Diffusion-GR2: Diffusion Generative Reasoning Re-ranker

• URL: http://arxiv.org/abs/2607.01170v3
• Date: Thu, 09 Jul 2026 08:33:13 GMT
• Title: Diffusion-GR2: Diffusion Generative Reasoning Re-ranker
• Abstract summary: Generative reasoning re-rankers achieve strong recommendation accuracy by emitting a chain-of-thought before re-ordering a candidate list.<n>An autoregressive (AR) decoder spends one sequential forward pass per reasoning token, and the reasoning trace far exceeds the ranking it produces.<n>To reduce this cost, block-diffusion language models decode many positions in parallel over a few denoising steps.
• Score: 36.3046891412256
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Abstract: Generative reasoning re-rankers achieve strong recommendation accuracy by emitting a chain-of-thought before re-ordering a candidate list, but they are slow at inference: an autoregressive (AR) decoder spends one sequential forward pass per reasoning token, and the reasoning trace far exceeds the ranking it produces. To reduce this cost, block-diffusion language models decode many positions in parallel over a few denoising steps and are substantially faster, yet naively converting an AR re-ranker into one opens two accuracy gaps: (1) a structural gap: answer positions are denoised in parallel and scored independently, so the decoder emits invalid rankings (duplicated, dropped, or out-of-set identifiers) that AR avoids through left-to-right masking; and (2) a distributional gap: fine-tuning the converted model on fixed teacher trajectories is off-policy relative to its own decoding at inference, leaving a residual accuracy gap. To close both gaps while keeping the speedup, we propose \textbf{Diffusion-GR2}, a recipe that converts our AR reasoning re-ranker (GR2) into a block-diffusion re-ranker. First, conversion fine-tuning (CFT) adapts the AR-initialized diffusion model to denoise the answer into a valid permutation on its own, without an external constrained decoder. Next, on-policy distillation (OPD) then supervises the model on its own decoded trajectories with dense per-token targets from the AR teacher. Finally, we apply a reinforcement-learning (RL) stage against a re-ranking reward on top of OPD's on-policy policy. Experiments on Amazon Beauty demonstrate that Diffusion-GR2 recovers to near-parity with the AR re-ranker, while block-parallel decoding raises decode throughput by $2.4$--$3.5\times$ at the model's reasoning output length. Ablations show that CFT recovers most of the conversion gap, and that on-policy distillation further closes it to the AR reference.

Related papers

• GEAR: Guided End-to-End AutoRegression for Image Synthesis [45.98061600562716]
  GEAR trains a vector-quantized (VQ) tokenizer and an autoregressive (AR) generator jointly and end-to-end.<n>GEAR speeds up ImageNet gFID convergence by up to 10x relative to the strong LlamaGen-REPA baseline.
  arXiv  Detail & Related papers  (2026-06-30T17:59:57Z)
• Test-Time Scaling with Diffusion Language Models via Reward-Guided Stitching [66.39914384073145]
  We propose a self-consistency framework that turns cheap diffusion-sampled reasoning into a reusable pool of step-level candidates.<n>We find that step-level recombination is most beneficial on harder problems.<n>Our training-free framework improves average accuracy by up to 2 across six math and coding tasks.
  arXiv  Detail & Related papers  (2026-02-26T11:08:39Z)
• Reversible Diffusion Decoding for Diffusion Language Models [69.10149777322108]
  Reversible Diffusion Decoding (RDD) is a decoding framework that introduces reversibility into block-wise diffusion generation.<n>RDD detects stagnation as a state-dependent failure of the reverse process and enables efficient backtracking to earlier blocks.<n> Experiments show that RDD improves generation robustness and quality over baselines with minimal computational overhead.
  arXiv  Detail & Related papers  (2026-01-29T12:52:33Z)
• From Next-Token to Next-Block: A Principled Adaptation Path for Diffusion LLMs [58.640039233470766]
  We show that principled AR-to-block-diffusion adaptation is an effective and compute-efficient alternative to training DLMs from scratch.<n> NBDiff-7B (Base and Instruct) could inherit the long-context modeling and reasoning capabilities, and achieve state-of-the-art performance.
  arXiv  Detail & Related papers  (2025-12-07T10:28:21Z)
• Uncertainty Quantification for Retrieval-Augmented Reasoning [40.43455995861054]
  Retrieval-augmented reasoning (RAR) is a recent evolution of retrieval-augmented generation (RAG) that employs multiple reasoning steps for retrieval and generation.<n>Uncertainty quantification (UQ) offers methods to estimate the confidence of systems' outputs.<n>We introduce Retrieval-Augmented Reasoning (R2C)-a novel UQ method for RAR.
  arXiv  Detail & Related papers  (2025-10-13T14:55:28Z)
• Continuous Speculative Decoding for Autoregressive Image Generation [27.308442169466975]
  Continuous visual autoregressive (AR) models have demonstrated promising performance in image generation.<n> speculative decoding has effectively accelerated discrete autoregressive inference.<n>This work addresses challenges from low acceptance rate, inconsistent output distribution, and modified distribution without analytic expression.
  arXiv  Detail & Related papers  (2024-11-18T09:19:15Z)
• FastLR: Non-Autoregressive Lipreading Model with Integrate-and-Fire [74.04394069262108]
  We propose FastLR, a non-autoregressive (NAR) lipreading model which generates all target tokens simultaneously. FastLR achieves the speedup up to 10.97$times$ compared with state-of-the-art lipreading model.
  arXiv  Detail & Related papers  (2020-08-06T08:28:56Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.