paper-with-me

홈 › Papers

A Mixture of $h-1$ Heads is Better than $h$ Heads

2020-05-13 · Hao Peng, Roy Schwartz, Dianqi Li, Noah A. Smith

Multi-head attentive neural architectures have achieved state-of-the-art results on a variety of natural language processing tasks. Evidence has shown that they are overparameterized; attention heads can be pruned without significant performance loss. In this work, we instead "reallocate" them -- the model learns to activate different heads on different inputs. Drawing connections between multi-head attention and mixture of experts, we propose the mixture of attentive experts model (MAE). MAE is trained using a block coordinate descent algorithm that alternates between updating (1) the responsibilities of the experts and (2) their parameters. Experiments on machine translation and language modeling show that MAE outperforms strong baselines on both tasks. Particularly, on the WMT14 English to German translation dataset, MAE improves over "transformer-base" by 0.8 BLEU, with a comparable number of parameters. Our analysis shows that our model learns to specialize different experts to different inputs.

📄 PDF Abstract BibTeX arXiv:2005.06537

Code (0)

등록된 구현이 없습니다.

Tasks

Language ModelingLanguage ModellingMachine TranslationMixture-of-ExpertsTranslation

Methods 이 논문이 사용한 방법론

Attention 설명 없음
Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…
Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.
Multi-Head Attention 설명 없음

Similar Papers 제목 키워드 기반

A Mixture of h - 1 Heads is Better than h Heads

2020-07-01 · ACL 2020 6 · Hao Peng, Roy Schwartz, Dianqi Li, Noah A. Smith

Multi-head attentive neural architectures have achieved state-of-the-art results on a variety of natural language processing tasks. Evidence has shown that they are overparameterized; attention heads can be pruned withou…

Language ModelingLanguage ModellingMachine TranslationMixture-of-Experts+1

Mixture of Attention Heads: Selecting Attention Heads Per Token

2022-10-11 · Xiaofeng Zhang, Yikang Shen, Zeyu Huang, Jie zhou 외

Mixture-of-Experts (MoE) networks have been proposed as an efficient way to scale up model capacity and implement conditional computing. However, the study of MoE components mostly focused on the feedforward layer in Tra…

Computational EfficiencyLanguage ModelingLanguage ModellingMachine Translation+2

Improving Transformers with Probabilistic Attention Keys

2021-10-16 · Tam Nguyen, Tan M. Nguyen, Dung D. Le, Duy Khuong Nguyen 외

Multi-head attention is a driving force behind state-of-the-art transformers, which achieve remarkable performance across a variety of natural language processing (NLP) and computer vision tasks. It has been observed tha…

Language ModelingLanguage Modelling

MISA: Mixture of Indexer Sparse Attention for Long-Context LLM Inference

2026-05-08 · Ruijie Zhou, Fanxu Meng, Yufei Xu, Tongxuan Liu 외 arxiv

DeepSeek Sparse Attention (DSA) sets the state of the art for fine-grained inference-time sparse attention by introducing a learned token-wise indexer that scores every prefix token and selects the most relevant ones for…

BACH: A Bayesian Admixture of Contrastive Heads for Multi-Interest Two-Tower Retrieval

2026-07-09 · Quoc Phong Nguyen, Paul Albert, Long Vuong, Vuong Le 외 arxiv

Two-tower retrievers compress each user into a single embedding, limiting their ability to serve diverse interests. Multi-interest models give each user several heads scored by a maximum inner product, but their hard-rou…