Patterns最新文献_第4页

Efficient attention mechanisms for large language models. 大型语言模型的有效注意机制。

IF 10.8

Patterns Pub Date : 2026-07-03 eCollection Date: 2026-09-11 DOI: 10.1016/j.patter.2026.101594

Yutao Sun, Zhenyu Li, Yike Zhang, Tengyu Pan, Bowen Dong, Yuyi Guo, Jianyong Wang

{"title":"Efficient attention mechanisms for large language models.","authors":"Yutao Sun, Zhenyu Li, Yike Zhang, Tengyu Pan, Bowen Dong, Yuyi Guo, Jianyong Wang","doi":"10.1016/j.patter.2026.101594","DOIUrl":"https://doi.org/10.1016/j.patter.2026.101594","url":null,"abstract":"Transformer-based architectures have become the prevailing backbone of large language models. However, the time and memory complexity of self-attention remains a fundamental obstacle to efficient long-context modeling. To address this limitation, recent research has introduced two principal categories of efficient attention mechanisms. Linear attention methods achieve linear complexity through kernel approximations, recurrent formulations, or fast-weight dynamics, thereby enabling scalable inference with reduced computational overhead. Sparse attention techniques, in contrast, limit attention computation to selected subsets of tokens based on fixed or dynamic strategies, enhancing efficiency while preserving contextual coverage. This survey provides a comprehensive overview of these developments, integrating both algorithmic innovations and hardware-level considerations. In addition, we analyze the incorporation of efficient attention into large-scale pre-trained language models, including both architectures built entirely on efficient attention and hybrid designs that combine local and global components. By aligning theoretical foundations with practical deployment strategies, this work aims to serve as a foundational reference for advancing the design of scalable and efficient language models.","PeriodicalId":36242,"journal":{"name":"Patterns","volume":"7 9","pages":"101594"},"PeriodicalIF":10.8,"publicationDate":"2026-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13576653/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148978212","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Multimodal spatial omics: From data acquisition to computational integration. 多模态空间组学：从数据获取到计算集成。

IF 10.8

Patterns Pub Date : 2026-06-30 eCollection Date: 2026-07-10 DOI: 10.1016/j.patter.2026.101592

Esra Busra Isik, Yusuf Hakan Usta, Maryam Riazi, Haozhe Liu, William Roach, Hongpeng Zhou, Anna Nicolaou, Magnus Rattray, Sokratia Georgaka

引用次数: 0

Design principles for integrated AI alignment. 集成AI对齐的设计原则。

IF 10.8

Patterns Pub Date : 2026-06-29 eCollection Date: 2026-07-10 DOI: 10.1016/j.patter.2026.101587

Ben Y Reis, William G La Cava

引用次数: 0

Brain-AI convergence: Generative world models and hierarchical attention for human intelligence. 脑-人工智能融合：人类智能的生成世界模型和分层关注。

IF 10.8

Patterns Pub Date : 2026-06-26 eCollection Date: 2026-08-14 DOI: 10.1016/j.patter.2026.101593

Shogo Ohmae, Keiko Ohmae

{"title":"Brain-AI convergence: Generative world models and hierarchical attention for human intelligence.","authors":"Shogo Ohmae, Keiko Ohmae","doi":"10.1016/j.patter.2026.101593","DOIUrl":"10.1016/j.patter.2026.101593","url":null,"abstract":"Recent advances in general-purpose AI provide new insights into how the neocortex and cerebellum, despite their uniform circuit architectures, support diverse functions and human intelligence. Beyond the traditional focus on visual processing, this paper offers a cross-domain comparison of the brain and AI through the lens of world-model-based computation. We argue that both the neocortex and cerebellum predict future world states from past inputs and construct predictive world models through prediction-error learning. These predictive world models are repurposed for sensory comprehension and motor output generation, thereby supporting multi-domain capabilities. Underlying these capabilities and even human-like adaptive intelligence, the neocortex implements hierarchical attention-based processing. Interestingly, autoregressive generative models in transformer-based AI have independently converged on similar computational principles. Together, these shared mechanisms suggest a common computational foundation through which uniform circuits can support cross-domain high-level intelligence in both biological and artificial systems.","PeriodicalId":36242,"journal":{"name":"Patterns","volume":"7 8","pages":"101593"},"PeriodicalIF":10.8,"publicationDate":"2026-06-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13494634/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148799026","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

