Rubric Dropout: A Simple Way to Mitigate Reward Hacking in Rubric-as-Reward RL

Reinforcement learning against rubrics, lists of criteria graded by an LLM judge, has become a standard way to post-train language models on tasks with no deterministic answer. The rubric, however, is a fixed proxy for quality, never a complete description of it, and a policy trained against it long enough will learn to exploit the difference. We measure this directly. Training Qwen3-8B with Group Relative Policy Optimization (GRPO) on medical and science rubrics and grading out-of-distribution (OOD) benchmarks with both the training judge and a stronger gold judge, we find that the two scores diverge during training. The training judge's score keeps climbing while the gold judge's score peaks and then falls, by 3 points on HealthBench-Hard and by 22 points on ResearchQA. A judge with a fixed bias would shift the gold curve by a constant, not send it down while the training score rises, so the divergence is reward hacking, not judge noise. We propose Rubric Dropout, a one-line fix borrowed from neuron dropout. At every step, we randomly drop a subset of the rubric's criteria before computing the reward, so the policy never optimizes the same rubric twice. The dropped subset is shared across each rollout group, so GRPO's group-relative advantages stay comparable, and evaluation always uses the full rubric. Comparing no dropout against dropout at 30% and 50% on both benchmark pairs, dropout raises the OOD gold score at every matched checkpoint (+1 to +2 points on HealthBench-Hard, +6 to +7 points on ResearchQA), lowers the two hacking measures we track, and costs nothing in domain. Sweeping the dropout fraction shows a broad 30-50% sweet spot, while the natural alternative, reweighting criteria by how useful they are to training, performs worse than no intervention at all in our setting.

CFM-Bench: A Unified Multi-Domain, Multi-Task Benchmark for Channel Foundation Models

Channel foundation models (CFMs) are developing rapidly, with recent studies reporting benefits from pretraining across downstream wireless tasks. Yet CFMs are commonly evaluated in model-specific pipelines with different data, radio configurations, partitions, adaptation procedures, task definitions, and metrics. Reported comparisons therefore tend to show that pretraining improves over supervised training from scratch within one pipeline, but neither rank CFMs nor compare them fairly with task-specific models. We release CFM-Bench, a unified multi-domain, multi-task benchmark designed to address this gap. It curates six channel configurations spanning 3GPP statistical simulation, two independent ray-tracing pipelines, industrial and aerial measurements, and synchronized vehicular multimodal simulation. Official partitions isolate complete trajectories, measurement sessions, vehicle links, simulation realizations, or buffered spatial regions. CFM-Bench does not prescribe an external pretraining corpus or strategy; no benchmark split may be used for foundation-model pretraining, and the official training split is reserved exclusively for downstream fine-tuning. The benchmark additionally requires disclosure of all data used during model development and prohibits training-stage use of official test units. Six task groups are organized along three CFM application dimensions: physical-layer (PHY) channel intelligence, radio-access-network (RAN) decision intelligence, and integrated sensing and communication (ISAC). They cover CSI feedback, frequency and temporal channel extrapolation, propagation-state classification, current- and future-beam prediction, and single-frame and temporal localization. CFM-Bench provides a common substrate for comparing the transferability of channel representations across models, domains, and tasks.

Feeling the Unexpected: ResTacVLA for Contact-Rich Manipulation via Residual Tactile Representation

Tactile perception is indispensable for contact-rich manipulation, yet integrating it into Vision-Language-Action (VLA) models often induces modality collapse, where high-bandwidth visual features overshadow sparse tactile cues. Inspired by Predictive Coding, a neural mechanism where the brain attenuates predictable inputs to prioritize surprising stimuli, we propose ResTacVLA. Rather than treating tactile data as raw input, we reformulate it as a Residual Tactile Representation capturing the discrepancy between visual priors and physical sensations. By filtering out visually predictable dynamics, this formulation transforms sparse tactile signals into dense, high-value information gain, thereby inherently resolving the bandwidth mismatch. These residuals are discretized through a Vector Quantized (VQ) bottleneck into Latent Contact Primitives that capture critical events missed by vision. Analogous to the neural surprise signal, we leverage the uncertainty of the visual prior to adaptively gate tactile integration, prioritizing residuals specifically during visually unreliable phases to explicitly prevent visual dominance. Experimental results show that ResTacVLA consistently outperforms all baselines on a diverse set of contact-rich manipulation tasks, while remaining robust to unexpected dynamic disturbances. Project page: https://awilekong.github.io/ResTacVLA-Website/

Dr. DocBench: A Comprehensive Benchmark for Expert-Level and Difficult Document Parsing

