BLADE: Scalable Bi-level Adaptive Data Selection for LLM Training

As Large Language Model (LLM) datasets scale to trillions of tokens, data selection has emerged as a critical frontier to filter out uninformative noise and construct adaptive learning trajectories. Beyond static heuristic filtering, advanced data selection methods for LLM training largely follow two paradigms, each with fundamental limitations. Influence-based methods provide principled bi-level objectives but require intractable inverse-Hessian computations, while excess-loss methods are computationally efficient but rely on a static reference model that becomes misaligned with the evolving proxy model during training. We propose BLADE (Bi-Level Adaptive Data sElection), a Hessian-free framework for data selection. BLADE reformulates the bi-level optimization problem underlying influence-based methods as a penalized single-level objective via Lagrange multipliers, avoiding inverse-Hessian computation while revealing a principled connection to excess-loss based data selection. The resulting objective recovers an excess-loss form but replaces the static reference model with a dynamic one that stays synchronized with training. Theoretically, we prove that this penalized formulation guarantees first-order convergence. For efficient online batch selection, we instantiate BLADE as a memoryless randomized block-coordinate Frank-Wolfe algorithm. Extensive experiments show that BLADE consistently outperforms state-of-the-art data selection baselines, providing a practical recipe for LLM training.

Unlocking the Potential of Movable Antennas: General and Practical Antenna Position Optimization

Recently, movable antenna (MA) has attracted wide attention in wireless communications due to its potential in enhancing wireless communication performance via local movement within a confined region. However, antenna position optimization (APO) has emerged as a major challenge for MAs, due to the lack of a tractable, analytical, and accurate channel model in terms of antenna positions. Although existing works have developed various algorithms for APO, most of them are based on simplified theoretical channel models, which limit their generality. To address this challenge, in this article, we present more general and effective APO algorithms for different purposes, categorized as continuous APO and discrete APO, respectively. Continuous APO is mainly applied for flexible array signal processing to boost large-scale communication performance, while discrete APO is applied for small-scale multi-path channel reshaping. Specifically, the discrete APO discretizes the antenna movement region into multiple sampling points and employs discrete algorithms to determine the optimal MA positions based on the point-wise channel state information (CSI), without the need for an analytical channel model. To reduce the overhead for CSI acquisition, we also present more efficient learning-based APO algorithms that operate without requiring full point-wise CSI. Finally, we compare the application scenarios of the proposed algorithms and validate their effectiveness with numerical results.

Mixture-of-Experts Diffusion Models for Adaptive Massive MIMO Channel Estimation via Variational Bayesian Inference

Channel estimation is essential to massive multiple-input multiple-output (MIMO) systems. While recent generative model-based approaches using lightweight diffusion models (DMs) have achieved superior performance, they typically rely on a single data-driven prior, which limits their adaptability to varying channel distributions in real-world scenarios. To address this deficiency, we propose a mixture-of-experts (MoE) diffusion model (DM) framework combined with variational Bayesian inference. Specifically, our approach employs multiple pre-trained DMs, with each trained on a specific type of propagation channels. We then propose a probabilistic graphical model in which the channel is modeled as a latent variable drawn from one of these candidate generative priors with a certain probability. By integrating variational Bayesian inference with DM-based data priors, the underlying channel along with the expert indicator variable are jointly inferred, thus enabling automatic model adaptation for channel estimation. The effectiveness of our approach is evaluated on 3GPP CDL channels. Simulation results demonstrate that our proposed approach achieves a clear performance improvement over the standard DM-based method that employs a single prior trained on aggregated data from all channel types, particularly when the channel samples from different propagation environments are imbalanced.

