Searcharxiv⌕ Search

arXiv subjects

Yiming Zhu

Publications and source records attributed to Yiming Zhu.

At least 19 recordsLinked to original sources
arXiv2026

Trim: Fee Level Matters More Than Skew in Lightning Routing

Lightning routing nodes lend their channel balance for a posted fee. Operators' tools move that fee with the balance around a hand-set level, and which part of the fee drives revenue is unclear. We present Trim, which models a channel as a two-sided inventory with costly refills. Its optimal fee is a level, set by demand and replenishment, plus a balance skew, the shadow price of liquidity. The level is first order: no fee rule beats the ceiling demand and replenishment set, and on exactly solved channels the best flat fee comes within $1.6\%$ of the optimum at the median, whereas a level off by a factor of two loses tens of percent. The skew is second-order insurance: it adds at most what the flow-balancing flat fee loses to refused forwards. So the level should be chosen: Trim's rule traces the revenue--success frontier, posts one level on every channel, and adds a skew only where refills stay frequent and costly. On a 2026 snapshot with synthetic payments, no tuned tool beats that frontier by over a few percent, and a skew changes gross revenue by a few percent.

cs.NI↗
arXiv2026

Axiom-0: Hierarchical Self-Refinement for Vision-Language-Action Policies

Reliable manipulation requires policies that generalize across tasks, layouts, and object interactions. Vision-language-action (VLA) models combine pretrained multimodal understanding with learned control, but their performance remains limited on unfamiliar task compositions. We present Axiom-0, a VLA policy that addresses both the quality of its representations and their use in action generation through hierarchical self-refinement. Intra-VLM Representation Refinement (IVRR) strengthens selected intermediate features through consistency supervision with learned feature projections during training. VLM-Action Interface Refinement (VAIR) recursively updates the resulting conditioning before action generation. Joint optimization of the supervised VLM layers, VAIR, and the action generator connects the two components. On RoboCasa365, Axiom-0 achieves 62.32% overall success in our evaluation, exceeding the highest published leaderboard score (Paimon-0, 58.1%) by 4.22 percentage points and the Xiaomi-Robotics-1 reference by 4.92 points. Under the evaluated profiling setup, mean action-chunk generation latency increases by 1.7% over Xiaomi-Robotics-1. These results demonstrate improved task completion with modest inference overhead.

cs.RO↗
arXiv2026

Beyond Compression: Diagnosing How Post-Training Changes Mathematical Reasoning

Post-training is central to mathematical reasoning in modern large language models (LLMs), but endpoint pass@1 alone underidentifies what has changed. Gains may reflect newly reachable solutions, cheaper sampling of latent solutions, surface robustness, or memorisation. We compare three post-training paths under a common diagnostic readout: our sufficiently trained off-policy distillation trajectories, released Qwen3 off-policy-plus-on-policy distillation endpoints, and a released DeepSeek-Math endpoint trained with Group Relative Policy Optimisation (GRPO). Our probe uses cross-surface pass@K over verbatim prompts, paraphrases, numerical isomorphisms, and translations, plus consistency, distribution-shape, and verified supervised-fine-tuning (SFT) membership analyses. We find two regimes. On easier AMC problems, large-K ceilings are near saturation, so post-training mainly compresses sample cost. On harder AIME problems, post-training expands the large-K ceiling over the base model: sufficient off-policy distillation already raises this ceiling, Qwen3 released endpoints raise it further, and DeepSeek-Math GRPO does not dominate sufficient off-policy distillation at large K. English-dominant distillation improves non-English reasoning but preserves language-tier gaps. A controlled-overfit audit finds limited sensitivity in current SFT-membership probes. Compression is one regime of post-training, not a universal explanation.

