arXiv subjects

Jie Ren

Publications and source records attributed to Jie Ren.

At least 19 recordsLinked to original sources

arXiv2026

Efficient GUI Agents: A Systems Survey of Observation, Memory, Action, and Runtime Optimization

GUI agents increasingly operate across websites, mobile apps, and desktop environments, yet the field still reports progress primarily through task success. We argue that practical deployment depends equally on efficiency: how much context, computation, action budget, and runtime overhead an agent consumes while succeeding. This survey studies efficient GUI agents through an end-to-end systems lens that preserves the current technical axes of observation efficiency, context and memory efficiency, action efficiency, and planner-side/system efficiency. For each subsection, we expand the seed literature through targeted search plus backward and forward citation chaining, then synthesize the dominant mechanisms, reported efficiency signals, and new overheads they introduce. Across the literature, recent progress converges on a small set of recurring ideas: selective reading instead of full-context ingestion, global-to-local visual allocation, recoverable memory rather than raw history replay, verification-aware control, and hybrid runtimes that can switch between GUI and non-GUI execution. We conclude by identifying the main open problems, including honest accounting of verifier cost, cross-benchmark comparability, and co-design of observation, memory, and execution layers under real latency and privacy constraints.

cs.CL↗

arXiv2026

From Inaudible Inputs to Model Failures: Low-Frequency Safety Risks in LALMs

Large audio-language models (LALMs) have demonstrated strong capabilities in understanding diverse audio inputs. This diversity includes low-frequency signals that are inaudible to humans but can still enter the model and influence its generation. However, the practical impact of such low-frequency inputs on LALMs remains largely unexplored. In this paper, we propose Intermittent Low-Frequency Lockout (ILL), an inaudible red teaming method that evaluates this risk using a universal waveform template in a black box setting. ILL uses Sentence Attention Scale Estimation to determine active intervals and Frequency Confusion Transfer to construct a low-frequency waveform with continuous phase from corpus spectral variation. To mitigate this risk, we propose Distributional Requery Guard (DRG) to detect low-frequency distribution shifts and conditionally request a second recording for semantic recovery. Across six LALMs and multiple audio understanding tasks, ILL reduces accuracy by up to 67 percentage points while receiving a mean human audibility rating of 1.33, close to 1.17 for clean audio; DRG raises mean attacked accuracy from 28.5\% to 46.1\% after clean reacquisition. These findings identify a previously overlooked safety risk for LALMs and provide a foundation for future research on robust audio understanding.

cs.SD↗

arXiv2026

C2Dex: Contact-Consistent Reconstruction and Retargeting for Dexterous Manipulation from Monocular Video

High-quality demonstrations for dexterous robot manipulation are costly and difficult to collect, whereas monocular human videos provide a scalable source of diverse manipulation behaviors. However, transferring such demonstrations to dexterous robots remains challenging: monocular hand-object interaction (HOI) reconstruction often produces temporally unstable contacts and physically implausible interactions, while conventional retargeting methods struggle to preserve task-relevant contacts and local interaction geometry across different hand embodiments. We present C2Dex, a video-to-dexterous-manipulation framework built around a shared interaction representation: stable object-side contacts recovered by aggregating noisy frame-wise observations in the canonical object space. These stable contacts serve a dual role: as trajectory-level constraints that guide reconstruction toward temporally coherent and physically plausible human HOI trajectories, and as explicit transfer targets for the dexterous hand, where Laplacian interaction optimization preserves the local hand-object geometry across embodiments and residual reinforcement learning refines the trajectory in simulation. Experiments on DexYCB and TACO show that C2Dex achieves end-to-end trajectory success rates of 57.78% and 26.67%, respectively, substantially outperforming the strongest baselines (17.78% and 10.00%) under identical evaluation criteria. Real-robot replay experiments further demonstrate physical feasibility across diverse contact-rich manipulation tasks. Project page: https://k-jie.github.io/C2Dex/

