[KongchangAI]
Benchmark神经信息处理系统大会2026 / NeurIPS2026

NeurIPS 2026

NeurIPS (Conference on Neural Information Processing Systems) is one of the most influential top-tier academic conferences in artificial intelligence and machine learning. Held annually, it features highly competitive paper acceptance and serves as a major academic forum for fields including deep learning and reinforcement learning.

Timeline (last 90 days)

Oct 14

RL-related papers accounted for a fairly high proportion of NeurIPS 2026 submissions, especially work combined with language models.

Unverified50%
Oct 14

Some researchers systematically reviewed the 9094 papers of NeurIPS 2026, and the analysis results show that agents, reinforcement learning fine-tuning and small models are the three directions repeatedly mentioned.

Unverified50%
Oct 14

Among the NeurIPS 2026 papers, work on model compression, knowledge distillation and quantized inference has clearly increased, and many come with reproducible code and pre-trained weights.

Unverified50%
Oct 14

Among all the analyzed papers, the growth in agent-related research was the most significant.

Unverified50%
Oct 12

The CL4SM (Constrained Learning for Scientific Modeling) workshop is a pre-conference event at NeurIPS 2026 in Paris, held as a track within the ELLIS UnConference.

Unverified50%
Oct 12

The CL4SM workshop venue is the Palais des Congrès de Paris, just before the official opening of NeurIPS 2026.

Unverified50%
Oct 5

M-plicits was accepted at NeurIPS 2026.

Unverified50%
Oct 1

The paper "Parallel-in-Time Training of Recurrent Neural Networks for Dynamical Systems Reconstruction" was selected as a NeurIPS 2026 Spotlight.

Unverified50%
Sep 27

NeurIPS officially expects designated authors to be physically present at the assigned location to present their work.

Unverified50%
Sep 27

NeurIPS 2026 adopts a multi-location conference format, and each paper will be assigned to a specific location.

Unverified50%

11 more timeline events

All Facts (18)

Unverified

Among all the analyzed papers, the growth in agent-related research was the most significant.

50%
Unverified

Among the NeurIPS 2026 papers, work on model compression, knowledge distillation and quantized inference has clearly increased, and many come with reproducible code and pre-trained weights.

50%
Unverified

RL-related papers accounted for a fairly high proportion of NeurIPS 2026 submissions, especially work combined with language models.

50%
Unverified

The NeurIPS official blog states that if the designated author has already registered for a location different from the one assigned to their paper, they must cancel the registration and re-register at the designated location.

50%
Unverified

NeurIPS officially expects designated authors to be physically present at the assigned location to present their work.

50%
Unverified

The NeurIPS registration preference form includes the option 'No presenter chosen yet'.

50%
Unverified

For key questions such as 'whether guaranteed quotas will expire or be randomly assigned,' the most reliable approach is to contact the conference registration support email or official FAQ directly, rather than relying on community speculation.

50%
Unverified

NeurIPS 2026 adopts a multi-location conference format, and each paper will be assigned to a specific location.

50%
Unverified

The guaranteed registration quota for NeurIPS 2026 is tied to the specific location to which the paper is assigned.

50%
Unverified

NeurIPS 2026 公布的 73 个工作坊中没有一个专门聚焦于因果推理主题

50%
Unverified

本文分析基于Reddit机器学习社区的单一讨论帖及NeurIPS 2026工作坊公开列表

50%
Unverified

Some researchers systematically reviewed the 9094 papers of NeurIPS 2026, and the analysis results show that agents, reinforcement learning fine-tuning and small models are the three directions repeatedly mentioned.

50%
Unverified

The CL4SM (Constrained Learning for Scientific Modeling) workshop is a pre-conference event at NeurIPS 2026 in Paris, held as a track within the ELLIS UnConference.

50%
Unverified

The CL4SM workshop venue is the Palais des Congrès de Paris, just before the official opening of NeurIPS 2026.

50%
Unverified

M-plicits was accepted at NeurIPS 2026.

50%
Unverified

The paper "Parallel-in-Time Training of Recurrent Neural Networks for Dynamical Systems Reconstruction" was selected as a NeurIPS 2026 Spotlight.

50%
Unverified

NeurIPS 2026 will introduce an Automatic Reference Checker.

50%
Unverified

The RealPDE competition is a competition under the NeurIPS 2026 framework, focusing on the solving of partial differential equations (PDEs) in scientific machine learning and alignment with real-world data.

50%

Source Articles