cs.LG↗
arXiv2026

Solving Without Stopping: On-Policy Distillation at Small Scale

On-policy distillation, where a student learns from a stronger teacher's feedback on its own outputs, is a common way to pass reasoning to smaller models. We analyze what it transfers at small scale, distilling Qwen3-8B into Qwen3 4B, 1.7B and 0.6B students, in thinking mode (reason at length, then end the reasoning and answer) and, for comparison, in non-thinking mode (no separate reasoning phase). Long reasoning needs two abilities, solving a problem and knowing when it is solved, and we find that distillation transfers the first, but in thinking mode not the second. Solving improves at every size, up to two ceilings, which we measure comprehensively across both modes and all student sizes: a student's single attempt never exceeds what it could already reach in many attempts before training, and the smaller the student, the further it stays below the teacher. Stopping is where the modes part. In non-thinking mode every student keeps stopping; in thinking mode students stop ending their reasoning early in training, and the smaller the student, the less of this ability survives: the teacher signals a stop almost only where a student already ends its reasoning, so distillation teaches no new stops; it only keeps the student's existing stops that land on a right answer, and a weak student has few such stops. The smallest students often reach the right value but do not commit to it: they either rarely mark it or mark it and write past it. Together, these results describe how small students behave under on-policy distillation, and a diagnostic that separates answer marking, correctness and stopping.

cs.AI↗
arXiv2026

Sluice: Global Invariant, Local Enforcement for Pooled Payment-Channel Liquidity

A routing node on the Lightning Network holds its liquidity in separate channels, so a payment can fail at a channel whose outbound balance is exhausted while the node's other channels still hold balance. Pooling the channels into one reserve fixes this only if the node's draws across all channels stay within the reserve: a global invariant that each counterparty must enforce from its own channel, with no shared counter that off-chain draws can update. A node must therefore split the reserve into per-channel quotas in advance or coordinate every draw with every counterparty. On three Lightning snapshots the advance split forfeits $16$ to $67\%$ of the pooling gain over unpooled channels, and the loss grows with channel count. Sluice recovers $19$ to $65\%$ of that loss with a nested reservation: each channel keeps an exclusive base that its counterparty checks alone, and the rest is a shared overflow drawn on with certificates from a capacity-weighted quorum of the node's counterparties. Two conflicting certificates share an honest signer, so over-drawing is prevented without a slashable stake, under a stated bound on the capacity the node controls in the signing set. Sluice loses at most $1.3$ points of payment success to coordination where deployed coin movers lose up to $10.3$, and improves them in eleven of twelve cells when stacked on them. Re-creating every base output each epoch costs $2.8$ to $5.1$ times Lightning's on-chain bytes; re-creating only those that overflowed costs $0.5$ to $1.2$ times and keeps part of the gain.

cs.CR↗
arXiv2026

Population Metrology of a Hidden Exciton Reservoir: Quasi-Thermalization versus Localization

Long-lived dark states can dominate the lowest-energy manifold of optically driven quantum materials, yet their occupation remains difficult to quantify, leaving it unclear whether it reflects thermal redistribution or kinetic trapping. We combine microsphere-enabled far-field access with quantitative optical-response calibration to retrieve dark-to-bright population ratios in monolayer WSe2. Temperature-dependent measurements and controlled defect enhancement separate mobile and localized contributions. Near room temperature, the mobile dark-to-bright ratio reaches approximately 65% of its Boltzmann limit, indicating substantial but incomplete quasi-thermalization, whereas the low-temperature excess is dominated by defect-assisted localization. Dark-state dominance alone therefore does not establish equilibration, a distinction essential for interpreting transport and collective phases in optically hidden quasiparticle reservoirs.

physics.optics↗
arXiv2026

Optical-NIR Multi-band Photometric Analysis and Characterization of Giant Exoplanets with CPI-C

We present a multi-band photometric approach to characterize giant exoplanets, which represents one of the anticipated core scientific outcomes of Cool Planet Imaging Coronagraph (CPI-C). CPI-C operates with two observational channels covering visible and near-infrared wavelengths, each equipped with four broadband filters. The planet--star flux ratio integrated over each filter bandpass is calculated for photometric analysis. For cool planets observed in the visible bands, the data are primarily used to fit the overall spectral shape and methane-induced modulation, providing sensitivity to metallicity- and cloud-dependent spectral variations while constraining the reflected-light spectral shape and the combined scaling involving planet radius, orbital separation, and orbital phase. In the near-infrared bands, which probe thermal emission, the data help to better constrain fundamental planetary parameters including the effective temperature, radius, surface gravity and mass. For a synthetic giant planet with measurable reflected-light and thermal-emission components, the combined VIS4+NIR4 data provide tighter same-target constraints than either filter set alone, especially for the planet radius and cloud sedimentation parameter. Our simulations incorporate realistic instrument throughput, detector noise, and residual speckle noise. The results demonstrate that the eight-band design spanning visible to near-infrared wavelengths supports reflected-light diagnostics, thermal-emission characterization, and joint optical--NIR analysis of giant exoplanets within CPI-C science observations.