cs.RO↗

arXiv2026

Geometric Mode Steering of the Quantum Mpemba Effect

The slowest Liouvillian mode often bottlenecks the relaxation of an open quantum system to its steady state. Standard strategies circumvent this bottleneck by selecting special initial states or engineering the dissipator. Here we show that neither is necessary. We introduce a pre-dissipative geometric steering protocol that reshapes any given pure or mixed state before relaxation begins -- coherent rotations interleaved with nonselective projective measurements -- at fixed Lindblad generator. By steering the state's Bloch direction along geodesic paths, the protocol suppresses its overlap with the slowest Liouvillian modes. The prepared state then starts farther from equilibrium yet relaxes faster, realizing the quantum Mpemba effect, whenever two computable conditions hold: reduced slow-mode overlap and a larger initial distance to stationarity. Our framework treats real and complex spectral gaps uniformly, and we demonstrate robust Mpemba acceleration in driven qubit and multiqubit systems using operations available in trapped-ion and superconducting platforms.

quant-ph↗

arXiv2026

Quantum many-body mixed phase space revealed by hybrid feedback control

Understanding how complex systems transition between order and chaos is a central challenge of nonequilibrium physics. While weak perturbations of classical integrable systems give rise to a mixed phase space of coexisting regular and chaotic trajectories, analogous behavior in interacting quantum many-body systems has remained elusive. Here we develop and experimentally implement a hybrid quantum-classical feedback protocol that autonomously discovers and stabilizes long-lived regular trajectories in a superconducting quantum processor. Each iteration combines short-time quantum evolution with classical optimization that projects the dynamics back onto a low-entanglement variational manifold, effectively distilling coherence from chaotic evolution. The stabilized trajectories reveal a quantum many-body mixed phase space emerging from nonlinear variational dynamics, without a direct analogue in classical or few-body quantum systems. Our results establish a versatile framework for algorithmic discovery and control of coherent dynamics previously inaccessible to experiment.

quant-ph↗

arXiv2026

Thermalization in a Height-Conserving Quantum Dimer Model

Strongly constrained quantum systems, in which local rules forbid most configurations, play a central role in condensed matter and lattice gauge theory. Their thermalization is often thought to be delicate: extensive conservation laws and dynamically frozen states can shatter the Hilbert space into many disconnected sectors. A natural question is whether, once the frozen states are removed, the dynamics within a single sector still thermalizes. We address this in the height-conserving quantum dimer model on the square lattice, whose local plaquette flips conserve an emergent height field. Resolving the winding numbers, the four sublattice heights, and lattice momentum , we isolate the dominant connected Krylov component of each fragmented sector and analyze its spectral spectral statistics, entanglement, and connectivity. The two standard chaos diagnostics then show different behavior:across momentum sectors the level-spacing statistics range from near-Poisoon to Wigner-Dyson, yet in every sector the eigenstate entanglement entropy collapses onto a narrow, dome-shaped curve characteristic of eigenstate thermalization. Only a handful of low-entanglement outliers interrupt this thermal pattern, in selected sectors. Thus, strong kinematic constraints can lead to a situation where spectral correlations and eigenstate thermalization need not follow the same universal signatures -- a manifestation of constrained quantum chaos.

cond-mat.str-el↗

arXiv2026

TS-Mask VLA: 2D Temporal-Spatial Masking for Vision-Language-Action Model with Effective Bridging

Vision-language-action (VLA) models aim to understand natural-language instructions and visual observations, and to generate and execute corresponding actions as embodied agents. Recently, autoregressive token-based action generation has driven the development of many representative VLA models. However, this paradigm often reduces action generation to next-token prediction, thereby lacking explicit modeling of the spatiotemporal structure of action sequences and the disentanglement between vision-language representations and actions, which can limit performance in long-horizon and complex scenarios. In this paper, we propose TS-Mask VLA, a vision-language-action framework for robot manipulation. TS-Mask VLA is built upon two key designs: (1) a Discrete Diffusion Action Expert equipped with a Bridge Attention conditioning bridge, which enables multi-layer conditioning from the VLM and facilitates more accurate and stable action generation; and (2) a temporal-spatial 2D masking strategy for discrete action tokens that strengthens the model's understanding of cross-time dependencies and inter-dimensional coupling, leading to more structurally consistent action sequences. We conduct extensive experiments on simulation benchmarks and real-world tasks. On LIBERO, TS-Mask VLA achieves a 95.7 percent average success rate with only 0.5B parameters, outperforming significantly larger models. On CALVIN, it attains the best average sequence length of 4.19 and strong long-horizon performance. Comprehensive analyses and ablations further validate the effectiveness of our design.

