arXiv subjects

Yueyang Zhang

Publications and source records attributed to Yueyang Zhang.

At least 19 recordsLinked to original sources

arXiv2026

Ultrafast and long-term stability Integrated Pockels laser with thin-film PZT

Integrated tunable lasers are central to coherent communications, wavelength-routed optical interconnects, spectroscopy and frequency-modulated continuous-wave LiDAR, yet chip-scale sources rarely combine broad wavelength coverage, nanosecond switching, high spectral purity and stable high-power operation. Here we demonstrate a frequency-agile hybrid external-cavity laser enabled by the Pockels effect in thin-film lead zirconate titanate (PZT). The strong linear electro-optic response of PZT provides direct, non-thermal tuning of compact microring resonators with a wavelength-tuning efficiency of 17 pm/V. In contrast to conventional anisotropic Pockels materials, the near-isotropic in-plane electro-optic behaviour of thin-film PZT relaxes crystal-axis layout constraints, allowing efficient Vernier wavelength selection in compact ring cavities. The PZT resonators also show no measurable photorefractive resonance distortion and no resolvable DC-bias drift during operation, preserving stable wavelength-selective feedback. The demonstrated laser achieves an 82 nm tuning range, a 5 mW fiber-coupled output power, a side-mode suppression ratio (SMSR) exceeding 56.7 dB, and a wavelength-switching time of 5.5 ns. These results establish thin-film PZT photonics as a powerful electro-optic platform for compact, high-power, and frequency-agile integrated laser sources.

physics.optics↗

arXiv2026

Thinking as Compression: Your Reasoning Model is Secretly a Context Compressor

Context compression aims to shorten long context inputs with minimal information loss for LLM inference acceleration. While existing methods have shown promise, they typically rely on complex compression modules or compression-specific training, leaving the intrinsic capabilities of LLMs underexplored. In contrast, this work reveals that a thinking model itself can naturally compress long contexts by organizing task-relevant information. We thus derive Thinking as Compression (TaC), a new compression paradigm that treats thinking itself as compressed context. Without relying on specific dedicated compressor, TaC directly prompts the thinking model to generate thinking traces as the shortened context, already outperforming most representative compression methods. Further, given that raw thinking output may struggle with budget control and shortcut behaviors, we introduce Thinking as Compression Constrained (TaC-C), leveraging a simple reward-driven optimization framework to elicit intrinsic thinking as compact and controllable compressed context. Experiments across four long-context QA benchmarks demonstrate that TaC-C consistently outperforms existing baselines. At 4x and 8x compression ratios, it surpasses the strongest competitor by 17.4% and 23.4% in average F1, and by 15.7% and 21.7% in average Exact Match Score (EM), respectively.

cs.AI↗

arXiv2026

Formal O(N3)-Scaling Second-Order Perturbation Theory by Block Tensor Decomposition: Implementation on MP2 and rPT2

Block tensor decomposition (BTD) and canonical polyadic decomposition (CPD) are combined into a unified $O(N^3)$-scaling framework for second-order perturbation theory (PT2), demonstrated on MP2 and renormalized PT2 (rPT2). BTD constructs the tensor hyper-contraction kernel at $O(N^3)$ via a dual-grid scheme; CPD factorizes the exchange channel through a block-based two-stage ALS. An asymmetric half-kernel design applies bare Coulomb to one vertex and coupling-constant-averaged screening to the other, capturing the SOSEX component of rPT2 without a frequency-dependent CPD. For MP2, BTD-CPD reproduces canonical RI-MP2 to 0.058~kcal/mol per heavy atom. For rPT2@PBE0 on the S66x8 benchmark, the mean absolute error is 0.36~kcal/mol (ME $-$0.19, RMSE 0.46) over 528 data points. The CPD-compressed intermediates yield $O(N^2)$ storage alongside $O(N^3)$ scaling.