Document parsing and recognition are fundamental capabilities for vision-language models (VLMs) and document processing systems. However, existing Optical Character Recognition (OCR) and document parsing benchmarks are increasingly limited in coverage and difficulty: many focus on common document genres or uniformly sampled pages where modern parsers already perform strongly, while offering limited annotation for expert-domain structures such as chemical formula, music notation, complex tables, and cross-page layouts. We introduce Dr. DocBench, a difficulty-aware benchmark for expert-level document parsing. Built from a large-scale multilingual book corpus, Dr. DocBench spans 52 BISAC subject domains and selects challenging documents through parser-failure-based sampling, targeting cases where multiple state-of-the-art systems struggle. It contains 4,514 annotated pages from long documents averaging around 100 pages, with 65k high-quality page- and block-level annotations for layout, reading order, hierarchical relations, and domain-specific visual contents. Evaluations of pipeline-based parsers and general-purpose VLMs show that strong performance on existing benchmarks does not transfer to our expert-level document parsing. Our analysis reveals substantial failures across subjects, content types, and structural attributes, highlighting Dr. DocBench as a comprehensive testbed for diagnosing and advancing document intelligence.

PriorVLA: Prior-Preserving Adaptation for Vision-Language-Action Models

Large-scale pretraining has made Vision-Language-Action (VLA) models promising foundations for generalist robot manipulation, yet adapting them to downstream tasks remains necessary. However, the common practice of full fine-tuning treats pretraining as initialization and can shift broad priors toward narrow training-distribution patterns. We propose PriorVLA, a novel framework that preserves pretrained priors and learns to leverage them for effective adaptation. PriorVLA keeps a frozen Prior Expert as a read-only prior source and trains an Adaptation Expert for downstream specialization. Expert Queries capture scene priors from the pretrained VLM and motor priors from the Prior Expert, integrating both into the Adaptation Expert to guide adaptation. Together, PriorVLA updates only 25% of the parameters updated by full fine-tuning. Across RoboTwin 2.0, LIBERO, and real-world tasks, PriorVLA achieves stronger overall performance than full fine-tuning and state-of-the-art VLA baselines, with the largest gains under out-of-distribution (OOD) and few-shot settings. PriorVLA improves over pi0.5 by 11 points on RoboTwin 2.0-Hard and achieves 99.1% average success on LIBERO. Across eight real-world tasks and two embodiments, PriorVLA reaches 81% in-distribution (ID) and 57% OOD success with standard data. With only 10 demonstrations per task, PriorVLA reaches 48% ID and 32% OOD success, surpassing pi0.5 by 24 and 22 points, respectively.

TextReasoningBench: Does Reasoning Really Improve Text Classification in Large Language Models?

Eliciting explicit, step-by-step reasoning traces from large language models (LLMs) has emerged as a dominant paradigm for enhancing model capabilities. Although such reasoning strategies were originally designed for problems requiring explicit multi-step reasoning, they have increasingly been applied to a broad range of NLP tasks. This expansion implicitly assumes that deliberative reasoning uniformly benefits heterogeneous tasks. However, whether such reasoning mechanisms truly benefit classification tasks remains largely underexplored, especially considering their substantial token and time costs. To fill this gap, we introduce TextReasoningBench, a systematic benchmark designed to evaluate the effectiveness and efficiency of reasoning strategies for text classification with LLMs. We compare seven reasoning strategies, namely IO, CoT, SC-CoT, ToT, GoT, BoC, and long-CoT across ten LLMs on five text classification datasets. Beyond traditional metrics such as accuracy and macro-F1, we introduce two cost-aware evaluation metrics that quantify the performance gain per reasoning token and the efficiency of performance improvement relative to token cost growth. Experimental results reveal three notable findings: (1) Reasoning does not universally improve classification performance: while moderate strategies such as CoT and SC-CoT yield consistent but limited gains (typically +1% to +3% on big models), more complex methods (e.g., ToT and GoT) often fail to outperform simpler baselines and can even degrade performance, especially on small models; (2) Reasoning is often inefficient: many reasoning strategies increase token consumption by 10$\times$ to 100$\times$ (e.g., SC-CoT and ToT) while providing only marginal performance improvements.

GenVideoLens: Where LVLMs Fall Short in AI-Generated Video Detection?

