UrbanAlign: Post-hoc Semantic Calibration for VLM-Human Preference Alignment

• URL: http://arxiv.org/abs/2602.19442v2
• Date: Wed, 04 Mar 2026 03:50:10 GMT
• Title: UrbanAlign: Post-hoc Semantic Calibration for VLM-Human Preference Alignment
• Abstract summary: We show that for subjective perception tasks, this alignment can be achieved without any model training.<n>We propose a training-free post-hoc concept-bottleneck pipeline consisting of three tightly coupled stages.
• Score: 17.430091762374744
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Aligning vision-language model (VLM) outputs with human preferences in domain-specific tasks typically requires fine-tuning or reinforcement learning, both of which demand labelled data and GPU compute. We show that for subjective perception tasks, this alignment can be achieved without any model training: VLMs are already strong concept extractors but poor decision calibrators, and the gap can be closed externally. We propose a training-free post-hoc concept-bottleneck pipeline consisting of three tightly coupled stages: concept mining, multi-agent structured scoring, and geometric calibration, unified by an end-to-end dimension optimization loop. Interpretable evaluation dimensions are mined from a handful of human annotations; an Observer-Debater-Judge chain extracts robust continuous concept scores from a frozen VLM; and locally-weighted ridge regression on a hybrid visual-semantic manifold calibrates these scores against human ratings. Applied to urban perception as UrbanAlign, the framework achieves 72.2% accuracy ($κ=0.45$) on Place Pulse 2.0 across six categories, outperforming the best supervised baseline by +15.1 pp and uncalibrated VLM scoring by +16.3 pp, with full dimension-level interpretability and zero model-weight modification.

