cs.LG（2024-08-29）

📊 共 13 篇论文 | 🔗 2 篇有代码

🎯 兴趣领域导航

支柱二：RL算法与架构 (RL & Architecture) (7 🔗1) 支柱九：具身大模型 (Embodied Foundation Models) (4 🔗1) 支柱六：视频提取与匹配 (Video Extraction) (1) 支柱四：生成式动作 (Generative Motion) (1)

🔬 支柱二：RL算法与架构 (RL & Architecture) (7 篇)

#题目一句话要点标签🔗⭐
1 Maven: A Multimodal Foundation Model for Supernova Science Maven：用于超新星科学的多模态基础模型，提升分类与红移估计性能。 contrastive learning foundation model multimodal
2 Seeking the Sufficiency and Necessity Causal Features in Multimodal Representation Learning 提出基于必要性和充分性因果特征的多模态表征学习方法，提升预测性能和鲁棒性。 representation learning multimodal
3 Evaluating Time-Series Training Dataset through Lens of Spectrum in Deep State Space Models 提出K-谱度量，用于评估深度状态空间模型训练数据集的有效性。 SSM state space model
4 GSTAM: Efficient Graph Distillation with Structural Attention-Matching 提出GSTAM，通过结构注意力匹配实现高效的图分类数据集蒸馏 distillation ✅
5 Subspace Representation Learning for Sparse Linear Arrays to Localize More Sources than Sensors: A Deep Learning Methodology 提出基于深度学习的子空间表示学习方法，用于稀疏线阵中多于传感器数量的声源定位 representation learning
6 Iterated Energy-based Flow Matching for Sampling from Boltzmann Densities 提出迭代能量基流匹配(iEFM)，用于从玻尔兹曼密度中采样。 flow matching
7 Towards Efficient Modelling of String Dynamics: A Comparison of State Space and Koopman based Deep Learning Methods 提出基于Koopman算子的深度学习模型，高效建模线性和非线性弦动力学 SSM state space model

🔬 支柱九：具身大模型 (Embodied Foundation Models) (4 篇)

#题目一句话要点标签🔗⭐
8 Multimodal ELBO with Diffusion Decoders 提出基于扩散解码器的多模态ELBO，提升多模态VAE的生成质量和一致性 multimodal
9 Preserving Diversity in Supervised Fine-Tuning of Large Language Models 提出GEM算法，通过博弈论方法提升大语言模型有监督微调中的多样性 large language model
10 CrisperWhisper: Accurate Timestamps on Verbatim Speech Transcriptions CrisperWhisper：通过优化Whisper模型提升语音转录时间戳精度 TAMP ✅
11 Illuminating the Diversity-Fitness Trade-Off in Black-Box Optimization 研究黑盒优化中多样性-质量权衡，揭示现有算法的潜在局限性 multimodal

🔬 支柱六：视频提取与匹配 (Video Extraction) (1 篇)

#题目一句话要点标签🔗⭐
12 Downscaling Neural Network for Coastal Simulations 提出DNNCS模型，用于海岸模拟的时空超分辨率重建，提升洪水预测精度。 feature matching spatiotemporal

🔬 支柱四：生成式动作 (Generative Motion) (1 篇)

#题目一句话要点标签🔗⭐
13 Blending Low and High-Level Semantics of Time Series for Better Masked Time Series Generation 提出NC-VQVAE框架，融合时序数据低级和高级语义以提升生成质量 VQ-VAE

⬅️ 返回 cs.LG 首页 · 🏠 返回主页