paper.dinq.meAI 与前沿科学顶会 · 学术新星雷达

更新于

返回新星榜

PA

Pierre Ablin

ICML 2026×5长上下文与架构效率与系统数据与预训练

新星分贡献者

59

一作贡献0

核心/参与31

方向广度6

合作网络12

0 一作 · 5 篇 · ICML 2026

获取完整研究档案

首次获取该候选人的完整档案及 StartLux 评估共 20 积分；每位用户单独解锁，之后查看免费。主动刷新 20 积分。

联系方式

实时查找电子邮箱，费用不超过20积分；未找到邮箱不扣费。

2026 顶会论文

5

作者位次 #4/9ICML 2026Oral

Learning Unmasking Policies for Diffusion Language Models

Metod Jazbec · Theo X. Olausson · Louis Béthune · Pierre Ablin · Michael Kirchhof · Joao Monteiro · Victor Guilherme Turrisi da Costa · Jason Ramapuram · Marco Cuturi

长上下文与架构OpenReview

作者位次 #3/4ICML 2026Poster· Deep Learning / Algorithms

Balanced LoRA: Removing Parameter Invariance to Accelerate Convergence

Valérie Castin · Kimia Nadjahi · Pierre Ablin · Gabriel Peyré

效率与系统OpenReview

作者位次 #2/7ICML 2026Poster· Deep Learning / Large Language Models

Optimal Splitting of Language Models from Mixtures to Specialized Domains

Skyler Seto · Pierre Ablin · Anastasiia Filippova · Jiayuan Ye · Louis Béthune · Angelos Katharopoulos · David Grangier

作者位次 #3/5ICML 2026Poster· General Machine Learning / Causality

Multi-View Causal Discovery without Non-Gaussianity: Identifiability and Algorithms

Ambroise Heurtebise · Lemir Omar Chehab · Pierre Ablin · Alexandre Gramfort · Aapo Hyvarinen

作者位次 #6/6ICML 2026Poster

Removing Noise, not Finding Gold: Quality Filtering for Large-Scale Pretraining

Thiziri Nait Saada · Louis Béthune · Michal Klein · David Grangier · Marco Cuturi · Pierre Ablin

数据与预训练OpenReview

常见合作者

LBLouis Béthune×3 MCMarco Cuturi×2 DGDavid Grangier×2 MJMetod Jazbec TOTheo X. Olausson MKMichael Kirchhof JMJoao Monteiro VCVictor Guilherme Turrisi da Costa JRJason Ramapuram VCValérie Castin KNKimia Nadjahi GPGabriel Peyré