In recent years, AI-generated videos have become increasingly realistic and sophisticated. Meanwhile, Large Vision-Language Models (LVLMs) have shown strong potential for detecting such content. However, existing evaluation protocols largely treat the task as a binary classification problem and rely on coarse-grained metrics such as overall accuracy, providing limited insight into where LVLMs succeed or fail. To address this limitation, we introduce GenVideoLens, a fine-grained benchmark that enables dimension-wise evaluation of LVLM capabilities in AI-generated video detection. The benchmark contains 400 highly deceptive AI-generated videos and 100 real videos, annotated by experts across 15 authenticity dimensions covering perceptual, optical, physical, and temporal cues. We evaluate eleven representative LVLMs on this benchmark. Our analysis reveals a pronounced dimensional imbalance. While LVLMs perform relatively well on perceptual cues, they struggle with optical consistency, physical interactions, and temporal-causal reasoning. Model performance also varies substantially across dimensions, with smaller open-source models sometimes outperforming stronger proprietary models on specific authenticity cues. Temporal perturbation experiments further show that current LVLMs make limited use of temporal information. Overall, GenVideoLens provides diagnostic insights into LVLM behavior, revealing key capability gaps and offering guidance for improving future AI-generated video detection systems.

Near-Field Positioning for XL-MIMO Uniform Circular Arrays: An Attention-Enhanced Deep Learning Approach

In the evolving landscape of sixth-generation (6G) mobile communication, multiple-input multiple-output (MIMO) systems are incorporating an unprecedented number of antenna elements, advancing towards Extremely large-scale multiple-input-multiple-output (XL-MIMO) systems. This enhancement significantly increases the spatial degrees of freedom, offering substantial benefits for wireless positioning. However, the expansion of the near-field range in XL-MIMO challenges the traditional far-field assumptions used in previous MIMO models. Among various configurations, uniform circular arrays (UCAs) demonstrate superior performance by maintaining constant angular resolution, unlike linear planar arrays. Addressing how to leverage the expanded aperture and harness the near-field effects in XL-MIMO systems remains an area requiring further investigation. In this paper, we introduce an attention-enhanced deep learning approach for precise positioning. We employ a dual-path channel attention mechanism and a spatial attention mechanism to effectively integrate channel-level and spatial-level features. Our comprehensive simulations show that this model surpasses existing benchmarks such as attention-based positioning networks (ABPN), near-field positioning networks (NFLnet), convolutional neural networks (CNN), and multilayer perceptrons (MLP). The proposed model achieves superior positioning accuracy by utilizing covariance metrics of the input signal. Also, simulation results reveal that covariance metric is advantageous for positioning over channel state information (CSI) in terms of positioning accuracy and model efficiency.

Efficient UAV trajectory prediction: A multi-modal deep diffusion framework

To meet the requirements for managing unauthorized UAVs in the low-altitude economy, a multi-modal UAV trajectory prediction method based on the fusion of LiDAR and millimeter-wave radar information is proposed. A deep fusion network for multi-modal UAV trajectory prediction, termed the Multi-Modal Deep Fusion Framework, is designed. The overall architecture consists of two modality-specific feature extraction networks and a bidirectional cross-attention fusion module, aiming to fully exploit the complementary information of LiDAR and radar point clouds in spatial geometric structure and dynamic reflection characteristics. In the feature extraction stage, the model employs independent but structurally identical feature encoders for LiDAR and radar. After feature extraction, the model enters the Bidirectional Cross-Attention Mechanism stage to achieve information complementarity and semantic alignment between the two modalities. To verify the effectiveness of the proposed model, the MMAUD dataset used in the CVPR 2024 UG2+ UAV Tracking and Pose-Estimation Challenge is adopted as the training and testing dataset. Experimental results show that the proposed multi-modal fusion model significantly improves trajectory prediction accuracy, achieving a 40% improvement compared to the baseline model. In addition, ablation experiments are conducted to demonstrate the effectiveness of different loss functions and post-processing strategies in improving model performance. The proposed model can effectively utilize multi-modal data and provides an efficient solution for unauthorized UAV trajectory prediction in the low-altitude economy.

Dynamic Channel Charting: An LSTM-AE-based Approach

With the development of the sixth-generation (6G) communication system, Channel State Information (CSI) plays a crucial role in improving network performance. Traditional Channel Charting (CC) methods map high-dimensional CSI data to low-dimensional spaces to help reveal the geometric structure of wireless channels. However, most existing CC methods focus on learning static geometric structures and ignore the dynamic nature of the channel over time, leading to instability and poor topological consistency of the channel charting in complex environments. To address this issue, this paper proposes a novel time-series channel charting approach based on the integration of Long Short-Term Memory (LSTM) networks and Auto encoders (AE) (LSTM-AE-CC). This method incorporates a temporal modeling mechanism into the traditional CC framework, capturing temporal dependencies in CSI using LSTM and learning continuous latent representations with AE. The proposed method ensures both geometric consistency of the channel and explicit modeling of the time-varying properties. Experimental results demonstrate that the proposed method outperforms traditional CC methods in various real-world communication scenarios, particularly in terms of channel charting stability, trajectory continuity, and long-term predictability.