OpenChallenges: A FAIR-aligned, centralized platform for crowdsourced challenges in biomedical research. 开放挑战：一个公平的、集中的平台，用于生物医学研究中的众包挑战。

IF 10.8

Patterns Pub Date : 2026-06-24 eCollection Date: 2026-09-11 DOI: 10.1016/j.patter.2026.101595

Verena Chung, Rongrong Chai, Hallie Swan, Maria Diaz, Shysta Bynum, Adam Hindman, Jacob Albrecht, Serghei Mangul, Susheel Varma, Jay Hodgson, Solveig K Sieberts, Alberto Pepe, Luca Foschini, Thomas Schaffter, Gaia Andreoletti

{"title":"OpenChallenges: A FAIR-aligned, centralized platform for crowdsourced challenges in biomedical research.","authors":"Verena Chung, Rongrong Chai, Hallie Swan, Maria Diaz, Shysta Bynum, Adam Hindman, Jacob Albrecht, Serghei Mangul, Susheel Varma, Jay Hodgson, Solveig K Sieberts, Alberto Pepe, Luca Foschini, Thomas Schaffter, Gaia Andreoletti","doi":"10.1016/j.patter.2026.101595","DOIUrl":"10.1016/j.patter.2026.101595","url":null,"abstract":"Open community challenges are critical for benchmarking computational methods in biomedical research, but the landscape is highly fragmented. Currently, over 2,000 challenges exist across a large and fragmented set of platforms, yet no centralized registry exists to index them. To address this, we developed OpenChallenges, a centralized, FAIR (Findable, Accessible, Interoperable, and Reusable)-aligned hub that aggregates and standardizes challenge metadata from diverse sources into a single, searchable registry. By implementing an updated metadata model explicitly mapped to FAIR dimensions, the platform advances challenge discoverability. A metadata completeness scoring framework assigns each challenge one of four tiers (baseline, standard, comprehensive, and gold standard), providing users with a transparent measure of metadata richness and guiding community curation priorities. Built on an open-source, API-driven architecture, OpenChallenges streamlines discovery, enhances visibility, and supports community-driven curation. This resource transforms a fragmented ecosystem into an interconnected knowledge base, advancing transparency in computational biology.","PeriodicalId":36242,"journal":{"name":"Patterns","volume":"7 9","pages":"101595"},"PeriodicalIF":10.8,"publicationDate":"2026-06-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13576655/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148978284","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Disentangling covariate effects on single-cell-resolved epigenomes with DeepDive. 用DeepDive解开单细胞表观基因组的协变量效应。

IF 10.8

Patterns Pub Date : 2026-06-17 eCollection Date: 2026-09-11 DOI: 10.1016/j.patter.2026.101589

Andreas Fønss Møller, Jesper Grud Skat Madsen

引用次数: 0

Density-based longitudinal neuron tracking in high-density electrophysiological recordings. 高密度电生理记录中基于密度的纵向神经元跟踪。

IF 10.8

Patterns Pub Date : 2026-06-17 eCollection Date: 2026-08-14 DOI: 10.1016/j.patter.2026.101590

Yue Huang, Hanbo Wang, Jiaming Cao, Yu Chen, Xuanning Wang, Yujie Zhao, Hengkun Ren, Qiang Zheng, Jianing Yu

