arXiv subjects

Yilong Chen

Publications and source records attributed to Yilong Chen.

At least 19 recordsLinked to original sources

arXiv2026

LLM Agents for Time-Series: A Survey

LLM-based agents are increasingly being developed for time-series problems, but their design choices vary substantially across task settings. This survey adopts a problem-driven taxonomy that organizes these systems by the time-series problems they address rather than by isolated technical components. We group existing systems into four categories: forecasting and reasoning, augmentation and synthesis, anomaly detection and diagnosis, and decision support. Within each category, we examine how task requirements shape agent architecture, tool use, and memory design. We further summarize representative datasets and environments, and compare reported model performance under shared or closely related settings. Overall, this survey offers a task-oriented guide to designing LLM-based agents for time-series problems and identifies open gaps for future work.

cs.AI↗

arXiv2026

Memory Augmentation Unlocks Efficient Chain-of-Thought Reasoning

Large language models often rely on Chain-of-Thought (CoT) reasoning to solve complex tasks, but verbose reasoning traces introduce substantial inference overhead. CoT compression shortens generation, yet aggressive compression may disrupt logical coherence and degrade performance. We formalize this trade-off as the Context-Generation Substitution Law, where explicit reasoning context substitutes for part of decode-time generation. Based on this principle, we propose Memory-Augmented Compression, a training-free framework that constructs reusable reasoning memories from historical traces and retrieves them as prefill-side scaffolds. Rather than using raw demonstrations, these memories summarize reusable reasoning patterns, key constraints, and critical operations to compensate for information lost during compression. Experiments show that Memory consistently improves prompt-based Chain-of-Draft (CoD) compression across mathematical reasoning, complex reasoning, and science question answering tasks, yielding accuracy gains of 21.4, 28.0, 29.5, and 6.61 points over CoD on GSM8K, MATH, BBH, and MMLU-Sci, while achieving a 1.14-1.49x latency speedup latency speedup over standard CoT. Memory is also compatible with token-level, reasoning-trace-level, and inference-state compression mechanisms.

cs.CL↗

arXiv2026

Construction and Dynamic Update of Channel Gain Maps via 3D Gaussian Splatting

Channel knowledge maps (CKMs) have emerged as a promising technique for providing scene-specific and location-dependent propagation knowledge to enable environment-aware wireless network design. This paper investigates the construction and dynamic updating of a particular type of CKM, namely grid-based channel gain maps (CGMs), for large-scale networks using three-dimensional Gaussian splatting (3DGS). First, we formulate a grid-based channel gain model, where each map entry is defined as the locally averaged channel gain over a receiver grid, thereby suppressing phase-sensitive small-scale fluctuations. The resulting channel gain is decomposed into distance-dependent attenuation, path transmittance, and effective scattering contributions. Based on this decomposition, we develop a physics-informed Gaussian-splatting-based channel gain (GS-CG) model, which represents the propagation environment as a set of Gaussian primitives. The proposed model maps Gaussian geometry, opacity, and directional features to propagation-related factors and renders grid-level channel gains through a differentiable process. To accommodate real-time environmental changes, we further propose an incremental learning mechanism that updates a static reference GS-CG representation into a dynamic CGM. Specifically, the reference Gaussian primitives are frozen, while a compact set of tunable Gaussians is introduced to capture newly induced local channel-gain variations from sparse measurements.Numerical results demonstrate that the proposed GS-CG methods accurately reconstruct grid-based CGMs, efficiently adapt to dynamic environmental changes, and achieve a favorable accuracy-complexity tradeoff for fast CGM refinement.

cs.IT↗

arXiv2026

WARA: A Closed-Loop Multi-Agent Framework for Wireless Optimization Autoresearch

Large language model (LLM) agents have shown growing capabilities in tool use, code execution, artifact inspection, and iterative revision, creating new opportunities for automating scientific research. To the best of our knowledge, this paper presents the first end-to-end autoresearch framework for the wireless domain, with a particular focus on wireless resource allocation optimization, an essential area for characterizing the fundamental performance limits of wireless systems and enhancing their practical performance under dynamic channel and network conditions. Specifically, we propose the Wireless AutoResearch Agent (WARA), a closed-loop multi-agent system that transforms an initial research topic into a complete research package. WARA organizes the research workflow into three phases: 1) research gap identification and problem proposal, 2) optimization modeling, algorithm design, and experimentation, and 3) research deliverable construction. Each phase follows an artifact-mediated process, in which structured upstream artifacts are consumed to generate downstream outputs. Controller-managed gates validate these artifacts and maintain consistency among problem formulations, algorithms, experiments, and research claims. When validation fails, WARA repairs only the affected artifact instead of restarting the entire workflow. We further design an LLM-based ScoringAgent to evaluate manuscript-level research validity. Comparative results show that WARA substantially outperforms one-shot LLM generation and approaches the quality profile of recently accepted peer-reviewed papers. These results demonstrate the potential of closed-loop artifact control for end-to-end LLM-assisted wireless optimization research. The source code is available at https://github.com/guoyuan-dotcom/WARA_CUHKSZ

eess.SP↗

arXiv2026

Loop the Loopies!

