results with limited training data.
Omar Khattab, Arnav Singhvi, Paridhi Maheshwari +10 authors
DSPy is a programming model for developing and optimizing language model pipelines by abstracting them as text transformation graphs with parameterized modules that learn through demonstrations.
Jie Huang, Xinyun Chen, Swaroop Mishra +4 authors
This paper evaluates the intrinsic self-correction capabilities of Large Language Models and identifies limitations in their ability to improve responses without external feedback.
Pei Zhou, Aman Madaan, Srividya Pranavi Potharaju +9 authors
LLMs perform well in inferring beliefs but struggle translating these into actions; Foresee and Reflect (FaR) improves their action guidance in social scenarios.
Ke Wang, Houxing Ren, Aojun Zhou +7 authors
A method is presented for fine-tuning language models to solve math problems using code, achieving state-of-the-art results on MATH and GSM8K datasets.
Ziqi Wang, Le Hou, Tianjian Lu +4 authors
A new framework, ImPlicit Self-ImprovemenT (PIT), uses human preference data to implicitly enhance the quality of Large Language Model responses, outperforming prompting-based methods.
Mihir Prabhudesai, Anirudh Goyal, Deepak Pathak +1 authors
AlignProp refines text-to-image diffusion models using backpropagation through the denoising process, leveraging low-rank adapters and gradient checkpointing to optimize for various objectives with higher efficiency.
Dongchao Yang, Jinchuan Tian, Xu Tan +9 authors
UniAudio, a unified language model system, generates various audio types effectively using tokenization, sequence concatenation, and next-token prediction, with a multi-scale Transformer model handling extended sequences.
Tu Vu, Mohit Iyyer, Xuezhi Wang +8 authors
FreshQA evaluates LLMs' factuality and hallucination in responding to current knowledge questions, revealing limitations and demonstrating improvements through FreshPrompt, an augmented prompting method using search engine data.
Max Ku, Tianle Li, Kai Zhang +4 authors
ImagenHub standardizes inference and evaluation for conditional image generation models, revealing inconsistencies and highlighting areas needing improvement.
Kangfu Mei, Mauricio Delbracio, Hossein Talebi +3 authors
A novel single-stage distillation method for generative diffusion models reduces sampling time while maintaining performance across tasks like super-resolution and image editing.
Michihiro Yasunaga, Xinyun Chen, Yujia Li +5 authors
Analogical Prompting enhances language model performance in reasoning tasks by self-generating relevant exemplars without labeled data.
Tianhong Li, Sangnie Bhardwaj, Yonglong Tian +6 authors
The ITIT model uses cycle consistency to train vision-language generative models on unpaired data, achieving performance comparable to those trained on large labeled datasets with significantly less paired data.
Murong Yue, Jie Zhao, Min Zhang +2 authors
An LLM cascade system reduces costs by using a less expensive LLM for simpler tasks and a more expensive LLM for challenging tasks, maintaining performance through answer consistency checks.
Yue Wu, Xuan Tang, Tom M. Mitchell +1 authors
SmartPlay is a benchmark and evaluation methodology for large language models as agents, featuring 6 games that test various capabilities like reasoning, planning, spatial awareness, and learning.
Prasann Singhal, Tanya Goyal, Jiacheng Xu +1 authors
Optimizing for response length significantly contributes to the improvements observed in Reinforcement Learning from Human Feedback (RLHF) when aligning large language models for helpfulness tasks.
Zhiwen Fan, Panwang Pan, Peihao Wang +6 authors
DragView is an interactive framework that generates novel views of unseen scenes using a single source image and sparse unposed multi-view images, utilizing an epipolar attention mechanism and transformer-based ray decoding without estimating camera poses.
Paloma Sodhi, S. R. K. Branavan, Ryan McDonald
Stacked LLM Policies for Web Actions dynamically compose policies to handle diverse web tasks, improving performance over baseline methods.
Jieyu Zhang, Ranjay Krishna, Ahmed H. Awadallah +1 authors
EcoAssistant framework enables more affordable and accurate responses from LLMs through iterative code refinement, LLM hierarchy, and solution retrieval.
北京市昌平区探索星信息技术及软件开发工作室
京ICP备2026059466号