mlsanityData to 14 Oct 2026

PARD-2: Target-Aligned Parallel Draft Model for Dual-Mode Speculative Decoding

arXiv 2605.08632 ↗First posted 9 May 2026cs.CL cs.AIcited by 8

On the map: Efficient LLM inference and serving → Speculative decoding

Abstract (arXiv)

Speculative decoding accelerates Large Language Models (LLMs) inference by using a lightweight draft model to propose candidate tokens that are verified in parallel by the target model. However, existing draft model training objectives are not directly aligned with the inference-time goal of maximizing consecutive token acceptance. To address this issue, we reformulate the draft model optimization objective, shifting the focus from token prediction accuracy to the overall acceptance length. In this paper, we build upon PARD to propose PARD-2, a dual-mode speculative decoding framework with Confidence-Adaptive Token (CAT) optimization. This approach adaptively reweights each token to better align with the verification process. Notably, PARD-2 enables a single draft model to support both target-dependent and target-independent modes. Experiments across diverse models and tasks demonstrate that PARD-2 achieves up to 6.94× lossless acceleration, surpassing EAGLE-3 by 1.9× and PARD by 1.3× on Llama3.1-8B. Our code is available at https://github.com/AMD-AGI/PARD.

Key facts. Cited by 8 papers in the map (data through 14 Oct 2026). Placed in Speculative decoding, one of 12 directions in Efficient LLM inference and serving. Builds on 29 of its 34 references that are in the map. First posted on arXiv 9 May 2026.

Builds on

29 of 34 references.

Closest ancestors · 21

Same direction, newest first

  1. 2026Speculative Speculative Decoding3 Mar
  2. 2026DFlash: Block Diffusion for Flash Speculative Decoding5 Feb
  3. 2026P-EAGLE: Parallel-Drafting EAGLE with Scalable Training1 Feb
  4. 2026DART: Diffusion-Inspired Speculative Decoding for Fast LLM Inference27 Jan
  5. 2025SpecDiff-2: Scaling Diffusion Drafter Alignment For Faster Speculative Decoding1 Nov
  6. 2025DiffuSpec: Unlocking Diffusion Language Models for Speculative Decoding28 Sep
  7. 2025PARD: Accelerating LLM Inference with Low-Cost PARallel Draft Model Adaptation23 Apr
  8. 2025EAGLE-3: Scaling up Inference Acceleration of Large Language Models via Training-Time Test3 Mar
  9. 2024Falcon: Faster and Parallel Inference of Large Language Models through Enhanced Semi-Autoregressive Drafting and Custom-Designed Decoding Tree17 Dec
  10. 2024ParallelSpec: Parallel Drafter for Efficient Speculative Decoding8 Oct
  11. 2024PEARL: Parallel Speculative Decoding with Adaptive Draft Length13 Aug
  12. 2024Speculative Diffusion Decoding: Accelerating Language Generation through Diffusion10 Aug
9 more
  1. 2024EAGLE-2: Faster Inference of Language Models with Dynamic Draft Trees24 Jun
  2. 2024Hydra: Sequentially-Dependent Draft Heads for Medusa Decoding7 Feb
  3. 2024GliDe with a CaPE: A Low-Hassle Method to Accelerate Speculative Decoding3 Feb
  4. 2024Break the Sequential Dependency of LLM Inference Using Lookahead Decoding3 Feb
  5. 2024EAGLE: Speculative Sampling Requires Rethinking Feature Uncertainty26 Jan
  6. 2024Medusa: Simple LLM Inference Acceleration Framework with Multiple Decoding Heads19 Jan
  7. 2023DistillSpec: Improving Speculative Decoding via Knowledge Distillation12 Oct
  8. 2023Accelerating Large Language Model Decoding with Speculative Sampling2 Feb
  9. 2022Fast Inference from Transformers via Speculative Decoding30 Nov

Other references · 8

  1. 2021Training Verifiers to Solve Math Word Problemscited by 10,109
  2. 2025Qwen3 Technical Reportcited by 8,820
  3. 2023Judging LLM-as-a-Judge with MT-Bench and Chatbot Arenacited by 8,626
  4. 2021Evaluating Large Language Models Trained on Codecited by 8,320
  5. 2023Let's Verify Step by Stepcited by 3,789
  6. 2024Magpie: Alignment Data Synthesis from Scratch by Prompting Aligned LLMs with Nothingcited by 346
  7. 2025NVIDIA Nemotron Nano 2: An Accurate and Efficient Hybrid Mamba-Transformer Reasoning Modelcited by 108
  8. 2025Nemotron 3 Nano: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoningcited by 104

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 8.

8 papers.

  1. 2026Verification-Aware Training for Speculative Decodingcited by 5
  2. 2026Trees from Marginals: Autoregressive drafting with factorized priorscited by 3
  3. 2026When Is a Draft Accepted? A Theory of Acceptance in Speculative Decodingcited by 1
  4. 2026OoO-Spec: Out-of-Order Semantic Speculation for Fast Tool Callingcited by 1
  5. 2026Fully Rotation-Equivariant Spectral-Spatial Learning for Multispectral Object Detectioncited by 0
  6. 2026DEdit: Iterative Draft Editing for Speculative Decodingcited by 0
  7. 2026Recovering Off-Policy Supervision for Speculative Decodingcited by 0
  8. 2026JetSpec++: Foresight Parallel Drafting with Self Refinementscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025GRIFFIN: Effective Token Alignment for Faster Speculative Decoding
  2. 2026D^2SD: Accelerating Speculative Decoding with Dual Diffusion Draft Models
  3. 2024Decoding Speculative Decoding
  4. 2025Speculative Decoding: Performance or Illusion?
  5. 2025Speculative Verification: Exploiting Information Gain to Refine Speculative Decoding
  6. 2026TSS: Target-Side Sparsification for Speculative Decoding in Domain-Specific Large Language Models
  7. 2025AdaSD: Adaptive Speculative Decoding for Efficient Language Model Inference
  8. 2024Recursive Speculative Decoding: Accelerating LLM Inference via Sampling Without Replacement