We present the Loopie series, consisting of two Mixture-of-Experts (MoE) models: a 20B-parameter model with 2B active parameters and a 6B-parameter model with 0.6B active parameters. Looped Transformers have long faced a challenge: given an N times increase in pre-training compute, increasing the parameter count by a factor of N usually outperforms looping a model N times. Loopie addresses this challenge. Extensive ablation studies, including comparisons with a vanilla 30B-A3B model, show that Loopie substantially outperforms vanilla Transformer baselines trained with the same compute budget. With a novel post-training method, Loopie develops strong reasoning abilities and achieves frontier-level reasoning performance.

cs.CL↗

arXiv2026

Dynamic Object Detection and Tracking in Construction: A Fisheye Camera and LiDAR Sensor Fusion Model

Robust dynamic object detection and tracking are essential for enabling robots to operate safely and effectively alongside humans in complex environments such as construction sites. While LiDAR-based SLAM and occupancy grid methods offer viable solutions for detecting and tracking motion, many state-of-the-art 3D vision approaches rely heavily on pre-trained neural networks and require additional post-processing to identify moving objects. Sensor fusion techniques, combining the precision of LiDAR with the semantic richness of RGB imagery, offer a promising alternative. In this work, we present a novel framework that enhances a quadruped robot equipped with a LiDAR sensor and an upward-facing fisheye camera for real-time dynamic object detection and tracking. After identifying moving objects within a registered point cloud, our method assigns semantic labels by projecting 3D coordinates onto a 2D cylindrical panorama, aligning with real-time image-based detections for observation update of the Kalman filter. The proposed system demonstrates high precision, simplicity, and robustness, particularly in handling objects transitioning between dynamic and static states, thus it is well-suited for deployment in real-world construction environments.

cs.RO↗

arXiv2026

ConSA: Controllable Sparsity in Hybrid Attention via Learnable Allocation

Hybrid architectures combining full attention (FA) and sliding-window attention (SWA) are a promising paradigm for efficient LLM inference. However, existing methods typically rely on hand-crafted rules or simple post-hoc heuristics for FA/SWA allocation and offer limited analysis of the attention behaviors underlying these designs. We propose Controllable Sparsity in Hybrid Attention (ConSA), a framework that learns optimal FA/SWA assignment under a user-specified sparsity target. ConSA employs L0 regularization to learn binary masks selecting between FA and SWA for each attention unit, while an augmented Lagrangian constraint enforces the target sparsity at either layer or KV-head granularity. We evaluate ConSA on two LLMs at the 0.6B and 1.7B scales. Learned allocations consistently outperform rule-based baselines, with KV-head-wise allocation yielding clear gains over layer-wise allocation. The learned patterns place SWA in the bottom layers and concentrate FA into contiguous middle-layer blocks, diverging from evenly interleaved patterns in rule-based methods. This structure persists across model scales, sparsity levels, and allocation granularities, revealing a fine-grained spectrum of intrinsic attention behaviors that underlies the learned allocation.

cs.CL↗

arXiv2026

WARA: Toward Automated Wireless Optimization Research with Closed-Loop LLM Agents

Large language model (LLM) agents are increasingly capable of tool use, code execution, artifact inspection, and iterative revision, creating new opportunities for automating scientific and engineering research. To the best of our knowledge, this paper presents the first end-to-end autoresearch framework for the wireless domain, with a focus on wireless resource allocation optimization. We propose the Wireless AutoResearch Agent (WARA), a closed-loop multi-agent system for automated wireless optimization research. Given only an initial topic, WARA decomposes the workflow into three phases: research gap identification and problem proposal, wireless optimization modeling, algorithm design and experimentation, and research deliverable construction. Across these phases, WARA uses artifact-mediated control: upstream artifacts are consumed as inputs, structured outputs are stored for downstream use, and controller-managed gates validate consistency among models, algorithms, experiments, and claims. When validation fails, WARA repairs only the responsible artifact instead of restarting the whole workflow. We present a representative wireless resource allocation case study showing how WARA converts an initial topic into a complete research package with executable evidence and a synthesized technical manuscript. We further design a structured LLM-based ScoringAgent to evaluate manuscript-level research validity and optimization research maturity. Comparative results show that WARA substantially outperforms one-shot LLM generation and approaches the quality profile of recently accepted peer-reviewed technical papers. These results indicate that closed-loop artifact control is a promising path toward end-to-end LLM-assisted wireless optimization research. The source code is available at https://github.com/guoyuan-dotcom/WARA_CUHKSZ.