Terahertz Beam Squint Mitigation via Six-Dimensional Movable Antennas

Analog beamforming holds great potential for future terahertz (THz) communications due to its ability to generate high-gain directional beams with low-cost phase shifters. However, conventional analog beamforming may suffer substantial performance degradation in wideband systems due to the beam squint effect. Instead of relying on high-cost true-time delayers, we propose an efficient six-dimensional movable antenna (6DMA) architecture to mitigate the beam-squint effect. In particular, we study a wideband wide-beam coverage problem in this paper, aiming to maximize the minimum beamforming gain over a given range of azimuth/elevation angles and frequencies by jointly optimizing the analog beamforming vector, the MA positions within a two-dimensional (2D) region, and the three-dimensional (3D) rotation angles of the antenna array. However, this problem is non-convex and intractable to solve optimally due to the coupling of the spatial and frequency domains and that of the antenna weights, positions and rotation. To tackle this problem, we first derive an optimal solution to it in a special case with azimuth or elevation angle coverage only. It is shown that rotating a uniform linear array (ULA) is sufficient to achieve global optimality and eliminate beam-squint effects. While for other general cases, an alternating optimization (AO) algorithm is proposed to obtain a high-quality suboptimal solution, where the antennas' beamforming weights, positions, and rotation angles are alternately optimized by combining successive convex approximation (SCA), sequential update with Gibbs sampling (GS), and hybrid coarse- and fine-grained search. Simulation results demonstrate that our proposed scheme can significantly outperform conventional antenna arrays without antenna movement or rotation, thus offering a cost-effective solution for wideband transmission over THz bands.

Compress the Easy, Explore the Hard: Difficulty-Aware Entropy Regularization for Efficient LLM Reasoning

Chain-of-Thought (CoT) has substantially empowered Large Language Models (LLMs) to tackle complex reasoning tasks, yet the verbose nature of explicit reasoning steps incurs prohibitive inference latency and computational costs, limiting real-world deployment. While existing compression methods - ranging from self-training to Reinforcement Learning (RL) with length constraints - attempt to mitigate this, they often sacrifice reasoning capability for brevity. We identify a critical failure mode in these approaches: explicitly optimizing for shorter trajectories triggers rapid entropy collapse, which prematurely shrinks the exploration space and stifles the discovery of valid reasoning paths, particularly for challenging questions requiring extensive deduction. To address this issue, we propose Compress responses for Easy questions and Explore Hard ones (CEEH), a difficulty-aware approach to RL-based efficient reasoning. CEEH dynamically assesses instance difficulty to apply selective entropy regularization: it preserves a diverse search space for currently hard questions to ensure robustness, while permitting aggressive compression on easier instances where the reasoning path is well-established. In addition, we introduce a dynamic optimal-length penalty anchored to the historically shortest correct response, which effectively counteracts entropy-induced length inflation and stabilizes the reward signal. Across six reasoning benchmarks, CEEH consistently reduces response length while maintaining accuracy comparable to the base model, and improves Pass@k relative to length-only optimization.

WavBench: Benchmarking Reasoning, Colloquialism, and Paralinguistics for End-to-End Spoken Dialogue Models

With the rapid integration of advanced reasoning capabilities into spoken dialogue models, the field urgently demands benchmarks that transcend simple interactions to address real-world complexity. However, current evaluations predominantly adhere to text-generation standards, overlooking the unique audio-centric characteristics of paralinguistics and colloquialisms, alongside the cognitive depth required by modern agents. To bridge this gap, we introduce WavBench, a comprehensive benchmark designed to evaluate realistic conversational abilities where prior works fall short. Uniquely, WavBench establishes a tripartite framework: 1) Pro subset, designed to rigorously challenge reasoning-enhanced models with significantly increased difficulty; 2) Basic subset, defining a novel standard for spoken colloquialism that prioritizes "listenability" through natural vocabulary, linguistic fluency, and interactive rapport, rather than rigid written accuracy; and 3) Acoustic subset, covering explicit understanding, generation, and implicit dialogue to rigorously evaluate comprehensive paralinguistic capabilities within authentic real-world scenarios. Through evaluating five state-of-the-art models, WavBench offers critical insights into the intersection of complex problem-solving, colloquial delivery, and paralinguistic fidelity, guiding the evolution of robust spoken dialogue models. The benchmark dataset and evaluation toolkit are available at https://naruto-2024.github.io/wavbench.github.io/.

Spectral Disentanglement and Enhancement: A Dual-domain Contrastive Framework for Representation Learning

Large-scale multimodal contrastive learning has recently achieved impressive success in learning rich and transferable representations, yet it remains fundamentally limited by the uniform treatment of feature dimensions and the neglect of the intrinsic spectral structure of the learned features. Empirical evidence indicates that high-dimensional embeddings tend to collapse into narrow cones, concentrating task-relevant semantics in a small subspace, while the majority of dimensions remain occupied by noise and spurious correlations. Such spectral imbalance and entanglement undermine model generalization. We propose Spectral Disentanglement and Enhancement (SDE), a novel framework that bridges the gap between the geometry of the embedded spaces and their spectral properties. Our approach leverages singular value decomposition to adaptively partition feature dimensions into strong signals that capture task-critical semantics, weak signals that reflect ancillary correlations, and noise representing irrelevant perturbations. A curriculum-based spectral enhancement strategy is then applied, selectively amplifying informative components with theoretical guarantees on training stability. Building upon the enhanced features, we further introduce a dual-domain contrastive loss that jointly optimizes alignment in both the feature and spectral spaces, effectively integrating spectral regularization into the training process and encouraging richer, more robust representations. Extensive experiments on large-scale multimodal benchmarks demonstrate that SDE consistently improves representation robustness and generalization, outperforming state-of-the-art methods. SDE integrates seamlessly with existing contrastive pipelines, offering an effective solution for multimodal representation learning.

