Search arXiv⌕ Search

arXiv subjects

Liang Chen

Publications and source records attributed to Liang Chen.

At least 91 records · Page 5Linked to original sources
arXiv2026

RoadmapBench: Evaluating Long-Horizon Agentic Software Development Across Version Upgrades

Coding agents are increasingly deployed in real software development, where a single version iteration requires months of coordinated work across many files. However, most existing benchmarks focus predominantly on single-issue bug fixes from Python repositories, with coarse pass/fail evaluation outcomes, and thus fail to capture long-horizon, multi-target development at real engineering scale. To address this gap, we present RoadmapBench, a benchmark of 115 long-horizon coding tasks grounded in real open-source version upgrades across 17 repositories and 5 programming languages. Each task places the agent on a source-version code snapshot and provides a multi-target roadmap instruction requiring it to implement the functionality introduced in the target version, with a median modification of 3,700 lines across 51 files. We conduct a systematic evaluation on thirteen frontier models and find that even the strongest, Claude-Opus-4.7, resolves only 39.1% of tasks, while the weakest achieves merely 5.2%, in stark contrast to existing bug-fix benchmarks, suggesting that long-horizon software development remains a largely unsolved problem.

cs.SE↗
arXiv2026

Improving MLLM Training Efficiency via Stage-Aware Sparsity

Multimodal Large Language Models (MLLMs) have demonstrated outstanding performance across a variety of domains. However, training MLLMs is often inefficient, as much of the computation is redundant due to the long input sequences from multimodal data and underutilized inter-layer operations. Notably, such redundancy is not static but varies across different stages of training. Building on this observation, we shift the focus to the training process itself and propose a training-efficient framework based on sparse representations, termed the Sparse Training Scheme (STS). Instead of applying a uniform sparsity strategy, STS adopts a stage-aware design that adapts to different sources of redundancy during training. Specifically, the framework consists of two complementary components: the Visual Token Compressor, which reduces the information load by compressing visual tokens during modality alignment, and the Layer Dynamic Skipper, which mitigates computational overhead by dynamically skipping unnecessary layers during instruction tuning. Our approach is broadly applicable to diverse MLLM architectures and has been extensively evaluated on multiple benchmarks, demonstrating its effectiveness and efficiency.

cs.LG↗
arXiv2026

Enhanced detection of electric field signals via squeezing-induced stochastic resonance

Stochastic resonance (SR) could amplify weak electric-field signals in nonlinear systems by means of the externally injected noises. Here we propose and experimentally demonstrate a modified SR method, termed squeezing-induced SR, implemented in the system involving a trapped ion behaving as a Duffing oscillator. We find that squeezing the phase noise of the oscillator results in amplified fluctuation of the corresponding amplitude, which helps achieve the SR. Since no auxiliary noise source is needed, the squeezing-induced SR may enhance the signal-to-noise ratio by 4.28 $\pm$ 0.39 dB compared to the conventional noise-induced SR under identical conditions of the electric-field detection. This technique offers a promising approach for developing atomic ion sensors for detecting weak electric-field signals.

physics.atom-ph↗
arXiv2026

Step-wise Rubric Rewards for LLM Reasoning

Reinforcement Learning with Verifiable Rewards (RLVR) is widely used to improve reasoning in large language models, but rewards only final-answer correctness with no supervision over intermediate steps. Rubric-based methods such as Rubrics as Rewards (RaR) introduce finer-grained supervision by scoring rollouts against structured criteria, yet the rubric scores are still aggregated into a single scalar applied to the entire response, causing three weaknesses: loss of multi-criterion structure, uniform supervision of correct and incorrect steps, and reward hacking through unbounded self-correction. On 1,000 problems, we find 18.2% of steps in correct-answer responses are wrong yet positively rewarded, while 49.9% of steps in incorrect-answer responses are correct yet penalized. We introduce Step-wise Rubrics as Rewards (SRaR), an RLVR framework that (i) uses an LLM judge to attribute each rubric item to a specific reasoning step, (ii) normalizes per-step rubric scores across rollouts so only steps whose quality varies produce a learning signal, and (iii) combines the per-step reward with the outcome reward through a decoupled advantage estimator that keeps the outcome baseline stable. We further build a 16K-problem rubric dataset by contrastively distilling rubric items from correct and flawed reasoning paths sampled from a strong model. Across six mathematical reasoning benchmarks, SRaR improves average accuracy over RaR by 3.57 points on Qwen3-8B and 2.75 points on Qwen3-32B, raises the Faithful Reasoning Rate on AIME 2025 from 34.5% to 46.7%, and reduces self-correction looping from 48.1% to 26.5%.