cs.NI↗

arXiv2026

VIP-COP: Context Optimization for Tabular Foundation Models

Tabular foundation models (TFMs) have emerged as a powerful paradigm for in-context learning on structured data, enabling direct prediction on new tabular tasks without task-specific training. However, their effectiveness is constrained by context length limits, restricting application to medium-scale data and degrading performance when inference-time data exceed pretraining size distributions. Our work introduces VIP-COP, estimating the Value of Importance for Prediction of training examples and features for hard Context OPtimization for TFMs. Its explicit selection mechanism suppresses noise and isolates influential data, enabling the model to also benefit from data augmentation by prioritizing high-value augmented samples and features. VIP-COP is (i) fast, boosting performance often within minutes of optimization, based on an online KernelSHAP-based regression with iterative refinement, value-guided context sampling, and multi-fidelity pruning; (ii) budget-aware and any-time, improving with additional test-time compute unlike heuristics that produce fixed contexts; (iii) model-aware yet fully black-box, requiring no access to model internals, making it compatible with both proprietary and open-source TFMs; (iv) interpretable, identifying discrete ``Very Important Predictors'' (samples and features) that maximize signal-to-noise, which makes it (v) robust, isolating high-value data from noise. In contrast, soft-prompt optimization requires model gradients, produces abstract latent tokens, and lacks explicit signal discrimination. Extensive experiments show that VIP-COP consistently outperforms heuristic and optimized baselines across large-scale high-dimensional testbeds, including data augmentation and data-noise settings, establishing a new state of the art in test-time context refinement for TFMs.

cs.LG↗

arXiv2026

Beyond N-gram: Data-Aware X-GRAM Extraction for Efficient Embedding Parameter Scaling

Large token-indexed lookup tables provide a compute-decoupled scaling path, but their practical gains are often limited by poor parameter efficiency and rapid memory growth. We attribute these limitations to Zipfian under-training of the long tail, heterogeneous demand across layers, and "slot collapse" that produces redundant embeddings. To address this, we propose X-GRAM, a frequency-aware dynamic token-injection framework. X-GRAM employs hybrid hashing and alias mixing to compress the tail while preserving head capacity, and refines retrieved vectors via normalized SwiGLU ShortConv to extract diverse local n-gram features. These signals are integrated into attention value streams and inter-layer residuals using depth-aware gating, effectively aligning static memory with dynamic context. This design introduces a memory-centric scaling axis that decouples model capacity from FLOPs. Extensive evaluations at the 0.73B and 1.15B scales show that X-GRAM improves average accuracy by as much as 4.4 points over the vanilla backbone and 3.2 points over strong retrieval baselines, while using substantially smaller tables in the 50% configuration. Overall, by decoupling capacity from compute through efficient memory management, X-GRAM offers a scalable and practical paradigm for future memory-augmented architectures. Code aviliable in https://github.com/Longyichen/X-gram.

cs.CL↗

arXiv2026

Environment-Aware Near-Field Channel Estimation Leveraging CKM and ISAC

This paper proposes an environment-aware near-field channel estimation framework for integrated sensing and communication (ISAC) systems equipped with extremely large-scale antenna arrays (ELAAs). The proposed framework jointly exploits channel knowledge maps (CKMs) and ISAC to obtain a priori information on static and dynamic environmental features for facilitating channel estimation. In particular, we propose a novel CKM representation, termed the virtual object map (VOM), which stores the locations of virtual environment objects (EOs) to characterize the dominant multipath components (MPCs) induced by static physical EOs. In addition, we design a sensing-assisted channel training protocol, in which the ISAC-enabled base station (BS) transmits downlink pilots while simultaneously collecting monostatic echoes for sensing dynamic targets in the environment, and the user equipment (UE) feeds back a quantized version of its received pilot observation. Based on the VOM prior and the sensed dynamic information, the BS jointly estimates the coefficients of the static and dynamic MPCs to recover the near-field channel. Numerical results demonstrate that the proposed joint VOM- and sensing-aided channel estimation scheme significantly outperforms conventional schemes without VOM-based priors and/or dynamic sensing in terms of both channel estimation accuracy and achievable rate.