astro-ph.EP↗
arXiv2026

Revealing the high redshift host galaxy of the short GRB 061201 with JWST

Using deep near-infrared and optical images from JWST and HST, we identify a new host galaxy candidate for GRB 061201. It lies ~2" from the optical afterglow position. Photometric redshift fitting yields z~1.2. We compare the previously proposed host at z=0.111 with the new candidate. The chance-coincidence probability is $P_{cc}=0.18$, above the classical threshold of 0.1 but consistent with a physical association given the extreme depth of JWST imaging. In contrast, evaluated with corresponding JWST observations, the previously claimed host has a lower $P_{cc}=0.11$, which is driven primarily by bright-tail statistics rather than a more plausible association. A high-z origin is favored by three independent lines of evidence. First, for the z=0.111 scenario, the beaming-corrected energy shows GRB 061201 is an outlier of the Ghirlanda ($E_{p,i}-E_γ$) relation for short GRBs, while for the z=1.2 scenario, it is well consistent with the Amati relation. Second, deep near-infrared observations rule out a kilonova similar to AT2017gfo at z=0.111. Third, afterglow modeling yields an AIC criterion of $Δ$AIC=16.35, providing strong evidence for the high-redshift scenario. Assuming the host candidate is the actual host galaxy of GRB 061201, the physical offset is 16.4-16.9 kpc (substantially reduced from ~42 kpc) and the host stellar age is ~2 Gyr, which are consistent with the host population of short GRBs. A low-redshift origin would lead to a very high binary neutron star merger rate of ~1400 Gpc$^{-3}$ yr$^{-1}$, which is contradictory to the gravitational-wave constraint. We suggest that GRB 061201 originates from a moderately high-redshift (z~1.2) host, significantly alleviating this apparent merger rate discrepancy. This case demonstrates the power of deep JWST exposures in revealing the host galaxies of historically hostless GRBs.

astro-ph.HE↗
arXiv2026

A General Framework for Multimodal LLM-Based Multimedia Understanding in Large-Scale Recommendation Systems

Conventional recommendation systems frequently fail to fully exploit the high-dimensional semantic signals inherent in multimedia content, thereby limiting the fidelity of user preference modeling. While Multimodal Large Language Models (MM-LLMs) offer robust mechanisms for interpreting such complex data, their integration into latency-constrained, industrial-scale architectures remains a significant challenge. To address this, we propose a generalized framework for MM-LLM-driven multimedia understanding. Our methodology employs a tripartite architecture encompassing content interpretation, representation extraction, and systematic pipeline integration, instantiated via a LLaMA2-based model that generates descriptive captions subsequently ingested as tokenized categorical features. Empirical evaluation demonstrates the efficacy of this approach, yielding a $0.35\%$ increase in offline AUC and a $0.02\%$ improvement in online metrics at scale, substantiating the practical viability of leveraging MM-LLMs to enhance large-scale recommendation performance.

cs.IR↗
arXiv2026

Semantic Communication for Multi-Satellite Massive MIMO Transmission: A Mixture of Cooperative Modes Framework