cs.LG↗
arXiv2026

Chasing the neutrino blazar candidates II: SED modeling with hadronic model

Blazars are promising candidates for high energy neutrino sources, yet the physical origin of their neutrino emission remains uncertain. In this work, we extend our previous study by modeling the broadband spectral energy distributions (SEDs) of 103 neutrino blazar candidates (NBCs) within a hadronic framework. To estimate the maximum possible neutrino output, we adopt an assumption in which the high energy emission is dominated by p gamma interactions and the contribution from leptonic inverse Compton scattering is strongly suppressed. From the SED modeling, we constrain nine key parameters describing the emission region and particle energy distributions. We perform a partial correlation analysis to investigate the relationship between neutrino luminosity and electromagnetic emission, and we found a weak or moderate correlation between optical R band and neutrino emission. Our model predicts prominent proton synchrotron emission peaking in the MeV band for most sources, with 99 out of 103 NBCs exhibiting proton synchrotron peaks within 0.1 to 100 MeV, highlighting the MeV band as a key window for distinguishing between leptonic and hadronic scenarios. Based on the model-predicted maximum neutrino fluxes, we find that three NBCs are potentially detectable by IceCube, while up to 22, 45, and 62 sources may be detectable by KM3NeT, NEON, and TRIDENT, respectively. These results provide testable predictions for future multi-messenger observations and offer new insights into the composition and radiation mechanisms of blazar jets.

astro-ph.HE↗
arXiv2026

Variational Matrix-Learning Fourier Networks for Parametric Multiphysics Surrogates

Multiphysics simulation is critical for system-technology co-optimization (STCO) in chiplet-based design, but repeated finite-element solutions of PDE-governed problems are computationally expensive in parametric design exploration. This paper proposes a variational matrix-learning Fourier network (VMLFN) for efficient parametric multiphysics surrogate modeling. VMLFN constructs a log-space sine neural representation with randomly sampled spectral frequencies, frequency-dependent decay regulation, and embedded Dirichlet boundary conditions. With fixed hidden-layer parameters, the output-layer weights are determined by reformulating the governing PDEs into variational weak forms and enforcing the stationarity condition of the resulting energy functional. This converts physics-informed training into a linear matrix-solving problem, requiring only first-order derivatives and avoiding both high-order automatic differentiation and penalty-coefficient tuning. A heuristic frequency-scanning algorithm is further introduced to select a problem-adaptive maximum frequency that covers the dominant spectral range of the target problem. The proposed method is validated on heat conduction, solid mechanics, and Helmholtz wave propagation problems. Results from five benchmark cases demonstrate that VMLFN delivers accurate full-field predictions with substantial speedup over conventional physics-informed neural networks and repeated finite-element simulations.

cs.LG↗
arXiv2026

A geometry aware framework enhances noninvasive mapping of whole human brain dynamics

Non-invasive electrophysiology lacks methods that accurately reconstruct whole-brain spatiotemporal dynamics while incorporating individual cortical geometry, leaving current electroencephalography and magnetoencephalography source imaging limited by simplistic or biologically implausible priors. Here, we show that embedding participant-specific Geometric Basis Functions (GBFs), eigenmodes derived from each individual's cortical surface, provides a powerful anatomic constraint that resolves the inverse problem and improves reconstruction fidelity. The method reconstructs neural sources as linear combinations of geometric basis functions, thereby aligning source estimates with the geometric organization of neural dynamics. We validate GBF across the Meta-Source Benchmark, task-evoked data, resting-state networks, intracranial stimulation, and epilepsy data. The results demonstrate that GBF yields high localization accuracy and captures fast spatiotemporal dynamics consistent with anatomical pathways. These findings suggest that both spontaneous and evoked whole-brain activity can be described by hundreds of geometric modes, providing a compact yet accurate representation of neural sources. By linking cortical geometry to electrophysiological dynamics, GBF offers a versatile source imaging tool for both scientific and clinical applications.

