Imagen de Portada

Draft Less, Retrieve More: Hybrid Tree Construction for Speculative Decoding

Fuente: arXiv

Guardado en:

Detalles Bibliográficos
Autores principales:	Shen, Yuhao, Liu, Tianyu, Hu, Xinyi, Kong, Quan, Zhang, Baolin, Dai, Jun, Zhang, Jun, Ge, Shuang, Chen, Lei, Li, Yue, Wan, Mingcheng, Wang, Cong
Formato:	Preprint
Publicado:	2026
Materias:	Machine Learning Artificial Intelligence
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

ECHO: Elastic Speculative Decoding with Sparse Gating for High-Concurrency Scenarios
por: Hu, Xinyi, et al.
Publicado: (2026)

SpecBranch: Speculative Decoding via Hybrid Drafting and Rollback-Aware Branch Parallelism
por: Shen, Yuhao, et al.
Publicado: (2025)

When Hidden States Drift: Can KV Caches Rescue Long-Range Speculative Decoding?
por: Liu, Tianyu, et al.
Publicado: (2026)

Double: Breaking the Acceleration Limit via Double Retrieval Speculative Parallelism
por: Shen, Yuhao, et al.
Publicado: (2026)

ParallelVLM: Lossless Video-LLM Acceleration with Visual Alignment Aware Parallel Speculative Decoding
por: Kong, Quan, et al.
Publicado: (2026)

Cost-Aware Diffusion Draft Trees for Speculative Decoding
por: Zhang, Shuai, et al.
Publicado: (2026)

When Drafts Evolve: Speculative Decoding Meets Online Learning
por: Qian, Yu-Yang, et al.
Publicado: (2026)

PEARL: Parallel Speculative Decoding with Adaptive Draft Length
por: Liu, Tianyu, et al.
Publicado: (2024)

LongSpec: Long-Context Lossless Speculative Decoding with Efficient Drafting and Verification
por: Yang, Penghui, et al.
Publicado: (2025)

Draft & Verify: Lossless Large Language Model Acceleration via Self-Speculative Decoding
por: Zhang, Jun, et al.
Publicado: (2023)

Learning to Draft: Adaptive Speculative Decoding with Reinforcement Learning
por: Zhang, Jiebin, et al.
Publicado: (2026)

RASD: Retrieval-Augmented Speculative Decoding
por: Quan, Guofeng, et al.
Publicado: (2025)

TALON: Confidence-Aware Speculative Decoding with Adaptive Token Trees
por: Liu, Tianyu, et al.
Publicado: (2026)

Domino: Decoupling Causal Modeling from Autoregressive Drafting in Speculative Decoding
por: Huang, Jianuo, et al.
Publicado: (2026)

Draft-OPD: On-Policy Distillation for Speculative Draft Models
por: Lei, Haodi, et al.
Publicado: (2026)

OPT-Tree: Speculative Decoding with Adaptive Draft Tree Structure
por: Wang, Jikai, et al.
Publicado: (2024)

Ouroboros: Generating Longer Drafts Phrase by Phrase for Faster Speculative Decoding
por: Zhao, Weilin, et al.
Publicado: (2024)

FastEagle: Cascaded Drafting for Accelerating Speculative Decoding
por: Huang, Haiduo, et al.
Publicado: (2025)

Accelerating Speculative Decoding with Block Diffusion Draft Trees
por: Ringel, Liran, et al.
Publicado: (2026)

Flatter Tokens are More Valuable for Speculative Draft Model Training
por: Fan, Jiaming, et al.
Publicado: (2026)

DFlare: Scaling Up Draft Capacity for Block Diffusion Speculative Decoding
por: Zhang, Jiebin, et al.
Publicado: (2026)

Speculative RAG: Enhancing Retrieval Augmented Generation through Drafting
por: Wang, Zilong, et al.
Publicado: (2024)

When, What, and How: Rethinking Retrieval-Enhanced Speculative Decoding
por: Fang, Min, et al.
Publicado: (2025)

TETRIS: Optimal Draft Token Selection for Batch Speculative Decoding
por: Wu, Zhaoxuan, et al.
Publicado: (2025)

MineDraft: A Framework for Batch Parallel Speculative Decoding
por: Tang, Zhenwei, et al.
Publicado: (2026)

SpecHub: Provable Acceleration to Multi-Draft Speculative Decoding
por: Sun, Ryan, et al.
Publicado: (2024)

Draft, Verify, and Improve: Toward Training-Aware Speculative Decoding
por: Bhansali, Shrenik, et al.
Publicado: (2025)

POSS: Position Specialist Generates Better Draft for Speculative Decoding
por: Huang, Langlin, et al.
Publicado: (2025)

FlexDraft: Flexible Speculative Decoding via Attention Tuning and Bonus-Guided Calibration
por: Zhang, Yaojie, et al.
Publicado: (2026)

DREAM-S: Speculative Decoding with Searchable Drafting and Target-Aware Refinement for Multimodal Generation
por: Liu, Zining, et al.
Publicado: (2026)

KOALA: Enhancing Speculative Decoding for LLM via Multi-Layer Draft Heads with Adversarial Learning
por: Zhang, Kaiqi, et al.
Publicado: (2024)

LogitSpec: Accelerating Retrieval-based Speculative Decoding via Next Next Token Speculation
por: Liu, Tianyu, et al.
Publicado: (2025)

FlexSpec: Frozen Drafts Meet Evolving Targets in Edge-Cloud Collaborative LLM Speculative Decoding
por: Li, Yuchen, et al.
Publicado: (2026)

DuoDecoding: Hardware-aware Heterogeneous Speculative Decoding with Dynamic Multi-Sequence Drafting
por: Lv, Kai, et al.
Publicado: (2025)

AHASD: Asynchronous Heterogeneous Architecture for LLM Adaptive Drafting Speculative Decoding on Mobile Devices
por: Zirui, Ma, et al.
Publicado: (2026)

AdaEAGLE: Optimizing Speculative Decoding via Explicit Modeling of Adaptive Draft Structures
por: Zhang, Situo, et al.
Publicado: (2024)

OmniDraft: A Cross-vocabulary, Online Adaptive Drafter for On-device Speculative Decoding
por: Ramakrishnan, Ramchalam Kinattinkara, et al.
Publicado: (2025)

Bridging Draft Policy Misalignment: Group Tree Optimization for Speculative Decoding
por: Hu, Shijing, et al.
Publicado: (2025)

SpecBlock: Block-Iterative Speculative Decoding with Dynamic Tree Drafting
por: Shi, Weijie, et al.
Publicado: (2026)

Draft on the Fly: Adaptive Self-Speculative Decoding using Cosine Similarity
por: Metel, Michael R., et al.
Publicado: (2024)