Lookahead Path Likelihood Optimization for Diffusion LLMs

• URL: http://arxiv.org/abs/2602.03496v1
• Date: Tue, 03 Feb 2026 13:12:41 GMT
• Title: Lookahead Path Likelihood Optimization for Diffusion LLMs
• Authors: Xuejie Liu, Yap Vit Chun, Yitao Liang, Anji Liu,
• Abstract summary: We introduce path log-likelihood (Path LL), a trajectory-conditioned objective that strongly correlates with downstream accuracy and enables principled selection of unmasking paths.<n>To optimize Path LL at inference time, we propose POKE, an efficient value estimator that predicts the expected future Path LL of a partial decoding trajectory.<n>We then integrate this lookahead signal into POKE-SMC, a Sequential Monte Carlo-based search framework for dynamically identifying optimal unmasking paths.
• Score: 31.01208893976334
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Diffusion Large Language Models (dLLMs) support arbitrary-order generation, yet their inference performance critically depends on the unmasking order. Existing strategies rely on heuristics that greedily optimize local confidence, offering limited guidance for identifying unmasking paths that are globally consistent and accurate. To bridge this gap, we introduce path log-likelihood (Path LL), a trajectory-conditioned objective that strongly correlates with downstream accuracy and enables principled selection of unmasking paths. To optimize Path LL at inference time, we propose POKE, an efficient value estimator that predicts the expected future Path LL of a partial decoding trajectory. We then integrate this lookahead signal into POKE-SMC, a Sequential Monte Carlo-based search framework for dynamically identifying optimal unmasking paths. Extensive experiments across 6 reasoning tasks show that POKE-SMC consistently improves accuracy, achieving 2%--3% average gains over strong decoding-time scaling baselines at comparable inference overhead on LLaDA models and advancing the accuracy--compute Pareto frontier.

Related papers

• LFPO: Likelihood-Free Policy Optimization for Masked Diffusion Models [48.68246945083386]
  Likelihood-Free Policy Optimization (LFPO) is a native framework that maps the concept of vector field flow matching to the discrete token space.<n>LFPO formulates alignment as geometric velocity rectification, which directly optimize denoising logits via contrastive updates.<n>Experiments demonstrate that LFPO not only outperforms state-of-the-art baselines on code and reasoning benchmarks but also accelerates inference by approximately 20% through reduced diffusion steps.
  arXiv  Detail & Related papers  (2026-03-02T07:42:55Z)
• Zero-Order Optimization for LLM Fine-Tuning via Learnable Direction Sampling [40.94400211806987]
  We propose a policy-driven ZO framework that treats the sampling distribution over perturbation directions as a learnable policy.<n>We show that learned sampling improves quality gradient information and relax the explicit dependence on $d$ in convergence bounds.<n>Our results suggest that adaptive direction sampling is a promising route to make ZO fine-tuning viable at scale.
  arXiv  Detail & Related papers  (2026-02-14T08:01:41Z)
• Predicting LLM Output Length via Entropy-Guided Representations [13.351384070796747]
  We introduce a lightweight framework that reuses the main model's internal hidden states for efficient length prediction.<n>Our framework features two core components: 1) Entropy-Guided Token Pooling (EGTP), which uses on-the-fly activations and token entropy for highly accurate static prediction.
  arXiv  Detail & Related papers  (2026-02-12T10:49:04Z)
• Lookahead Unmasking Elicits Accurate Decoding in Diffusion Language Models [51.12873073612084]
  Masked Diffusion Models (MDMs) as language models generate by iteratively unmasking tokens, yet their performance depends on the inference time order of unmasking.<n>We propose Lookahead Unmasking (LookUM), which addresses these concerns by reformulating sampling as path selection over all possible unmasking orders.<n>LookUM requires only two to three paths to achieve peak performance, demonstrating remarkably efficient path selection.
  arXiv  Detail & Related papers  (2025-11-04T02:37:37Z)
• Boundary-Guided Policy Optimization for Memory-efficient RL of Diffusion Large Language Models [53.339700196282905]
  A key challenge in applying reinforcement learning to large language models (dLLMs) is the intractability of their likelihood functions.<n>We propose a memory-efficient RL algorithm that maximizes a specially constructed lower bound of the ELBO-based objective.<n> Experiments show that BGPO significantly outperforms previous RL algorithms for dLLMs in math problem solving, code generation, and planning tasks.
  arXiv  Detail & Related papers  (2025-10-13T17:47:50Z)
• DAMR: Efficient and Adaptive Context-Aware Knowledge Graph Question Answering with LLM-Guided MCTS [28.828541350757714]
  This paper proposes Dynamically Adaptive MCTS-based Reasoning (DAMR) for Knowledge Graph Question Answering (KGQA)<n>DAMR integrates Monte Carlo Tree Search (MCTS) with adaptive path evaluation to enable context-aware KGQA.<n>Experiments on multiple KGQA benchmarks show DAMR significantly outperforms SOTA methods.
  arXiv  Detail & Related papers  (2025-08-01T15:38:21Z)
• Rational Tuning of LLM Cascades via Probabilistic Modeling [0.9208007322096532]
  We present a probabilistic model for the joint performance distribution of a sequence of large language models (LLMs)<n>Compared to selecting confidence thresholds using Bayesian optimization, our Markov parametric-copula model yields more favorable error-cost trade-offs.<n>Our framework's inductive assumptions about the interactions between the error rates of different LLMs enhance sample efficiency.
  arXiv  Detail & Related papers  (2025-01-16T07:58:33Z)
• Decoding-Time Language Model Alignment with Multiple Objectives [116.42095026960598]
  Existing methods primarily focus on optimizing LMs for a single reward function, limiting their adaptability to varied objectives. Here, we propose $textbfmulti-objective decoding (MOD)$, a decoding-time algorithm that outputs the next token from a linear combination of predictions. We show why existing approaches can be sub-optimal even in natural settings and obtain optimality guarantees for our method.
  arXiv  Detail & Related papers  (2024-06-27T02:46:30Z)
• LLM-A*: Large Language Model Enhanced Incremental Heuristic Search on Path Planning [91.95362946266577]
  Path planning is a fundamental scientific problem in robotics and autonomous navigation.<n>Traditional algorithms like A* and its variants are capable of ensuring path validity but suffer from significant computational and memory inefficiencies as the state space grows.<n>We propose a new LLM based route planning method that synergistically combines the precise pathfinding capabilities of A* with the global reasoning capability of LLMs.<n>This hybrid approach aims to enhance pathfinding efficiency in terms of time and space complexity while maintaining the integrity of path validity, especially in large-scale scenarios.
  arXiv  Detail & Related papers  (2024-06-20T01:24:30Z)
• Bypass Back-propagation: Optimization-based Structural Pruning for Large Language Models via Policy Gradient [57.9629676017527]
  We propose an optimization-based structural pruning that learns the pruning masks in a probabilistic space directly by optimizing the loss of the pruned model.<n>We achieve this by learning an underlying Bernoulli distribution to sample binary pruning masks.<n>Experiments conducted on LLaMA, LLaMA-2, LLaMA-3, Vicuna, and Mistral models demonstrate the promising performance of our method in efficiency and effectiveness.
  arXiv  Detail & Related papers  (2024-06-15T09:31:03Z)
• Low-Pass Filtering SGD for Recovering Flat Optima in the Deep Learning Optimization Landscape [15.362190838843915]
  We show that LPF-SGD converges to a better optimal point with smaller generalization error than SGD. We show that our algorithm achieves superior generalization performance compared to the common DL training strategies.
  arXiv  Detail & Related papers  (2022-01-20T07:13:04Z)

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.