q-bio.NC↗
arXiv2026

Multiplicative Contractions, Additive Recoveries: Functional-Form Restrictions on Risk Exposure Dynamics

We test a regime-conditional functional-form restriction on aggregate risk-exposure dynamics implied by VaR-constrained intermediary models: exposures contract multiplicatively when capital constraints bind and grow additively (level-independent) when slack. The contraction half follows from binding VaR constraints (Brunnermeier and Pedersen 2009; Adrian and Shin 2010; He and Krishnamurthy 2013). The additive-rebuild prediction is derived under constant-rate capital replenishment; we test the joint restriction on FINRA monthly margin debt (1997-2026). Two findings. First, regime-interacted regression of detrended margin growth on lagged level (T=350 months) yields calm slope -0.040 (p=0.082, additive) and stress slope -0.205 (p<0.001, multiplicative); Wald test on regime x level interaction rejects equal dependence (p=0.0016). Second, the restriction implies drawdown-recovery duration ratio increases with crash depth. On 73 S&P 500 episodes (1950-2026), Cox model gives depth coefficient -13.75 (p<10^{-7}): 75% lower recovery hazard per 10pp deeper drawdown. Continuous-depth regression yields beta=1.22 (p=0.047); beta=1.59 (p<0.001) excluding 1980-82 Volcker. Median duration ratio for crashes >30% is 3.1x; replicates across eight other equity indices. Calibrated Heston, Markov-switching, and block bootstrap nulls match price-level duration asymmetry but lack an exposure state variable, so cannot speak to the regime-conditional flip on direct exposures. We do not claim the exposure test identifies the intermediary mechanism: FINRA margin debt is a noisy proxy. We claim only that the regime-conditional functional form is a sharper target than return-level moments alone, and confirming it on margin debt is consistent with -- not proof of -- the constrained-intermediary mechanism. A companion test on CFTC weekly speculative positioning is left for future work (Sections 5.2 and F).

q-fin.RM↗
arXiv2026

Ultra-high-energy $γ$-ray imprints from PeV particles accelerated by supernova remnants

The quest for the origin of cosmic ray (CRs) is a fundamental issue in astrophysics. Shocks of supernova remnants (SNRs) have been considered as the dominant contributors to Galactic CRs below the spectral knee near $\sim 3$ petaelectronvolt (PeV). Whether SNRs are efficient accelerators of particles beyond PeV energies has long been debated. Here we report observations of very-high-energy $γ$-ray emission up to hundreds of TeV from two middle age shell-type SNRs, G150.3$+$4.5 and $γ$-Cygni, with the Large High Altitude Air Shower Observatory (LHAASO). Two (or three) distinct morphological/spectral components with convex spectral shapes are observed in both sources, with the low-energy one being more extended than the high-energy one. %Although it is possible that these high-energy components may be driven by powerful pulsars, The likely association of the high-energy component with molecular clouds at similar distances, and the weakness/absence of pulsar wind nebulae (PWNe) inside these SNRs clearly indicate for the first time that the highest energy emission is produced by collision of hadronic CRs up to PeV energies with the clouds. These results are compatible with the classic model prediction that PeV particles accelerated near the end of the free expansion phase of SNR evolution can illuminate nearby molecular clouds (MCs) to produce strong $γ$-ray emission.

