cs.LG（2026-09-09）

📊 共 18 篇论文 | 🔗 2 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (8 🔗2) 支柱二：RL算法与架构 (RL & Architecture) (6) 支柱一：机器人控制 (Robot Control) (2) 支柱三：空间感知与语义 (Perception & Semantics) (1) 支柱四：生成式动作 (Generative Motion) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (8 篇)

#题目一句话要点标签🔗⭐
1 OmniMed-FL: A Robust Multimodal Federated Learning Framework for Clinical Diagnosis 提出OmniMed-FL以解决临床诊断中的多模态数据融合问题 multimodal
2 EEGBind: Detecting Source-Level Interictal Epileptiform Discharges via EEG-Centric Multimodal Binding 提出EEGBind以解决源级间歇性癫痫放电检测问题 multimodal ✅
3 Forgetting Only What Matters: Layer-Selective Unlearning toward Robust LLMs 提出层选择性遗忘方法以增强大语言模型的鲁棒性 large language model
4 A Later Test Set Is Not a New Domain: Pretraining Familiarity Survives a Contamination-Free Hold-Out 提出时间序列模型评估新方法以解决预训练数据影响问题 foundation model
5 Maverick: Private and Verifiable LLM Inference Made Practical via Matrix-Vector Multiplication Delegation 提出Maverick以解决大规模LLM推理中的隐私与可验证性问题 large language model
6 Forward-Free LLM Depth Pruning via Weight Redundancy 提出权重冗余剪枝方法以解决LLM推理成本问题 large language model
7 Privacy-Preserving Split Learning for Federated LLM Fine-Tuning 提出隐私保护的分割学习方法以解决联邦大语言模型微调问题 large language model
8 PELM: Power Efficient On-Device LLM Inference with Speculative Decoding and Dynamic Voltage Frequency Scaling 提出PELM以解决移动设备上LLM推理的能效问题 large language model ✅

🔬 支柱二：RL算法与架构 (RL & Architecture) (6 篇)

#题目一句话要点标签🔗⭐
9 Semigroup-JEPA: Latent Dynamics Consistency for Zero-Shot Physics Generalization 提出SemiGroup-JEPA以解决物理建模的零-shot泛化问题 diffusion policy world model worldmodel
10 Distillation of Synthetic Data for Time Series Foundation Models 提出合成数据蒸馏方法以提升时间序列基础模型训练效率 distillation foundation model
11 Searching for New Physics with Reinforcement Learning 提出强化学习方法以寻找新物理现象的SMEFT运算符 reinforcement learning
12 CompassOPD: Cross-Family On-Policy Distillation via Within-Family Likelihood Shifts 提出CompassOPD以解决跨家族蒸馏效果下降问题 distillation
13 Optimal Value Inference for Reinforcement Learning 提出一种新方法以优化强化学习中的价值推断 reinforcement learning
14 Teacher Geometry Shapes Learnability in Teacher-Student Networks 提出教师几何形状可学习性以优化教师-学生网络 teacher-student

🔬 支柱一：机器人控制 (Robot Control) (2 篇)

#题目一句话要点标签🔗⭐
15 Adversarial Training for Tabular Credit Scoring: A Multi-Attack Robustness Evaluation in P2P Lending 提出对抗训练以提升P2P信贷评分的鲁棒性 manipulation
16 Cascading Gradient Inversion via LT-Code Inspired Peeling in Federated Learning 提出基于LT码的渐进梯度反演方法以解决联邦学习隐私泄露问题 manipulation

🔬 支柱三：空间感知与语义 (Perception & Semantics) (1 篇)

#题目一句话要点标签🔗⭐
17 View-Structured Conformal Prediction for 3D Gaussian Splatting 提出视结构一致性预测以解决3D高斯渲染的不确定性问题 3D gaussian splatting 3DGS gaussian splatting

🔬 支柱四：生成式动作 (Generative Motion) (1 篇)

#题目一句话要点标签🔗⭐
18 Field-level prediction of mid-plane stress tensor fields in concrete target penetration: a cross-velocity graph neural operator surrogate 提出图神经算子代理以预测混凝土靶材的应力张量场 penetration

⬅️ 返回 cs.LG 首页 · 🏠 返回主页