cs.RO↗

arXiv2026

Optically Switched Phonon Superradiance of Surface Acoustic Wave in Diamond

Surface acoustic wave (SAW) phonon coupling with nitrogen-vacancy (NV) center spins in diamond offers a promising platform for on-chip quantum phononic manipulations. Although an ensemble of NV centers coupled to a common SAW phonon mode enables superradiance and collective quantum control, achieving a tunable superradiant phase transition remains challenging. Here, we show that optically driving NV centers level transitions enhances the effective spin-phonon coupling, triggering a SAW phonon superradiant phase transition in the weak-coupling regime. We also demonstrate that above a critical threshold, the driving light rapidly switches on the phonon superradiance--a dynamic effect that persists in finite-number NV ensembles. Our results provide a controllable route to coherent phonon-NV spin manipulation in solid state quantum devices.

cond-mat.mes-hall↗

arXiv2026

Axial-Vector Lattice Benchmarks Reveal a Common Medium Response of Meson Screening in Hot QCD

Meson screening masses trace the evolution of hadronic correlations toward quasi-free quark--antiquark screening in hot QCD. Combining lattice QCD (LQCD) benchmarks with a symmetry-preserving Dyson--Schwinger baseline, we identify a flavor-dependent axial-vector quasi-free onset, $x^\ast_{fg}=T^\ast_{fg}/T_c$: an operational high-temperature matching scale at which the axial-vector screening mass has approached the corresponding free-field value after ordinary chiral restoration or parity-partner convergence has set in. On the finite interval $1\le x\lesssim x^\ast_{fg}$, independent light/strange and charm-containing lattice benchmarks are organized by a common medium-response function with one flavor-sector parameter. One axial-vector point fixes this parameter; the remaining axial-vector data test its temperature dependence, and vector screening masses validate it without vector input. A reduced-mass interpolation then yields lattice-testable quasi-free onsets and screening spectra for light-charm and bottom-containing sectors. The resulting onset scales provide common reference points for future lattice and continuum studies of meson dissolution across flavor.

hep-ph↗

arXiv2026

ROSE: Benchmarking the Perception-to-Action Gap in Multimodal Models

Multimodal large language models (MLLMs) are increasingly expected to act on visual information, yet the same scene may require different actions under different task contexts. How reliably can a model turn the same visual evidence into the action required by the current context? To answer this question, we introduce \textsc{ROSE} (\textbf{R}eference-conditioned \textbf{O}ddity and \textbf{S}ymbolic \textbf{E}xecution), a controlled benchmark that holds the visual scene fixed while varying region constraints and required symbolic outputs. Through coupled counting and coordinate-action tasks, \textsc{ROSE} tests whether models can infer an implicit majority reference and act on the resulting fine-grained visual evidence under changing contexts. Across nine recent MLLMs, performance drops by as much as 44.5 percentage points from counting-oriented tasks to region-conditioned action, despite 98.8\% human performance. The gap persists on paired scenes and regions for which the same model returns the correct count, while global-click and matched local controls show that coordinate grounding explains only part of the loss, revealing a distinct, model-dependent bottleneck in turning shared visual evidence into context-specific actions.

cs.CV↗

arXiv2026

AutoPass: Evidence-Guided LLM Agents for Compiler Performance Tuning

Large Language Models (LLMs) show promise for code compilation tasks, but applying them to runtime performance tuning is difficult due to complex microarchitectural effects and noisy runtime measurements. We present AutoPass, a multi-agent framework for compiler performance tuning that uses compiler and runtime evidence to guide LLM-generated optimization decisions. Rather than treating the compiler as a black box like prior auto-tuning schemes, AutoPass opens up the compiler to the LLM, enabling it to query compiler-internal optimization states and analyze the intermediate representation to orchestrate compiler options. The search process iteratively refines optimization configurations using measured runtime feedback to diagnose regressions and guide latency-improving edits. AutoPass operates in an inference-only, training-free setting and requires no offline training or task-specific fine-tuning, making it readily applicable to new benchmarks and platforms. We implement AutoPass on the LLVM compiler and evaluate it on server-grade x86-64 and embedded ARM64 systems. AutoPass outperforms expert-tuned heuristics and classical autotuning methods, achieving geometric-mean speedups of 1.043x and 1.117x over LLVM -O3 on x86-64 and ARM64, respectively.