astro-ph.HE↗
arXiv2026

Fully multiplexed photonic tensor computing

Tensor operations dominate modern computational workloads, yet their further acceleration demands hardware platforms with greater parallelism. Although photonic computing provides a compelling route for parallel processing, fully exploiting all native multiplexing dimensions of optical fields is impeded by the challenges in routing and programming light in all dimensions simultaneously. Here we introduce FieldCore, a fully multiplexed photonic tensor core that jointly harnesses wavelength, radio-frequency, guided-mode, time and space dimensions, thereby enabling parallelism to scale multiplicatively within a single optical field. Enabled by inverse-designed silicon photonics, FieldCore preserves a uniform programmed computation across all multiplexed channels in parallel. Experimentally, we validate and benchmark its performance from ultra-high-baudrate arithmetic operations to high-fidelity image convolution and parallel handwritten-digit recognition. We further use FieldCore to unlock applications that naturally require high-dimensional data processing, such as high-dimensional hyperspectral classification and massively parallel mechanical fault diagnosis. Our FieldCore supports an estimated aggregate compute throughput of 69.12 tera operations per second (TOPS) and accommodates up to 1,800 parallel input streams within a single core, establishing a scalable paradigm for fully multiplexed photonic tensor computing and AI inference.

physics.optics↗
arXiv2026

Expressive yet Efficient Feature Expansion with Adaptive Cross-Hadamard Products

Recent theoretical advances reveal that the Hadamard product induces nonlinear representations and implicit high-dimensional mappings for the field of deep learning, yet their practical deployment in resource-constrained vision models remains largely unexplored. To address this gap, we introduce the Adaptive Cross-Hadamard (ACH) module, a novel operator that embeds learnability through differentiable discrete sampling and dynamic softsign normalization. This facilitates highly efficient feature reuse without incurring additional convolutional parameters, while ensuring stable gradient flow. Integrated into Hadaptive-Net (Hadamard Adaptive Network) via neural architecture search, our approach achieves unprecedented efficiency. Comprehensive experiments demonstrate state-of-the-art accuracy/speed trade-offs on image classification tasks, establishing Hadamard operations as specific building blocks for efficient vision models.

cs.CV↗
arXiv2026

EEPO: Exploration-Enhanced Policy Optimization via Sample-Then-Forget

Balancing exploration and exploitation remains a central challenge in reinforcement learning with verifiable rewards (RLVR) for large language models (LLMs). Current RLVR methods often overemphasize exploitation, leading to entropy collapse, diminished exploratory capacity, and ultimately limited performance gains. Although techniques that increase policy stochasticity can promote exploration, they frequently fail to escape dominant behavioral modes. This creates a self-reinforcing loop -- repeatedly sampling and rewarding dominant modes -- that further erodes exploration. We introduce Exploration-Enhanced Policy Optimization (EEPO), a framework that promotes exploration via two-stage rollouts with adaptive unlearning. In the first stage, the model generates half of the trajectories; it then undergoes a lightweight unlearning step to temporarily suppress these sampled responses, forcing the second stage to explore different regions of the output space. This sample-then-forget mechanism disrupts the self-reinforcing loop and promotes wider exploration during rollouts. Across five reasoning benchmarks, EEPO outperforms GRPO, achieving average relative gains of 24.3% on Qwen2.5-3B, 33.0% on Llama3.2-3B-Instruct, and 10.4% on Qwen3-8B-Base.

cs.CL↗
arXiv2026

Cygnus X-3: A variable petaelectronvolt gamma-ray source

We report the discovery of variable $γ$-rays up to petaelectronvolt from Cygnus X-3, an iconic X-ray binary. The $γ$-ray signal was detected with a statistical significance of approximately 10 $σ$ by the Large High Altitude Air Shower Observatory (LHAASO). Its intrinsic spectral energy distribution (SED), extending from 0.06 to 3.7 PeV, shows a pronounced rise toward 1 PeV after accounting for absorption by the cosmic microwave background radiation. We find variability on month-long timescales at a significance of $8.6 σ$, coinciding with a high state of the GeV gamma-ray flux detected by the Fermi-LAT. This,together with a 3.2$σ$ evidence for orbital modulation, suggests that the PeV $γ$-rays originate within, or in close proximity to, the binary system itself. The observed energy spectrum and temporal modulation can be naturally explained by $γ$-ray production through photomeson processes in the innermost region of the relativistic jet, where protons need to be accelerated to tens of PeV energies.