Agent-Omit: Training Efficient LLM Agents for Adaptive Thought and Observation Omission via Agentic Reinforcement Learning

Managing agent thought and observation during multi-turn agent-environment interactions is an emerging strategy to improve agent efficiency. However, existing studies treat the entire interaction trajectories equally, overlooking the thought necessity and observation utility varies across turns. To this end, we first conduct quantitative investigations into how thought and observation affect agent effectiveness and efficiency. Based on our findings, we propose Agent-Omit, a unified training framework that empowers LLM agents to adaptively omit redundant thoughts and observations. Specifically, we first synthesize a small amount of cold-start data, including both single-turn and multi-turn omission scenarios, to fine-tune the agent for omission behaviors. Furthermore, we introduce an omit-aware agentic reinforcement learning approach, incorporating a dual sampling mechanism and a tailored omission reward to incentivize the agent's adaptive omission capability. Theoretically, we prove that the deviation of our omission policy is upper-bounded by KL-divergence. Experimental results on five agent benchmarks show that our constructed Agent-Omit-8B could obtain performance comparable to seven frontier LLM agent, and achieve the best effectiveness-efficiency trade-off than seven efficient LLM agents methods. Our code and data are available at https://github.com/usail-hkust/Agent-Omit.

Darwinian Memory: A Training-Free Self-Regulating Memory System for GUI Agent Evolution

Multimodal Large Language Model (MLLM) agents facilitate Graphical User Interface (GUI) automation but struggle with long-horizon, cross-application tasks due to limited context windows. While memory systems provide a viable solution, existing paradigms struggle to adapt to dynamic GUI environments, suffering from a granularity mismatch between high-level intent and low-level execution, and context pollution where the static accumulation of outdated experiences drives agents into hallucination. To address these bottlenecks, we propose the Darwinian Memory System (DMS), a self-evolving architecture that constructs memory as a dynamic ecosystem governed by the law of survival of the fittest. DMS decomposes complex trajectories into independent, reusable units for compositional flexibility, and implements Utility-driven Natural Selection to track survival value, actively pruning suboptimal paths and inhibiting high-risk plans. This evolutionary pressure compels the agent to derive superior strategies. Extensive experiments on real-world multi-app benchmarks validate that DMS boosts general-purpose MLLMs without training costs or architectural overhead, achieving average gains of 18.0% in success rate and 33.9% in execution stability, while reducing task latency, establishing it as an effective self-evolving memory system for GUI tasks.

Integrated Channel Estimation and Sensing for Near-Field ELAA Systems

In this paper, we study the problem of uplink channel estimation for near-filed orthogonal frequency division multiplexing (OFDM) systems, where a base station (BS), equipped with an extremely large-scale antenna array (ELAA), serves multiple users over the same time-frequency resource block. A non-orthogonal pilot transmission scheme is considered to accommodate a larger number of users that can be supported by ELAA systems without incurring an excessive amount of training overhead. To facilitate efficient multi-user channel estimation, we express the received signal as a third-order low-rank tensor, which admits a canonical polyadic decomposition (CPD) model for line-of-sight (LoS) scenarios and a block term decomposition (BTD) model for non-line-of-sight (NLoS) scenarios. An alternating least squares (ALS) algorithm and a non-linear least squares (NLS) algorithm are employed to perform CPD and BTD, respectively. Channel parameters are then efficiently extracted from the recovered factor matrices. By exploiting the geometry of the propagation paths in the estimated channel, users' positions can be precisely determined in LoS scenarios. Moreover, our uniqueness analysis shows that the proposed tensor-based joint multi-user channel estimation framework is effective even when the number of pilot symbols is much smaller than the number of users, revealing its potential in training overhead reduction. Simulation results demonstrate that the proposed method achieves markedly higher channel estimation accuracy than compressed sensing (CS)-based approaches.