牛哥精选 · 三个月

1

📝 深度技术 arXiv AI 2026-06-09

InA-Probe: Instruction-Aware Active Probing for Time Series Forecasting with LLMs

为时间序列预测量身定制大模型，指令感知主动探测方法提升预测精度与效率。

arXiv:2606.08601v1 Announce Type: new Abstract: Large Language Models (LLMs) have recently demonstrated impressive potential for time series forecasti…

时间序列预测大语言模型主动探测指令感知 ina-probe

2

📝 深度技术 Hacker News LLM 2026-06-09

LLM Are Universal Simulators

揭示LLM不仅是下一个token预测，而是具备前瞻规划的通用模拟器，颠覆认知。

Article URL: https://invertedpassion.com/llm-are-universal-simulators/ Comments URL: https://news.ycombinator.com/item?id=48449409 Points: 1 # Comment…

llm 大语言模型通用模拟器序列预测注意力机制

3

📝 深度技术 arXiv 机器学习 2026-06-09

Causal Semantic Alignment for LLM-based Time Series Forecasting

提出因果语义对齐方法，让大语言模型更精准捕捉时间序列中的因果关系，提升预测性能。

arXiv:2606.08262v1 Announce Type: new Abstract: Recent advances in Large Language Models (LLMs) have opened new possibilities for time series forecast…

时间序列预测因果语义对齐大语言模型 llm 深度学习

4

📝 深度技术 arXiv AI 2026-06-02

Bridging the Last Mile of Time Series Forecasting with LLM Agents

用LLM智能体打通时间序列预测的“最后一公里”，前沿方法探索。

arXiv:2606.02497v1 Announce Type: new Abstract: Time series forecasting has advanced rapidly, especially with the emergence of foundation models that …

时间序列预测 llm智能体人工智能研究论文深度学习

5

🤖 AI·大模型 arXiv 机器学习 2026-06-02

Byte Pair Encoding for Efficient Time Series Forecasting

将自然语言处理中的字节对编码创新应用于时间序列预测，开辟效率提升新路径。

arXiv:2505.14411v4 Announce Type: replace Abstract: Existing time series tokenization methods predominantly encode a constant number of samples into i…

byte pair 时间序列预测效率提升创新应用

6

📝 深度技术 arXiv 机器学习 2026-05-29

Rethinking Post-Training Recipes for Multimodal Time-Series Forecasting

突破性后训练优化方案，让多模态时间序列预测表现更精准、更鲁棒。

arXiv:2605.29401v1 Announce Type: new Abstract: Time-Series Foundation Models (TSFMs) excel at zero-shot unimodal forecasting using numerical data, bu…

多模态时间序列预测后训练优化方法 ai前沿

7

📝 深度技术 arXiv AI 2026-05-29

KairosAgent: Agentic Time Series Forecasting with Fused Semantic Reasoning

创新融合语义推理与智能体技术，突破传统时间序列预测局限的arXiv最新论文。

arXiv:2605.30002v1 Announce Type: new Abstract: Cross-domain multimodal time series forecasting is a challenging task, requiring models to integrate p…

时间序列预测语义推理智能体 ai论文深度学习

8

🤖 AI·大模型 Hacker News LLM 2026-05-27

LLM Driven AutoForecasting with Sktime's `Craft()`

手把手教你用Sktime的Craft()函数，让大模型自动搞定时间序列预测

Article URL: https://pub.towardsai.net/llm-driven-autoforecasting-with-sktimes-craft-0355f5c720e8 Comments URL: https://news.ycombinator.com/item?id=4…

llm 自动预测时间序列 sktime python

9

📄 文档手册 arXiv 机器学习 2026-05-27

Probabilistic Recurrent Intention Switching Model

提出概率递归意图切换模型，用概率方法建模意图动态转换，为序列决策和时序预测提供新思路

arXiv:2605.26998v1 Announce Type: new Abstract: Inverse reinforcement learning (IRL) recovers reward functions from observed behavior, yet traditional…

概率模型递归神经网络意图切换序列预测决策建模

10

🤖 AI·大模型 arXiv 机器学习 2026-05-26

FuXi-Nowcast: Environment-conditioned deep learning for severe convection nowcasting

AI模型结合环境条件提升恶劣对流临近预报精度，气象预测新突破

arXiv:2512.08974v2 Announce Type: replace-cross Abstract: Severe convection produces localized hazards that often require warnings before radar echoes…

