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。
50 篇论文 · 按点赞排序
Mingzhe Zheng, Yongqi Xu, Haojian Huang +8 authors
VGoT generates high-quality, coherent multi-shot videos by structuring the process into script, keyframe, and shot-level generation, using IP embeddings and smoothing mechanisms for consistency and narrative flow.
Junjie Zhou, Zheng Liu, Ze Liu +6 authors
MegaPairs, a data synthesis method using vision language models and open-domain images, generates high-quality multimodal retrieval data, outperforming larger existing datasets and achieving state-of-the-art zero-shot performance.
Yihao Meng, Hao Ouyang, Hanlin Wang +6 authors
AniDoc uses video diffusion models to automate colorization and in-betweening in 2D animation, improving efficiency by leveraging correspondence matching.
Jiarui Zhang, Ollie Liu, Tianyu Yu +2 authors
A benchmark called Geoperception is introduced to evaluate MLLMs' geometric description accuracy, leading to the development of Euclid, a model optimized for low-level geometric perception using synthetic data and a data curriculum.
Yuqing Wang, Shuhuai Ren, Zhijie Lin +6 authors
A parallel generation strategy for autoregressive models improves inference speed without significantly compromising quality in visual generation tasks.
Xuekai Zhu, Daixuan Cheng, Hengli Li +7 authors
The use of synthetic data in language model training leads to model collapse, which is mitigated by token-level editing of human-produced data to create semi-synthetic data.
LG AI Research, Soyoung An, Kyunghoon Bae +30 authors
EXAONE 3.5 language models, available in three configurations, demonstrate exceptional instruction following, long-context comprehension, and competitive performance across various benchmarks.
Frank F. Xu, Yufan Song, Boxuan Li +18 authors
A benchmark platform evaluates AI agents' autonomy in performing professional tasks in a simulated workplace environment, showing that a significant portion of simpler tasks can be automated but complex, long-term tasks remain challenging.
Yutao Sun, Hangbo Bao, Wenhui Wang +5 authors
LatentLM integrates continuous and discrete data using causal Transformers, VAEs, and next-token diffusion, achieving superior performance in multimodal tasks like image generation, large language model integration, and text-to-speech synthesis.
Zhisheng Zhong, Chengyao Wang, Yuqi Liu +12 authors
As Multi-modal Large Language Models (MLLMs) evolve, expanding beyond single-domain capabilities is essential to meet the demands for more versatile and efficient AI. However, previous omni-models have insufficiently explored speech, neglecting its integration with multi-modality. We introduce Lyra, an efficient MLLM that enhances multimodal abilities, including advanced long-speech comprehension, sound understanding, cross-modality efficiency, and seamless speech interaction. To achieve efficiency and speech-centric capabilities, Lyra employs three strategies: (1) leveraging existing open-source large models and a proposed multi-modality LoRA to reduce training costs and data requirements; (2) using a latent multi-modality regularizer and extractor to strengthen the relationship between speech and other modalities, thereby enhancing model performance; and (3) constructing a high-quality, extensive dataset that includes 1.5M multi-modal (language, vision, audio) data samples and 12K long speech samples, enabling Lyra to handle complex long speech inputs and achieve more robust omni-cognition. Compared to other omni-methods, Lyra achieves state-of-the-art performance on various vision-language, vision-speech, and speech-language benchmarks, while also using fewer computational resources and less training data.
Jianzong Wu, Chao Tang, Jingbo Wang +3 authors
DiffSensei, a diffusion-based framework with a multimodal large language model, enhances manga generation by precisely controlling multi-character appearances and interactions using masked cross-attention and text cues.
Jian Yang, Jiaxi Yang, Ke Jin +7 authors
A human-curated benchmark (CodeArena) and a large synthetic instruction corpus (SynCode-Instruct) are introduced to evaluate code LLMs based on human preference alignment, revealing performance differences between open-source and proprietary models.
Weihao Zeng, Yuzhen Huang, Lulu Zhao +3 authors
B-STaR framework autonomously adjusts configurations to enhance exploration and exploitation, improving performance in self-improving 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.
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.
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.
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.
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.
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.
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.
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