This paper investigates semantic communications (SemComs) for multi-satellite cooperative massive multiple-input multiple-output (MIMO) transmission, where multiple massive-MIMO satellites jointly serve a common set of multi-antenna user terminals. For the first time, SemComs with image transmission task are integrated into satellite massive MIMO and multi-satellite cooperative transmission. For the two representative cooperative modes, namely coherent transmission (CT) and non-coherent transmission (NCT), we develop multi-satellite CT (MSCT) and multi-satellite NCT (MSNCT) SemCom frameworks, respectively. MSCT adopts a symmetric architecture, whereas MSNCT introduces transmitter-side stream allocation and a two-stage receiver design that combines per-stream semantic extraction with cross-stream semantic-interference exploitation. To instantiate MSCT, we further design a symmetric encoder and decoder network based on hybrid Swin-Transformer and lightweight bottleneck convolutional neural network (CNN) blocks, termed HSTC, where Swin Transformer provides scalable computation and the CNN branch improves performance and convergence. For MSNCT, a Transformer-based backbone is employed to support cross-stream interference exploitation through global attention. Building on these two frameworks, we propose a mixture of cooperative modes (MoCM) framework, in which a permutation-invariant network dynamically switches between MSCT and MSNCT using multi-satellite statistical channel state information, thereby balancing semantic performance and complexity. Simulation results under practical configurations demonstrate the performance gains of the proposed frameworks.

eess.SP↗
arXiv2026

CPI-C: Cool Planet Imaging Coronagraph on Chinese Space Station Survey Telescope

Cool Planet Imaging Coronagraph (CPI-C) on Chinese Space Station Survey Telescope (CSST) is proposed to direct image the cool planets around nearby solar-type stars (within 40 pc). The core scientific objective of CPI-C is to conduct high-contrast directly imaging surveys of exoplanets ranging in size from Neptune-like to Jupiter-like, located at separations of 0.5 to 5 AU from their host stars, and to perform systematic spectroscopic analysis of the detected planets through high-precision multi-band photometry. CPI-C employs a step-transmission apodization technique to suppress the diffraction noises from the telescope pupil and a precise phase correction technique to eliminate the speckle noises due to imperfections of the optical surfaces. The contrast requirement is better than $10^{-8}$ at an inner working angle (IWA) of $3-4λ/D$, in the visible wavelength from 600 nm to 900 nm. CPI-C will be the first space-based instrument capable of directly imaging the reflection light from the cool exoplanets in the visible wavelength enabling the measurement of key physical parameters such as the effective temperature, surface gravity, radius, mass, and other key parameters. The potential observation results will significantly contribute to further understand the formation and evolution mechanisms of planets, which will also lay a solid foundation for future confirmation of the Earth-twins in the next generation space flagship missions.

astro-ph.EP↗
arXiv2026

Characterizing LLM-driven Social Network: The Chirper.ai Case

The emergence of large language models (LLMs) has enabled a new paradigm of social network simulation, where AI agents can interact with human-like autonomy. Recent research has explored collective behavioral patterns and structural characteristics of LLM agents within simulated networks. However, empirical comparisons between LLM-driven and human-driven online social networks remain scarce, limiting our understanding of how LLM agents differ from human users. This paper presents a large-scale analysis of Chirper.ai, an X/Twitter-like social network entirely populated by LLM agents, comprising over 65,000 agents and 7.7 million AI-generated posts. For comparison, we collect a parallel dataset from Mastodon, a human-driven decentralized social network, with over 117,000 users and 16 million posts. We examine key differences between LLM agents and humans in posting behaviors, abusive content, and social network structures. Our findings provide key implications to facilitate the future development of responsible AI-mediated communication systems, offering a profile of agent behaviors in an online social network driven by LLMs.

cs.SI↗
arXiv2026

Defending against Patch-Based and Texture-Based Adversarial Attacks with Spectral Decomposition