深度学习恶劣对流天气临近预报环境条件气象预测

11

📝 深度技术 arXiv AI 2026-05-23

ChronoMedicalWorld: A Medical World Model for Learning Patient Trajectories from Longitudinal Care Data

首个时间感知医疗世界模型，从纵向护理数据学习患者轨迹，解决慢性病长期模拟中的模型漂移难题。

arXiv:2605.21963v1 Announce Type: cross Abstract: Long-horizon clinical simulation -- predicting how a patient's physiology evolves over years under s…

医疗世界模型患者轨迹纵向护理数据慢性病模拟 ehr

12

🤖 AI·大模型 arXiv 机器学习 2026-05-21

Time-Prompt: Integrated Heterogeneous Prompts for Unlocking LLMs in Time Series Forecasting

提出Time-Prompt方法，通过集成异构提示让大语言模型高效处理时间序列预测，效果显著。

arXiv:2506.17631v4 Announce Type: replace Abstract: Time series forecasting aims to model temporal dependencies among variables for future state infer…

时间序列预测大语言模型提示工程深度学习异质提示集成

13

📝 深度技术 arXiv 机器学习 2026-05-20

LEAF: A Living Benchmark for Event-Augmented Forecasting

针对预测任务，LEAF动态基准填补了多维事件评估空白，让大模型预测能力测试更贴近现实。

arXiv:2605.16358v1 Announce Type: new Abstract: Large Language Models (LLMs) are increasingly applied to forecasting. To evaluate this capability whil…

机器学习事件增强预测活基准大模型评估时间序列预测

14

📝 深度技术 arXiv 机器学习 2026-05-20

Revisiting Long-term Time Series Forecasting: An Investigation on Linear Mapping

重新审视线性映射在长期时间序列预测中的有效性，挑战了复杂模型的必要地位。

arXiv:2305.10721v2 Announce Type: replace Abstract: Introduction: Long-term time series forecasting (LTSF) has gained significant attention in recent …

时间序列预测线性映射深度学习长期预测 arxiv论文

15

📝 深度技术 arXiv 机器学习 2026-05-20

L-Drive: Beyond a Single Mapping-Latent Context Drives Time Series Forecasting

提出L-Drive方法，突破单映射限制，用潜在上下文驱动时间序列预测，为时序建模提供新思路。

arXiv:2605.17730v1 Announce Type: new Abstract: Mainstream methods for multivariate time-series forecasting largely follow the Direct-Mapping paradigm…

时间序列预测潜在上下文 l-drive 深度学习预测方法

16

📝 深度技术 arXiv AI 2026-05-19

Time Series Forecasting Through the Lens of Dynamics

从动力学视角提出PRO-DYN，挑战Transformer同质化，直接学习过去到未来的映射。

arXiv:2507.15774v3 Announce Type: replace-cross Abstract: While deep learning is facing an homogenization across modalities led by Transformers, they …

时间序列预测动力学深度学习 pro-dyn 机器学习

🐂 牛哥精选

InA-Probe: Instruction-Aware Active Probing for Time Series Forecasting with LLMs

LLM Are Universal Simulators

Causal Semantic Alignment for LLM-based Time Series Forecasting

Bridging the Last Mile of Time Series Forecasting with LLM Agents

Byte Pair Encoding for Efficient Time Series Forecasting

Rethinking Post-Training Recipes for Multimodal Time-Series Forecasting

KairosAgent: Agentic Time Series Forecasting with Fused Semantic Reasoning

LLM Driven AutoForecasting with Sktime's `Craft()`

Probabilistic Recurrent Intention Switching Model

FuXi-Nowcast: Environment-conditioned deep learning for severe convection nowcasting

ChronoMedicalWorld: A Medical World Model for Learning Patient Trajectories from Longitudinal Care Data

Time-Prompt: Integrated Heterogeneous Prompts for Unlocking LLMs in Time Series Forecasting

LEAF: A Living Benchmark for Event-Augmented Forecasting

Revisiting Long-term Time Series Forecasting: An Investigation on Linear Mapping

L-Drive: Beyond a Single Mapping-Latent Context Drives Time Series Forecasting

Time Series Forecasting Through the Lens of Dynamics

📅 日期