cs.IT↗

arXiv2026

Sparse Growing Transformer: Training-Time Sparse Depth Allocation via Progressive Attention Looping

Existing approaches to increasing the effective depth of Transformers predominantly rely on parameter reuse, extending computation through recursive execution. Under this paradigm, the network structure remains static along the training timeline, and additional computational depth is uniformly assigned to entire blocks at the parameter level. This rigidity across training time and parameter space leads to substantial computational redundancy during training. In contrast, we argue that depth allocation during training should not be a static preset, but rather a progressively growing structural process. Our systematic analysis reveals a deep-to-shallow maturation trajectory across layers, where high-entropy attention heads play a crucial role in semantic integration. Motivated by this observation, we introduce the Sparse Growing Transformer (SGT). SGT is a training-time sparse depth allocation framework that progressively extends recurrence from deeper to shallower layers via targeted attention looping on informative heads. This mechanism induces structural sparsity by selectively increasing depth only for a small subset of parameters as training evolves. Extensive experiments across multiple parameter scales demonstrate that SGT consistently outperforms training-time static block-level looping baselines under comparable settings, while reducing the additional training FLOPs overhead from approximately 16--20% to only 1--3% relative to a standard Transformer backbone.

cs.CL↗

arXiv2026

Mixture of Universal Experts: Scaling Virtual Width via Depth-Width Transformation

Mixture-of-Experts (MoE) decouples model capacity from per-token computation, yet their scalability remains limited by the physical dimensions of depth and width. To overcome this, we propose Mixture of Universal Experts (MOUE),a MoE generalization introducing a novel scaling dimension: Virtual Width. In general, MoUE aims to reuse a universal layer-agnostic expert pool across layers, converting depth into virtual width under a fixed per-token activation budget. However, two challenges remain: a routing path explosion from recursive expert reuse, and a mismatch between the exposure induced by reuse and the conventional load-balancing objectives. We address these with three core components: a Staggered Rotational Topology for structured expert sharing, a Universal Expert Load Balance for depth-aware exposure correction, and a Universal Router with lightweight trajectory state for coherent multi-step routing. Empirically, MoUE consistently outperforms matched MoE baselines by up to 1.3% across scaling regimes, enables progressive conversion of existing MoE checkpoints with up to 4.2% gains, and reveals a new scaling dimension for MoE architectures.

cs.LG↗

arXiv2026

MacrOData: New Benchmarks of Thousands of Datasets for Tabular Outlier Detection

Quality benchmarks are essential for fairly and accurately tracking scientific progress and enabling practitioners to make informed methodological choices. Outlier detection (OD) on tabular data underpins numerous real-world applications, yet existing OD benchmarks remain limited. The prominent OD benchmark AdBench is the de facto standard in the literature, yet comprises only 57 datasets. In addition to other shortcomings discussed in this work, its small scale severely restricts diversity and statistical power. We introduce MacrOData, a large-scale benchmark suite for tabular OD comprising three carefully curated components: OddBench, with 790 datasets containing real-world semantic anomalies; OvrBench, with 856 datasets featuring real-world statistical outliers; and SynBench, with 800 synthetically generated datasets spanning diverse data priors and outlier archetypes. Owing to its scale and diversity, MacrOData enables comprehensive and statistically robust evaluation of tabular OD methods. Our benchmarks further satisfy several key desiderata: We provide standardized train/test splits for all datasets, public/private benchmark partitions with held-out test labels for the latter reserved toward an online leaderboard, and annotate our datasets with semantic metadata. We conduct extensive experiments across all benchmarks, evaluating a broad range of OD methods comprising classical, deep, and foundation models, over diverse hyperparameter configurations. We report detailed empirical findings, practical guidelines, as well as individual performances as references for future research. All benchmarks containing 2,446 datasets combined are open-sourced, along with a publicly accessible leaderboard hosted at https://huggingface.co/MacrOData-CMU.