cs.SE↗

arXiv2026

Moir\'e Strain Skyrmions in Sliding Twisted Bilayers

Strain defect is crucial to the physical properties of solid materials. Among them, strain glass induced by defect engineering provides an important paradigm for nanoscale domain manipulation. Here, we propose purely mechanical moir\'e strain Skyrmions, a topologically protected elastic textures whose motion can be controlled by interlayer sliding and the chirality of the moir\'e bilayer. Using an empirical continuum elastic model combined with symmetry analysis, we demonstrate the Skyrmion lattice structure as the elastic ground state. Under interlayer sliding, these moir\'e strain Skyrmions exhibit the Skyrmion Hall effect of transverse motion, with a Hall angle determined by bilayer chirality and inversely proportional to the moir\'e twist angle. Our work establishes interlayer sliding as an efficient, low-energy control knob for topological excitations, offering a new paradigm for designing chiral-material-based information transport devices.

cond-mat.mtrl-sci↗

arXiv2026

Why Retrieval-Augmented Generation Fails: A Graph Perspective

Retrieval-Augmented Generation (RAG) has become a powerful and widely used approach for improving large language models by grounding generation in retrieved evidence. However, RAG systems still produce incorrect answers in many cases. Why RAG fails despite having access to external information remains poorly understood. We present a model-internal study of retrieval-augmented generation that examines how retrieved evidence influences answer generation. Using circuit tracing, we construct attribution graphs that model the flow of information through transformer layers during decoding. These graphs represent interactions among retrieved context, intermediate model activations, and generated tokens, providing a graph, circuit-level view of how external evidence is integrated into the model's reasoning process across multiple question answering benchmarks, we observe consistent structural differences: correct predictions exhibit deeper reasoning paths, more distributed evidence flow, and a more structured pattern of local connectivity, while failed predictions show shallower, fragmented, and overly concentrated evidence flow. Building on these findings, we develop a graph-based error detection framework that uses attribution-graph topology features. Furthermore, we show that attribution graphs enable targeted interventions. By reinforcing question-constrained evidence grounding, we reshape internal routing so that answer generation remains guided by the question, leading to more effective integration of retrieved information and fewer errors.

cs.CL↗

arXiv2026

Distributed Generative Inference of LLM at Internet Scales with Multi-Dimensional Communication Optimization

Decentralized LLM inference distributes computation among heterogeneous nodes across the internet, offering a performant and cost-efficient solution, alternative to traditional centralized inference. However, the low cross-node network bandwidth makes communication the primary bottleneck. In this paper, we introduce BloomBee, an internet-scale distributed LLM inference framework. BloomBee integrates LLM-layer assignment, micro-batching and tensor offloading to optimize communication from multiple dimensions. Additionally, BloomBee formulates the coordination of these techniques as an optimization problem and solves it using dynamic programming. BloomBee also customizes lossless compression and speculative decoding according to low-bandwidth network settings to reduce communication overhead. We evaluate BloomBee across a spectrum of network environments and show that it improves service throughput by up to 1.76x. It also reduces average latency by up to 43.20% compared to state-of-the-art decentralized LLM inference systems. BloomBee is open-sourced.

cs.DC↗

arXiv2026

Beyond Pixels: Introspective and Interactive Grounding for Visualization Agents

Vision-Language Models (VLMs) frequently misread values, hallucinate details, and confuse overlapping elements in charts. Current approaches rely solely on pixel interpretation, creating a Pixel-Only Bottleneck: agents treat interactive charts as static images, losing access to the structured specification that encodes exact values. We introduce Introspective and Interactive Visual Grounding (IVG), a framework that combines (1) spec-grounded introspection, which queries the underlying specification for deterministic evidence, with (2) view-grounded interaction, which manipulates the view to resolve visual ambiguity. To enable evaluation without VLM bias, we present iPlotBench, a benchmark of 500 interactive Plotly figures with 6,706 binary questions and ground-truth specifications. Experiments show that introspection improves data reconstruction fidelity, while the combination with interaction achieves the highest QA accuracy (0.81), with +6.7 % gains on overlapping geometries. We further demonstrate IVG in deployed agents that explore data autonomously and collaborate with human users in real time.

