paper-with-me

홈 › Papers

Efficient Many-Shot In-Context Learning with Dynamic Block-Sparse Attention

2025-03-11 · Emily Xiao, Chin-Jou Li, Yilin Zhang, Graham Neubig, Amanda Bertsch

Many-shot in-context learning has recently shown promise as an alternative to finetuning, with the major advantage that the same model can be served for multiple tasks. However, this shifts the computational burden from training-time to inference-time, making deployment of many-shot ICL challenging to justify in-practice. This cost is further increased if a custom demonstration set is retrieved for each inference example. We present Dynamic Block-Sparse Attention, a training-free framework for retrieval-based many-shot in-context learning. By combining carefully designed block-sparse attention and retrieval of cached groups of demonstrations, we achieve comparable per-example latency to finetuning while maintaining on average >95% of the best method's accuracy across strong ICL and finetuning baselines. We hope that this will further enable the deployment of many-shot ICL at scale.

📄 PDF Abstract BibTeX arXiv:2503.08640

Code (1)

millix19/dbsa 공식 구현 pytorch

Tasks

In-Context LearningRetrieval

Methods 이 논문이 사용한 방법론

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 $x$…
Attention 설명 없음
SET Dynamic Sparse Training method where weight mask is updated randomly periodically

Similar Papers 제목 키워드 기반

Dynamic sparsity in tree-structured feed-forward layers at scale

2026-03-18 · Reza Sedghi, Robin Schiewer, Anand Subramoney, David Kappel arxiv

At typical context lengths, the feed-forward MLP block accounts for a large share of a transformer's compute budget, motivating sparse alternatives to dense MLP blocks. We study sparse, tree-structured feed-forward layer…

Question Answering

CompactAttention: Accelerating Chunked Prefill with Block-Union KV Selection

2026-05-16 · Jiwon Song, Dongwon Jo, Beomseok Kang, Jae-Joon Kim arxiv

Chunked prefill has become a widely adopted serving strategy for long-context large language models, but efficient attention computation in this regime remains challenging. Existing sparse attention methods are primarily…

Simultaneous Block-Sparse Signal Recovery Using Pattern-Coupled Sparse Bayesian Learning

2017-11-06 · Hang Xiao, Zhengli Xing, Linxiao Yang, Jun Fang 외

In this paper, we consider the block-sparse signals recovery problem in the context of multiple measurement vectors (MMV) with common row sparsity patterns. We develop a new method for recovery of common row sparsity MMV…

Predict, Reuse, and Repair: Accelerating Dynamic Sparse Attention for Long-Context LLM Decoding

2026-06-29 · Tianyu Wang, Gourav Rattihalli, Aditya Dhakal, Junbo Li 외 arxiv

Dynamic sparse attention (DSA) accelerates long-context LLM decoding by attending to only the top-K KV blocks relevant to each query, but it introduces a serialized selection-to-attention dependency that emerges as a new…

Parametric Models for DOA Trajectory Localization

2022-04-20 · Ruchi Pandey, Santosh Nannuru

Directions of arrival (DOA) estimation or localization of sources is an important problem in many applications for which numerous algorithms have been proposed. Most localization methods use block-level processing that c…