cs.LG↗

arXiv2026

Near-Field Multi-Cell ISCAP with Extremely Large-Scale Antenna Array

This paper investigates a coordinated multi-cell integrated sensing, communication, and powering (ISCAP) system operating in the electromagnetic near field, where each base station (BS) employs an extremely large-scale antenna array (ELAA) to simultaneously support downlink communication, wireless power transfer (WPT), and environmental sensing. Three categories of communication users (CUs) with different interference cancellation capabilities are considered, and sensing is enabled through a distributed multiple-input multiple-output (MIMO) radar architecture. To address the resulting design challenges, a robust optimization framework is proposed by optimizing the beamforming strategy to maximize the worst-case detection probability over a prescribed sensing region, subject to per-user signal-to-interference-plus-noise ratio (SINR) constraints and energy harvesting requirements at energy receivers (ERs), while explicitly capturing the uncertainty in ER locations. By leveraging semidefinite relaxation (SDR), the original non-convex problem is reformulated as a convex semidefinite program with a provably tight relaxation. Furthermore, a low-complexity maximum ratio transmission (MRT)-based suboptimal scheme is developed, yielding a closed-form solution in the asymptotic regime as the number of antenna elements approaches infinity. Extensive numerical results reveal the fundamental trade-offs among sensing accuracy, communication reliability, and WPT efficiency.

cs.NI↗

arXiv2025

Blink: Dynamic Visual Token Resolution for Enhanced Multimodal Understanding

Multimodal large language models (MLLMs) have achieved remarkable progress on various vision-language tasks, yet their visual perception remains limited. Humans, in comparison, perceive complex scenes efficiently by dynamically scanning and focusing on salient regions in a sequential "blink-like" process. Motivated by this strategy, we first investigate whether MLLMs exhibit similar behavior. Our pilot analysis reveals that MLLMs naturally attend to different visual regions across layers and that selectively allocating more computation to salient tokens can enhance visual perception. Building on this insight, we propose Blink, a dynamic visual token resolution framework that emulates the human-inspired process within a single forward pass. Specifically, Blink includes two modules: saliency-guided scanning and dynamic token resolution. It first estimates the saliency of visual tokens in each layer based on the attention map, and extends important tokens through a plug-and-play token super-resolution (TokenSR) module. In the next layer, it drops the extended tokens when they lose focus. This dynamic mechanism balances broad exploration and fine-grained focus, thereby enhancing visual perception adaptively and efficiently. Extensive experiments validate Blink, demonstrating its effectiveness in enhancing visual perception and multimodal understanding.

cs.CV↗

arXiv2025

Channel Knowledge Map Enabled Low-Altitude ISAC Networks: Joint Air Corridor Planning and Base Station Deployment

This letter addresses the joint air corridor planning and base station (BS) deployment problem for low-altitude integrated sensing and communication (ISAC) networks. In the considered system, unmanned aerial vehicles (UAVs) operate within a structured air corridor composed of connected cubic segments, and multiple BSs need to be selectively deployed at a set of candidate locations to ensure both sensing and communication coverage throughout the corridor. In particular, we leverage the channel knowledge map (CKM) to characterize wireless channels for candidate BS sites prior to deployment, thereby facilitating the offline planning. Under this setup, we minimize the system cost in terms of the weighted sum of the air corridor length and the number of deployed BSs, subject to the constraints on both sensing and communication performance across the corridor. To solve the formulated large-scale nonconvex integer programming problem, we develop a hierarchical coarse-to-fine grid decomposition algorithm. Simulation results demonstrate the benefit of the proposed joint design in reducing the overall deployment cost while ensuring the coverage of the low-altitude ISAC networks.

eess.SP↗

arXiv2025

Intelligent Reflecting Surfaces for Integrated Sensing and Communications: From System Coexistence to Networked Mutualism

The rapid development of sixth-generation (6G) wireless networks requires seamless integration of communication and sensing to support ubiquitous intelligence and real-time, high-reliability applications. Integrated sensing and communication (ISAC) has emerged as a key solution for achieving this convergence, offering joint utilization of spectral, hardware, and computing resources. However, realizing high-performance ISAC remains challenging due to environmental line-of-sight (LoS) blockage, limited spatial resolution, and the inherent coverage asymmetry and resource coupling between sensing and communication. Intelligent reflecting surfaces (IRSs), featuring low-cost, energy-efficient, and programmable electromagnetic reconfiguration, provide a promising solution to overcome these limitations. This article presents a comprehensive overview of IRS-aided wireless sensing and ISAC technologies, including IRS architectures, target detection and estimation techniques, beamforming designs, and performance metrics. It further explores IRS-enabled new opportunities for more efficient performance balancing, coexistence, and networking in ISAC systems, focuses on current design bottlenecks, and outlines future research directions. This article aims to offer a unified design framework that guides the development of practical and scalable IRS-aided ISAC systems for the next-generation wireless network.

eess.SP↗