cs.CL↗

arXiv2026

Taming GPU Underutilization via Static Partitioning and Fine-grained CPU Offloading

Advances in GPU compute throughput and memory capacity brings significant opportunities to a wide range of workloads. However, efficiently utilizing these resources remains challenging, particularly because diverse application characteristics may result in imbalanced utilization. Multi-Instance GPU (MIG) is a promising approach to improve utilization by partitioning GPU compute and memory resources into fixed-size slices with isolation. Yet, its effectiveness and limitations in supporting HPC workloads remain an open question. We present a comprehensive system-level characterization of different GPU sharing options using real-world scientific, AI, and data analytics applications, including NekRS, LAMMPS, Llama3, and Qiskit. Our analysis reveals that while GPU sharing via MIG can significantly reduce resource underutilization, and enable system-level improvements in throughput and energy, interference still occurs through shared resources, such as power throttling. Our performance-resource scaling results indicate that coarse-grained provisioning for tightly coupled compute and memory resources often mismatches application needs. To address this mismatch, we propose a memory-offloading scheme that leverages the cache-coherent Nvlink-C2C interconnect to bridge the gap between coarse-grained resource slices and reduce resource underutilization.

cs.DC↗

arXiv2026

ChunQiuTR: Time-Keyed Temporal Retrieval in Classical Chinese Annals

Retrieval shapes how language models access and ground knowledge in retrieval-augmented generation (RAG). In historical research, the target is often not an arbitrary relevant passage, but the exact record for a specific regnal month, where temporal consistency matters as much as topical relevance. This is especially challenging for Classical Chinese annals, where time is expressed through terse, implicit, non-Gregorian reign phrases that must be interpreted from surrounding context, so semantically plausible evidence can still be temporally invalid. We introduce \textbf{ChunQiuTR}, a time-keyed retrieval benchmark built from the \textit{Spring and Autumn Annals} and its exegetical tradition. ChunQiuTR organizes records by month-level reign keys and includes chrono-near confounders that mirror realistic retrieval failures. We further propose \textbf{CTD} (Calendrical Temporal Dual-encoder), a time-aware dual-encoder that combines Fourier-based absolute calendrical context with relative offset biasing. Experiments show consistent gains over strong semantic dual-encoder baselines under time-keyed evaluation, supporting retrieval-time temporal consistency as a key prerequisite for faithful downstream historical RAG. Our code and datasets are available at \href{https://github.com/xbdxwyh/ChunQiuTR}{\texttt{github.com/xbdxwyh/ChunQiuTR}}.

cs.CL↗

arXiv2026

KappaFormer: Physics-aware Transformer for lattice thermal conductivity via cross-domain transfer learning

Machine learning has been widely used for predicting material properties. However, efficient prediction of lattice thermal conductivity ($\kappa_\mathrm{L}$) remains a long-standing challenge, primarily due to the scarcity of high-quality training data. Here we introduce KappaFormer, a physics-aware Transformer architecture that embeds the harmonic-anharmonic decomposition of $\kappa_\mathrm{L}$ within the network. KappaFormer comprises a harmonic branch pre-trained on large-scale elastic property data and an anharmonic branch fine-tuned on limited experimental $\kappa_\mathrm{L}$ data, enabling effective knowledge transfer and enhanced generalization. High-throughput screening with KappaFormer identifies multiple candidates with ultralow $\kappa_\mathrm{L}$, which are further confirmed by first-principles calculations. Physics interpretability further elucidates the vibrational mechanisms governing thermal transport suppression, linking structural motifs to strong anharmonicity. This study provides a generalizable framework for physics-guided machine learning to accelerate the discovery of new materials.

cond-mat.mtrl-sci↗