physics.chem-ph↗

arXiv2026

Hybrid-plasticity Photonic Synapses Enabling Hardware-Level Neural Reuse

Biological intelligence is distinguished by neural reuse, the capacity to preserve established learning memory while repurposing it for new tasks and dynamic environments. Bringing this capability to photonic hardware requires hybrid plasticity, namely the coexistence of long-term synaptic plasticity for persistent weight storage and short-term synaptic plasticity for rapid, reversible adaptation within a single synaptic element; however, current photonic architectures lack such a unified mechanism. Here, we demonstrate a hybrid-plasticity photonic synapse on thin-film lead zirconate titanate (PZT) that couples non-volatile and volatile modes to enable hardware-level neural reuse. Crucially, high-speed refresh operations can be superimposed without perturbing the stored weight. Such a neural-reuse framework yields a convergence speedup of over 20-fold and reduces the weight updates by approximately 30-fold compared with random initialization. These results establish hybrid-plasticity photonic synapses as a pathway toward on-chip learning systems that are both memory-preserving and rapidly adaptable.

physics.optics↗

arXiv2026

When Less is More: The LLM Scaling Paradox in Context Compression

Scaling up model parameters has long been a prevalent training paradigm driven by the assumption that larger models yield superior generation capabilities. However, under lossy context compression in a compressor--decoder setup, we find a \textbf{\textit{Size-Fidelity Paradox}}: increasing compressor size can lessen the faithfulness of reconstructed contexts though reconstruction error decreases. Across 27 compressor setups spanning model families, scales, and compression rates, we coin this paradox arising from two dominant factors: 1) \textit{knowledge overwriting}: larger models increasingly replace source facts with their own prior beliefs, \textit{e.g.}, ``the white strawberry`` $\to$ ``the red strawberry``; and 2) \textit{semantic drift}: larger models tend to paraphrase or restructure content instead of reproducing it verbatim, \textit{e.g.}, ``Alice hit Bob`` $\to$ ``Bob hit Alice``. Interestingly, this paradox persists across varied settings, with mid-sized compressors often outperforming larger ones in faithful recovery. By analyzing the compressed memory via embedding geometry and reconstruction determinacy, we further reveal that compressors tend to organize memory across broader semantic subspaces, yielding more ambiguous representations prone to overwriting, drift, and weakened recovery. These findings complement existing evaluations of context compression and expose a breakdown of scaling laws when the objective shifts from plausible generation to faithful preservation.

cs.LG↗

arXiv2026

ConsistRM: Improving Generative Reward Models via Consistency-Aware Self-Training

Generative reward models (GRMs) have emerged as a promising approach for aligning Large Language Models (LLMs) with human preferences by offering greater representational capacity and flexibility than traditional scalar reward models. However, GRMs face two major challenges: reliance on costly human-annotated data restricts scalability, and self-training approaches often suffer from instability and vulnerability to reward hacking. To address these issues, we propose ConsistRM, a self-training framework that enables effective and stable GRM training without human annotations. ConsistRM incorporates the Consistency-Aware Answer Reward, which produces reliable pseudo-labels with temporal consistency, thereby providing more stable model optimization. Moreover, the Consistency-Aware Critique Reward is introduced to assess semantic consistency across multiple critiques and allocates fine-grained and differentiated rewards. Experiments on five benchmark datasets across four base models demonstrate that ConsistRM outperforms vanilla Reinforcement Fine-Tuning (RFT) by an average of 1.5%. Further analysis shows that ConsistRM enhances output consistency and mitigates position bias caused by input order, highlighting the effectiveness of consistency-aware rewards in improving GRMs. Our implementation is available at https://github.com/yuliangCarmelo/ConsistRM.

cs.AI↗

arXiv2026

ReflectRM: Boosting Generative Reward Models via Self-Reflection within a Unified Judgment Framework

Reward Models (RMs) are critical components in the Reinforcement Learning from Human Feedback (RLHF) pipeline, directly determining the alignment quality of Large Language Models (LLMs). Recently, Generative Reward Models (GRMs) have emerged as a superior paradigm, offering higher interpretability and stronger generalization than traditional scalar RMs. However, existing methods for GRMs focus primarily on outcome-level supervision, neglecting analytical process quality, which constrains their potential. To address this, we propose ReflectRM, a novel GRM that leverages self-reflection to assess analytical quality and enhance preference modeling. ReflectRM is trained under a unified generative framework for joint modeling of response preference and analysis preference. During inference, we use its self-reflection capability to identify the most reliable analysis, from which the final preference prediction is derived. Experiments across four benchmarks show that ReflectRM consistently improves performance, achieving an average accuracy gain of +3.7 on Qwen3-4B. Further experiments confirm that response preference and analysis preference are mutually reinforcing. Notably, ReflectRM substantially mitigates positional bias, yielding +10.2 improvement compared with leading GRMs and establishing itself as a more stable evaluator. Our code is available at https://github.com/yuliangCarmelo/ReflectRM.