astro-ph.HE↗
arXiv2026

Precise Measurement of the Cosmic Ray Helium Spectrum above 0.1 PeV

We report a measurement of the cosmic ray helium energy spectrum in the energy interval 0.16 -- 13~PeV, derived by subtracting the proton spectrum from the light component~(proton and helium) spectrum obtained with observations made by the Large High Altitude Air Shower Observatory~(LHAASO) under a consistent energy scale. The helium spectrum shows a significant hardening centered at $E \simeq$ 1.1~PeV, followed by a softening at $\sim$ 7 PeV, indicating the appearance of a helium `knee'. Comparing the proton and helium spectra in the LHAASO energy range reveals some remarkable facts. In the lower part of this range, in contrast to the behavior at lower energies, the helium spectrum is significantly softer than the proton spectrum. This results in protons overtaking helium nuclei and becoming the largest cosmic ray component at $E \simeq$ 0.7 PeV. A second crossing of the two spectra is observed at $E \simeq$ 5 PeV, above the proton knee, when helium nuclei overtake protons to become the largest cosmic ray component again. These results have important implications for our understanding of the Galactic cosmic ray sources.

astro-ph.HE↗
arXiv2026

Synchrotron Self-Absorption Spectral Modeling Reveals a Magnetically Driven Shock-in-Jet Scenario in Blazar 1156+295

Unveiling the launching and driving mechanisms of powerful jets in active galactic nuclei (AGNs) is crucial for understanding the co-evolution of supermassive black holes (SMBHs) and their host galaxies. 1156+295 is a blazar at a redshift of z=0.729 and exhibits significant variability in long-term radio monitoring. Using multi-frequency Effelsberg single-dish flux density data from 2007 to 2012, we performed synchrotron self-absorption (SSA) spectral modeling and extracted the turnover frequency and turnover flux density. By combining SSA spectral modeling with the core size and brightness temperature from quasi-simultaneous very long baseline interferometry (VLBI) images, we estimated the jet magnetic-field strength and magnetic flux, and investigated their temporal evolution in 1156+295. The evolution of radio flux density, spectral shape, and jet structure is consistent with the shock-in-jet framework. The inferred magnetic flux reaching or exceeding the magnetically arrested disk (MAD) threshold, together with evidence that magnetic energy release precedes the radio flares, supports a magnetically driven jet scenario. Overall, our results place magnetic-field measurements, spectral evolution, and inner-jet structural changes on a common timeline, providing observational constraints on their coupled evolution during flares.

astro-ph.HE↗
arXiv2026

Towards unified brain-to-text decoding across speech production and perception

Speech production and perception are the main ways humans communicate daily. Prior brain-to-text decoding studies have largely focused on a single modality and alphabetic languages. Here, we present a unified brain-to-sentence decoding framework for both speech production and perception in Mandarin Chinese. The framework exhibits strong generalization ability, enabling sentence-level decoding when trained only on single-character data and supporting characters and syllables unseen during training. In addition, it allows direct and controlled comparison of neural dynamics across modalities. Mandarin speech is decoded by first classifying syllable components in Hanyu Pinyin, namely initials and finals, from neural signals, followed by a post-trained large language model (LLM) that maps sequences of toneless Pinyin syllables to Chinese sentences. To enhance LLM decoding, we designed a three-stage post-training and two-stage inference framework based on a 7-billion-parameter LLM, achieving overall performance that exceeds larger commercial LLMs with hundreds of billions of parameters or more. In addition, several characteristics were observed in Mandarin speech production and perception: speech production involved neural responses across broader cortical regions than auditory perception; channels responsive to both modalities exhibited similar activity patterns, with speech perception showing a temporal delay relative to production; and decoding performance was broadly comparable across hemispheres. Our work not only establishes the feasibility of a unified decoding framework but also provides insights into the neural characteristics of Mandarin speech production and perception. These advances contribute to brain-to-text decoding in logosyllabic languages and pave the way toward neural language decoding systems supporting multiple modalities.

q-bio.NC↗
arXiv2026

Derived equivalences and delooping levels

We construct a finite-dimensional algebra derived equivalent to the example of Kershaw--Rickard. For the Kershaw--Rickard example the delooping level and the sub-derived delooping level are both infinite, while for our algebra both invariants are $0$. Thus the finiteness of these invariants is not preserved under derived equivalences.

math.RT↗