Adversarial examples present significant challenges to the security of Deep Neural Network (DNN) applications. Specifically, there are patch-based and texture-based attacks that are usually used to craft physical-world adversarial examples, posing real threats to security-critical applications such as person detection in surveillance and autonomous systems, because those attacks are physically realizable. Existing defense mechanisms face challenges in the adaptive attack setting, i.e., the attacks are specifically designed against them. In this paper, we propose Adversarial Spectrum Defense (ASD), a defense mechanism that leverages spectral decomposition via Discrete Wavelet Transform (DWT) to analyze adversarial patterns across multiple frequency scales. The multi-resolution and localization capability of DWT enables ASD to capture both high-frequency (fine-grained) and low-frequency (spatially pervasive) perturbations. By integrating this spectral analysis with the off-the-shelf Adversarial Training (AT) model, ASD provides a comprehensive defense strategy against both patch-based and texture-based adversarial attacks. Extensive experiments demonstrate that ASD+AT achieved state-of-the-art (SOTA) performance against various attacks, outperforming the APs of previous defense methods by 21.73%, in the face of strong adaptive adversaries specifically designed against ASD. Code available at https://github.com/weiz0823/adv-spectral-defense .

cs.CV↗
arXiv2026

Toward Multi-Satellite Cooperative Transmission: A Joint Framework for CSI Acquisition, Feedback, and Phase Synchronization

The stringent link budget, caused by long propagation distances and payload constraints, poses a fundamental bottleneck for single-satellite transmission. Although LEO mega-constellations make multi-satellite cooperative transmission (MSCT), such as distributed precoding (DP), increasingly feasible, its cooperative gains critically rely on stringent time-frequency-phase synchronization (TFP-Sync), which is difficult to maintain under rapid channel variation and feedback latency. To address this issue, this paper proposes a joint CSI acquisition, feedback, and phase-level synchronization (JCAFPS) framework for MSCT. Specifically, to enable reliable, overhead-efficient CSI acquisition, we design a beam-domain adjustable phase-shift tracking reference signal (TRS) transmission scheme, along with criteria for the TRS and CSI-feedback periods. Then, exploiting deterministic orbital motion and dominant LoS propagation, we establish a polynomial model for the temporal evolution of delay and Doppler shift, and derive an OFDM-based multi-satellite signal model under non-ideal synchronization. The analysis reveals that, unlike the single-satellite case, the composite multi-satellite channel exhibits nonlinear time-frequency-varying phase behavior, necessitating symbol- and subcarrier-wise phase precompensation for coherent transmission. Based on these results, we develop a practical closed-loop realization integrating single-TRS-based channel parameter estimation, multi-TRS-based channel prediction, predictive CSI feedback, and user-specific TFP precompensation. Numerical results demonstrate that the proposed framework achieves accurate CSI acquisition and precise TFP-Sync, enabling DP-based dual-satellite cooperative transmission to approach the theoretical 6 dB power gain over single-satellite transmission, while remaining robust under extended prediction durations and enlarged TRS periods.

eess.SP↗
arXiv2026

DSCSNet: A Dynamic Sparse Compression Sensing Network for Closely-Spaced Infrared Small Target Unmixing

Due to the limitations of optical lens focal length and detector resolution, distant clustered infrared small targets often appear as mixed spots. The Close Small Object Unmixing (CSOU) task aims to recover the number, sub-pixel positions, and radiant intensities of individual targets from these spots, which is a highly ill-posed inverse problem. Existing methods struggle to balance the rigorous sparsity guarantees of model-driven approaches and the dynamic scene adaptability of data-driven methods. To address this dilemma, this paper proposes a Dynamic Sparse Compressed Sensing Network (DSCSNet), a deep-unfolded network that couples the Alternating Direction Method of Multipliers (ADMM) with learnable parameters. Specifically, we embed a strict $\ell_1$-norm sparsity constraint into the auxiliary variable update step of ADMM to replace the traditional $\ell_2$-norm smoothness-promoting terms, which effectively preserves the discrete energy peaks of small targets. We also integrate a self-attention-based dynamic thresholding mechanism into the reconstruction stage, which adaptively adjusts the sparsification intensity using the sparsity-enhanced information from the iterative process. These modules are jointly optimized end-to-end across the three iterative steps of ADMM. Retaining the physical logic of compressed sensing, DSCSNet achieves robust sparsity induction and scene adaptability, thus enhancing the unmixing accuracy and generalization in complex infrared scenarios. Extensive experiments on the synthetic infrared dataset CSIST-100K demonstrate that DSCSNet outperforms state-of-the-art methods in key metrics such as CSO-mAP and sub-pixel localization error.

cs.CV↗