cs.AI↗

arXiv2026

Advancing General-Purpose Reasoning Models with Modular Gradient Surgery

Reinforcement learning (RL) has played a central role in recent advances in large reasoning models (LRMs), yielding strong gains in verifiable and open-ended reasoning. However, training a single general-purpose LRM across diverse domains remains challenging due to pronounced domain heterogeneity. Through a systematic study of two widely used strategies, Sequential RL and Mixed RL, we find that both incur substantial cross-domain interference at the behavioral and gradient levels, resulting in limited overall gains. To address these challenges, we introduce **M**odular **G**radient **S**urgery (**MGS**), which resolves gradient conflicts at the module level within the transformer. When applied to Llama and Qwen models, MGS achieves average improvements of 4.3 (16.6\%) and 4.5 (11.1\%) points, respectively, over standard multi-task RL across three representative domains (math, general chat, and instruction following). Further analysis demonstrates that MGS remains effective under prolonged training. Overall, our study clarifies the sources of interference in multi-domain RL and presents an effective solution for training general-purpose LRMs.

cs.CL↗

arXiv2026

TRE: Encouraging Exploration in the Trust Region

Entropy regularization is a standard technique in reinforcement learning (RL) to enhance exploration, yet it yields negligible effects or even degrades performance in Large Language Models (LLMs). We attribute this failure to the cumulative tail risk inherent to LLMs with massive vocabularies and long generation horizons. In such environments, standard global entropy maximization indiscriminately dilutes probability mass into the vast tail of invalid tokens rather than focusing on plausible candidates, thereby disrupting coherent reasoning. To address this, we propose Trust Region Entropy (TRE), a method that encourages exploration strictly within the model's trust region. Extensive experiments across mathematical reasoning (MATH), combinatorial search (Countdown), and preference alignment (HH) tasks demonstrate that TRE consistently outperforms vanilla PPO, standard entropy regularization, and other exploration baselines. Our code is available at https://github.com/WhyChaos/TRE-Encouraging-Exploration-in-the-Trust-Region.

cs.CL↗

arXiv2026

On the modulus of solutions of a first order differential equation

Let $P(z)=z^{n}+a_{n-2}z^{n-2}+\cdots+a_0$ be a nonconstant polynomial and $S(z)$ be a nonzero rational function and denote $h(z)=S(z)e^{P(z)}$. Let $θ\in(0,π/2n)$ be a constant and $\varepsilon>0$ be a small constant. It is shown that if $f(z)$ is a solution of the first order differential equation $f'(z)=h(z)f(z)+1$, then there is a sequence $\{r_{k}\}$ such that the set $E=\cup_{l=0}^{\infty}[r_{2l},r_{2l+1}]$ has infinite logarithmic measure and for all $r\in E$, \begin{equation}\tag† \begin{split} |f(re^{iθ})|\geq (1-\varepsilon)\frac{\sqrt[n]{\sin nθ}}{n}r\exp\left(e^{(1-\varepsilon)r^n\cos nθ}\sin\varepsilon\right). \end{split} \end{equation} When $h(z)=e^{z}$, we also give a lower bound for $|f(re^{iθ})|$ for other values of $r$. The estimate in $(†)$ yields that the hyper-order $ς(f)$ of $f(z)$ is equal to $n$, giving a partial answer to Brück's conjecture in uniqueness theory of meromorphic functions. An extension of the method also yields a complete description on the order of growth of entire solutions of a second order algebraic differential equation of Hayman in the autonomous case.

math.CV↗

arXiv2025

Formal O(N3) scaling GW calculations by block tensor decomposition for large molecule systems

