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。
598 篇论文 · 按点赞排序
Long Xing, Qidong Huang, Xiaoyi Dong +8 authors
PyramidDrop, a novel strategy for reducing visual token redundancy in large vision-language models, enhances training and inference efficiency with minimal performance loss.
Lucas Lehnert, Sainbayar Sukhbaatar, Paul Mcvay +2 authors
Searchformer, a Transformer model, outperforms traditional $A^*$ search by optimizing Sokoban puzzle solutions with fewer steps and demonstrates superior performance on maze navigation and more complex tasks.
David Grangier, Angelos Katharopoulos, Pierre Ablin +1 authors
The study evaluates various machine learning approaches under constrained inference and training budgets, demonstrating that hyper-networks and mixture of experts perform well with large pretraining budgets, while small models trained on importance sampled datasets are better with large specialization budgets.
Xudong Lu, Yinghao Chen, Cheng Chen +19 authors
BlueLM-V-3B optimizes the deployment of multimodal large language models on mobile platforms through redesigned dynamic resolution and hardware-aware system optimizations, achieving high performance with a small model size and fast generation speed.
Xiangxiang Chu, Jianlin Su, Bo Zhang +1 authors
A LLaMA-like vision transformer, VisionLLaMA, is introduced for processing 2D images and achieves state-of-the-art performance in various vision tasks.
Ruiqi Gao, Aleksander Holynski, Philipp Henzler +5 authors
CAT3D generates consistent 3D views from limited input using a multi-view diffusion model, enabling fast and accurate 3D reconstruction.
Wenqi Zhang, Zhenglin Cheng, Yuanyu He +8 authors
A multi-modal self-instruct benchmark using large language models and code capabilities enhances understanding and reasoning for abstract images and visual tasks, improving performance metrics like chart understanding and map navigation.
Jinghuan Shang, Karl Schmeckpeper, Brandon B. May +4 authors
Theia, a vision foundation model for robot learning, outperforms existing models with less data and smaller sizes by distilling knowledge from multiple vision foundation models.
Seungone Kim, Juyoung Suk, Xiang Yue +7 authors
AgoraBench evaluates language models' data generation abilities, revealing their distinct strengths and demonstrating various factors that impact effectiveness.
Zijian Zhang, Kaiyuan Zheng, Zhaorun Chen +6 authors
GRAPE improves vision-language-action models' performance by aligning policies via preference modeling, enhancing generalizability and allowing customization of objectives like safety and efficiency.
Zihan Wang, Deli Chen, Damai Dai +3 authors
Research on parameter-efficient fine-tuning for Large Language Models with Mixture-of-Experts architecture reveals that task-specific expert tuning can enhance performance and efficiency.
Huajian Xin, Daya Guo, Zhihong Shao +6 authors
Generating extensive Lean 4 proof data from mathematical competition problems improved DeepSeekMath 7B's theorem-proving capabilities over GPT-4 and other methods.
Moo Jin Kim, Karl Pertsch, Siddharth Karamcheti +15 authors
OpenVLA, a 7B-parameter open-source vision-language-action model, demonstrates strong performance in generalist manipulation and efficient fine-tuning for new tasks, outperforming larger closed models and from-scratch imitation learning methods.
Jinhyuk Lee, Zhuyun Dai, Xiaoqi Ren +17 authors
Gecko is a compact and versatile text embedding model that distills knowledge from large language models to achieve strong retrieval performance on benchmarks.
Taekyung Ki, Dongchan Min, Gyoungsu Chae
FLOAT method generates high-quality, temporally consistent, and emotion-enhanced talking portraits using flow matching in a learned motion latent space with a transformer-based vector field predictor.
Gongfan Fang, Hongxu Yin, Saurav Muralidharan +5 authors
MaskLLM introduces a learnable pruning method using Gumbel Softmax sampling to achieve N:M sparsity in LLMs, improving inference efficiency and transferability across tasks.
Zhiyong Wu, Chengcheng Han, Zichen Ding +5 authors
OS-Copilot framework enables FRIDAY, a self-improving generalist agent, to automate a wide range of computer tasks with enhanced generalization and minimal supervision.
Zhen Huang, Haoyang Zou, Xuefeng Li +7 authors
Knowledge distillation from O1 model combined with fine-tuning achieves superior performance on complex mathematical reasoning and generalizes to open-ended QA tasks, promoting transparency in AI research.
Lifan Yuan, Ganqu Cui, Hanbin Wang +12 authors
Eurus, a suite of reasoning-optimized large language models, achieves state-of-the-art performance on various benchmarks through UltraInteract, a large-scale, high-quality alignment dataset, and a novel reward modeling objective.
Jarvis Guo, Tuney Zheng, Yuelin Bai +7 authors
A large-scale multimodal instruction-tuning dataset with intermediate rationales improves reasoning capabilities in large language models, achieving state-of-the-art performance across benchmarks.
Tingxu Han, Chunrong Fang, Shiyu Zhao +3 authors
A token-budget-aware framework dynamically estimates and allocates token budgets for LLM reasoning, reducing costs with minimal performance loss.
Chien Van Nguyen, Xuan Shen, Ryan Aponte +25 authors
A survey of small language models covers their architectures, training methods, and model compression techniques, with a taxonomy for optimization, benchmark datasets, evaluation metrics, and open challenges.
Tiannan Wang, Jiamin Chen, Qingrui Jia +43 authors
Weaver, a family of specialized large language models, achieves superior writing capabilities through pre-training and fine-tuning methods, surpasses GPT-4 in various writing tasks, and supports retrieval-augmented generation and tool usage.
Xiao Yang, Kai Sun, Hao Xin +24 authors
The Comprehensive RAG Benchmark (CRAG) evaluates RAG solutions with diverse QA tasks, revealing gaps in LLM accuracy and future research directions.
Chunjiang Ge, Sijie Cheng, Ziming Wang +6 authors
ConvLLaVA addresses excessive visual tokens and quadratic complexity in high-resolution multimodal models by using ConvNeXt as a hierarchical backbone, optimizing pretrained ConvNeXt for high resolution, and adding a successive stage for further compression, achieving competitive performance.
Wei Huang, Xudong Ma, Haotong Qin +7 authors
LLaMA3 performance degrades significantly when quantized to low bit-width, highlighting challenges in low-bit quantization for large language models.
Sumeet Ramesh Motwani, Chandler Smith, Rocktim Jyoti Das +6 authors
Multi-agent LLM training improves performance on reasoning tasks by assigning specialized roles and utilizing joint outcome-based rewards to enhance collaboration among models.
Zichen Liu, Yihao Meng, Hao Ouyang +4 authors
Dynamic Typography generates coherent and semantically meaningful text animations using neural displacement fields, end-to-end optimization, and perceptual loss regularization, outperforming baseline methods.
Qichen Fu, Minsik Cho, Thomas Merth +3 authors
LazyLLM dynamically selects important prompt tokens to accelerate the prefilling stage in transformer-based language models without sacrificing accuracy.
Yibin Wang, Zhiyu Tan, Junyan Wang +3 authors
LiFT fine-tunes text-to-video generative models using human feedback to improve video alignment with text descriptions, demonstrating superior performance over a larger model.
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