Related papers

• Back to the Feature: Zero-Shot 6DoF Pose Estimation via Dense Local Features [0.0]
  We present B2ose, a training-free zero-shot method for 6DoF pose estimation of unseen objects from RGB images.<n>B2ose extracts dense patch-level features that generalize across the synthetic-to-real domain gap without task-specific fine-tuning.<n>On the seven core datasets of the BOP Benchmark, B2ose achieves 40.7 mean AR without refinement and 56.4 refinement.
  arXiv  Detail & Related papers  (2026-09-06T16:57:27Z)
• Perturb the Thought, Not the Pixels: Latent-Space Rollout Diversification for Reinforcement Learning of Vision-Language Models [0.0]
  Noise-Contrastive GRPO injects scale-calibrated Gaussian noise into the last hidden layer of the prompt-encoding pass.<n>On Qwen2.5-VL-7B trained on Geometry3K, NC-GRPO significantly improves out-of-domain mathematical reasoning.
  arXiv  Detail & Related papers  (2026-08-21T19:56:50Z)
• GIRL: Generative Imagination Reinforcement Learning via Information-Theoretic Hallucination Control [0.0]
  We introduce GIRL (Generative Imagination Reinforcement Learning), a latent world-model framework that addresses this failure mode with two key components.<n> GIRL reduces latent rollout drift by 38 to 61 percent across tasks relative to DreamerV3, improves return, and requires fewer environment interactions on long-horizon tasks.<n>A distilled-prior variant reduces overhead and improves computational efficiency relative to the full model.
  arXiv  Detail & Related papers  (2026-04-08T17:14:21Z)
• ER-Pose: Rethinking Keypoint-Driven Representation Learning for Real-Time Human Pose Estimation [2.2248409468073143]
  Single-stage multi-person pose estimation aims to jointly perform human localization and keypoint prediction.<n>Box-driven single-stage pose estimation is implicitly constrained by bounding-box supervision during training.<n>We propose a keypoint-driven learning paradigm that elevates pose estimation to a primary prediction objective.<n>On MS COCO and CrowdPose, ER-Pose-n achieves AP improvements of 3.2/6.7 without pre-training and 7.4/4.9 with pre-training respectively.
  arXiv  Detail & Related papers  (2026-03-09T17:49:46Z)
• K-Sort Eval: Efficient Preference Evaluation for Visual Generation via Corrected VLM-as-a-Judge [51.93484138861584]
  The rapid development of visual generative models raises the need for more scalable and human-aligned evaluation methods.<n>We propose K-Sort Eval, a reliable and efficient VLM-based evaluation framework that integrates posterior correction and dynamic matching.<n>Experiments show that K-Sort Eval delivers evaluation results consistent with K-Sort Arena, typically requiring fewer than 90 model runs.
  arXiv  Detail & Related papers  (2026-02-10T05:07:46Z)
• Multi-dimensional Preference Alignment by Conditioning Reward Itself [32.33870784484853]
  Multi Reward Conditional DPO resolves reward conflicts by introducing a disentangled Bradley-Terry objective.<n>Experiments on Stable 1.5 and SDXL demonstrate that MCDPO achieves superior performance on benchmarks.
  arXiv  Detail & Related papers  (2025-12-11T02:44:31Z)
• From Noisy Traces to Stable Gradients: Bias-Variance Optimized Preference Optimization for Aligning Large Reasoning Models [90.45197506653341]
  Large reasoning models generate intermediate reasoning traces before producing final answers.<n> aligning LRMs with human preferences, a crucial prerequisite for model deployment, remains underexplored.<n>A common workaround optimized a single sampled trajectory, which introduces substantial gradient variance from trace sampling.
  arXiv  Detail & Related papers  (2025-10-06T17:58:01Z)
• GrAInS: Gradient-based Attribution for Inference-Time Steering of LLMs and VLMs [56.93583799109029]
  GrAInS is an inference-time steering approach that operates across both language-only and vision-language models and tasks.<n>During inference, GrAInS hidden activations at transformer layers guided by token-level attribution signals, and normalizes activations to preserve representational scale.<n>It consistently outperforms both fine-tuning and existing steering baselines.
  arXiv  Detail & Related papers  (2025-07-24T02:34:13Z)
• SHAPE : Self-Improved Visual Preference Alignment by Iteratively Generating Holistic Winner [35.843587407696006]
  Large Visual Language Models (LVLMs) increasingly rely on preference alignment to ensure reliability.<n>We present projectname, a self-supervised framework capable of transforming the already abundant supervised text-image pairs into holistic preference triplets.
  arXiv  Detail & Related papers  (2025-03-06T08:33:11Z)
• UniCBE: An Uniformity-driven Comparing Based Evaluation Framework with Unified Multi-Objective Optimization [19.673388630963807]
  We propose UniCBE, a unified uniformity-driven CBE framework.<n>On the AlpacaEval benchmark, UniCBE saves over 17% of evaluation budgets while achieving a Pearson correlation with ground truth exceeding 0.995.<n>In scenarios where new models are continuously introduced, UniCBE can even save over 50% of evaluation costs.
  arXiv  Detail & Related papers  (2025-02-17T05:28:12Z)
• ALOcc: Adaptive Lifting-Based 3D Semantic Occupancy and Cost Volume-Based Flow Predictions [91.55655961014027]
  3D semantic occupancy and flow prediction are fundamental to understanding scene scene.<n>This paper proposes a vision-based framework with three targeted improvements.<n>Our purely convolutional architecture establishes new SOTA performance on multiple benchmarks for both semantic occupancy and joint semantic-flow prediction.
  arXiv  Detail & Related papers  (2024-11-12T11:32:56Z)
• 360 Layout Estimation via Orthogonal Planes Disentanglement and Multi-view Geometric Consistency Perception [56.84921040837699]
  Existing panoramic layout estimation solutions tend to recover room boundaries from a vertically compressed sequence, yielding imprecise results. We propose an orthogonal plane disentanglement network (termed DOPNet) to distinguish ambiguous semantics. We also present an unsupervised adaptation technique tailored for horizon-depth and ratio representations. Our solution outperforms other SoTA models on both monocular layout estimation and multi-view layout estimation tasks.
  arXiv  Detail & Related papers  (2023-12-26T12:16:03Z)
• Uncertainty-Aware Adaptation for Self-Supervised 3D Human Pose Estimation [70.32536356351706]
  We introduce MRP-Net that constitutes a common deep network backbone with two output heads subscribing to two diverse configurations. We derive suitable measures to quantify prediction uncertainty at both pose and joint level. We present a comprehensive evaluation of the proposed approach and demonstrate state-of-the-art performance on benchmark datasets.
  arXiv  Detail & Related papers  (2022-03-29T07:14:58Z)

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.