Within the framework of many-body perturbation theory based on Green's functions, the $GW$ approximation has emerged as a pivotal method for computing quasiparticle energies and excitation spectra. However, its high computational cost and steep scaling present significant challenges for applications to large molecular systems. In this work, we extend the block tensor decomposition (BTD) algorithm, recently developed in our previous work [J. Chem. Phys. 163, 174109 (2025)] for low-rank tensor compression, to enable a formally $O(N^3)$-scaling $GW$ algorithm. By integrating BTD with an imaginary-time $GW$ formalism and introducing a real space screening strategy for the polarizability, we achieve an observed scaling of approximately $O(N^2)$ in test systems. Key parameters of the algorithm are optimized on the S66 dataset using the JADE algorithm, ensuring a balanced compromise between accuracy and efficiency. Our BTD-based random phase approximation also exhibits $O(N^2)$ scaling, and eigenvalue-self-consistent $GW$ calculations become feasible for systems with over 3000 basis functions. This work establishes BTD as an efficient and scalable approach for large-scale $GW$ calculations in molecular systems.

physics.chem-ph↗

arXiv2025

On transcendental meromorphic solutions of Hayman's equation

We present a complete description of the form of transcendental meromorphic solutions of the second order differential equation \begin{equation}\tag† w''w-w'^2+a w'w+b w^2=αw+βw'+γ, \end{equation} where $a$, $b$, $α$, $β$ and $γ$ are all rational functions. Together with the Wiman--Valiron theory, we then show that any transcendental meromorphic solution $w$ of equation $(†)$ has hyper-order $ς(w)\leq n$ for some integer $n\geq 0$. Moreover, if $w$ has finite order $σ(w)$, then $2σ(w)$ is a positive integer; if $β\equivγ\equiv0$ and $w$ has infinite order or if $γ\not\equiv0$ and $w$ has infinite order, then the hyper-order $ς(w)$ is a positive integer.

math.CV↗

arXiv2025

Block Tensor Decomposition: A dual grid scheme with formal O(N3) for THC decomposition of molecular systems

Accurate and fast treatment of electron-electron interactions remains a central challenge in electronic structure theory because post-Hartree-Fock methods often suffered from the computational cost for 4-index electron repulsion integrals (ERIs). Low-rank approaches such as tensor hyper-contraction (THC) and interpolative separable density fitting (ISDF) have been proposed for Hartree-Fock exchange and correlation's calculations. Their application to molecular systems remains inefficient due to the construction of THC kernel whose time scale increases as quartic with the number of basis functions. In this work, we present an algorithm named block tensor decomposition (BTD) based on a dual grid scheme that combines Hilbert sort and pivoted Cholesky decomposition to generate compact interpolative grids, allowing strict $O(N^3)$ scaling for THC/ISDF kernel construction. The key parameters in BTD are optimized via differential evolution, balancing efficiency and accuracy. Furthermore, we apply BTD in scaled opposite-spin MP2 (SOS-MP2), leveraging sparse mapping in real space to achieve quadratic scaling for electron correlation calculation and linear scaling for exchange calculation. This work advances low-scaling THC/ISDF methodologies for molecular systems, offering a robust framework for efficient and accurate electronic structure computations.