{"title":"Density-based longitudinal neuron tracking in high-density electrophysiological recordings.","authors":"Yue Huang, Hanbo Wang, Jiaming Cao, Yu Chen, Xuanning Wang, Yujie Zhao, Hengkun Ren, Qiang Zheng, Jianing Yu","doi":"10.1016/j.patter.2026.101590","DOIUrl":"10.1016/j.patter.2026.101590","url":null,"abstract":"Tracking neurons across days in high-density extracellular recordings is essential for investigating the mechanisms of learning and representational drift. However, in weeks-long recordings, identifying matches across sessions is hindered by changes in spike waveforms and unit turnover. We introduce DANT (density-based across-day neuron tracking), a framework that iterates between density-based clustering in feature space and probe-motion estimation inferred from provisional matches. The estimated motion is then used to reregister spike waveforms across sessions before clustering is recomputed in the next iteration. Within this loop, DANT learns a decision boundary from match and non-match labels and uses it in post hoc curation. Applied to weeks-long Neuropixels recordings from cortex and striatum in freely moving rats during reaction-time and self-timing tasks, DANT substantially increases match yield while maintaining a low false-positive rate relative to existing approaches. These results establish DANT as a general unsupervised solution for longitudinal tracking in chronic recordings.","PeriodicalId":36242,"journal":{"name":"Patterns","volume":"7 8","pages":"101590"},"PeriodicalIF":10.8,"publicationDate":"2026-06-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13494646/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148799128","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

CLEAR-HPV: Interpretable concept discovery for human-papillomavirus-associated morphology in whole-slide histology. CLEAR-HPV：人类乳头瘤病毒相关形态学在全切片组织学上的可解释概念发现。

IF 10.8

Patterns Pub Date : 2026-06-16 eCollection Date: 2026-09-11 DOI: 10.1016/j.patter.2026.101588

Weiyi Qin, Yingci Liu-Swetz, Shiwei Tan, Hao Wang

{"title":"CLEAR-HPV: Interpretable concept discovery for human-papillomavirus-associated morphology in whole-slide histology.","authors":"Weiyi Qin, Yingci Liu-Swetz, Shiwei Tan, Hao Wang","doi":"10.1016/j.patter.2026.101588","DOIUrl":"10.1016/j.patter.2026.101588","url":null,"abstract":"Human papillomavirus (HPV) status is a critical determinant of prognosis and treatment response in head and neck and cervical cancers. Although attention-based multiple instance learning (MIL) achieves strong slide-level prediction for HPV-related whole-slide histopathology, it provides limited morphologic interpretability. To address this limitation, we introduce concept-level explainable attention-guided representation for HPV (CLEAR-HPV), a framework that restructures the MIL latent space to enable concept discovery without requiring concept labels during training. Within an attention-weighted latent space, CLEAR-HPV automatically discovers keratinizing, basaloid, and stromal morphologic concepts; generates spatial concept maps; and represents each slide with a compact concept-fraction vector. Its concept-fraction vectors preserve the predictive information of the original MIL embeddings while reducing the high-dimensional feature space (e.g., 1,536 dimensions) to only 10 interpretable concepts. CLEAR-HPV demonstrates consistent concept structure across The Cancer Genome Atlas (TCGA)-HNSCC, TCGA-CESC, and Clinical Proteomic Tumor Analysis Consortium (CPTAC) -HNSCC, providing compact, concept-level interpretability through a general, backbone-agnostic framework for attention-based MIL models of whole-slide histopathology.","PeriodicalId":36242,"journal":{"name":"Patterns","volume":"7 9","pages":"101588"},"PeriodicalIF":10.8,"publicationDate":"2026-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13576518/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148978087","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Moving from disclosure to substance in AI transparency. 从披露到人工智能透明度的实质。

IF 10.8

Patterns Pub Date : 2026-06-12 DOI: 10.1016/j.patter.2026.101585

Huixin Zhan, Jason H Moore

引用次数: 0

Resisting epistemic loss in AI image generation. 抗人工智能图像生成中的认知损失。

IF 10.8

Patterns Pub Date : 2026-06-12 DOI: 10.1016/j.patter.2026.101568

Abdullah Hasan Safir, Alan F Blackwell, Ramit Debnath

引用次数: 0