physics.chem-ph↗

arXiv2025

Bank-Laine functions with preassigned number of zeros

A Bank--Laine function $E$ is written as $E=f_1f_2$ for two normalized solutions $f_1$ and $f_2$ of the second order differential equation $f''+Af=0$, where $A$ is an entire function. In this paper, we first complete the construction of Bank--Laine functions by Bergweiler and Eremenko. Then, letting $n\in \mathbb{N}$ be a positive integer, we show the existence of entire functions $A$ for which the associated Bank--Laine functions $E=f_1f_2$ have preassigned exponent of convergence of number of zeros $λ(E)$ of three types: (1) for every two numbers $λ_1,λ_2\in[0,n]$ such that $λ_1\leq λ_2$, there exists an entire function $A$ of order $ρ(A)=n$ such that $E=f_1f_2$ satisfies $λ(f_1)=λ_1$, $λ(f_2)=λ_2$ and $λ(E)=λ_2\leq ρ(E)=n$; (2) for every number $ρ\in(n/2,n)$ and $λ\in[0,\infty)$, there exists an entire function $A$ of order $ρ(A)=ρ$ such that $E=f_1f_2$ satisfies $λ(f_1)=λ$, $λ(f_2)=\infty$ and, moreover, $E_c=f_1(cf_1+f_2)$ satisfies $λ(E_c)=\infty$ for any constant $c$; (3) for every number $λ\in[0,n]$, there exists an entire function $A$ of order $ρ(A)=n$ such that $E=f_1f_2$ satisfies $λ(f_1)=λ$, $λ(f_2)=\infty$ and, moreover, $E_c=f_1(cf_1+f_2)$ satisfies $λ(E_c)=\infty$ for any constant $c$. The construction for the three types of Bank--Laine functions requires new developments of the method of quasiconformal surgery by Bergweiler and Eremenko.

math.CV↗

arXiv2024

PZT Optical Memristors

Optical memristors represent a monumental leap in the fusion of photonics and electronics, heralding a new era of applications from neuromorphic computing to artificial intelligence. However, current technologies are hindered by complex fabrication, limited endurance, high optical loss or low modulation depth. For the first time, we reveal optical non-volatility in thin-film Lead Zirconate Titanate (PZT) by electrically manipulating the ferroelectric domains to control the refractive index, providing a brand-new routine for optical memristors. The developed PZT optical memristors offer unprecedented advantages more than exceptional performance metrics like low loss of <2 dB/cm, high precision exceeding 6-bits, large modulation depth with an index change as large as 4.6x10-3. Additionally, these devices offer impressive stability, maintaining minimal wavelength variation for over three weeks and enduring more than 10,000 cycles, and require a mere 0.8 pJ of energy for non-volatile operation. The wafer-scale sol-gel fabrication process also ensures compatible with standardized mass fabrication processes and high scalability for photonic integration. Specially, these devices also demonstrate unique functional duality: setting above a threshold voltage enables non-volatile behaviors, below this threshold allows volatile high-speed optical modulation. This marks the first-ever optical memristor capable of performing high-speed (48 Gbps) and energy-efficient (450 fJ/bit) signal processing and non-volatile retention on a single platform, and is also the inaugural demonstration of scalable functional systems. The PZT optical memristors developed here facilitate the realization of novel paradigms for high-speed and energy-efficient optical interconnects, programmable PICs, quantum computing, neural networks, in-memory computing and brain-like architecture.

physics.optics↗

arXiv2024

GenCRF: Generative Clustering and Reformulation Framework for Enhanced Intent-Driven Information Retrieval

Query reformulation is a well-known problem in Information Retrieval (IR) aimed at enhancing single search successful completion rate by automatically modifying user's input query. Recent methods leverage Large Language Models (LLMs) to improve query reformulation, but often generate limited and redundant expansions, potentially constraining their effectiveness in capturing diverse intents. In this paper, we propose GenCRF: a Generative Clustering and Reformulation Framework to capture diverse intentions adaptively based on multiple differentiated, well-generated queries in the retrieval phase for the first time. GenCRF leverages LLMs to generate variable queries from the initial query using customized prompts, then clusters them into groups to distinctly represent diverse intents. Furthermore, the framework explores to combine diverse intents query with innovative weighted aggregation strategies to optimize retrieval performance and crucially integrates a novel Query Evaluation Rewarding Model (QERM) to refine the process through feedback loops. Empirical experiments on the BEIR benchmark demonstrate that GenCRF achieves state-of-the-art performance, surpassing previous query reformulation SOTAs by up to 12% on nDCG@10. These techniques can be adapted to various LLMs, significantly boosting retriever performance and advancing the field of Information Retrieval.

cs.IR↗

arXiv2024

Zero order meromorphic solutions of $q$-difference equations of Malmquist type

We consider the first order $q$-difference equation \begin{equation}\tag† f(qz)^n=R(z,f), \end{equation} where $q\not=0,1$ is a constant and $R(z,f)$ is rational in both arguments. When $|q|\not=1$, we show that, if $(†)$ has a zero order transcendental meromorphic solution, then $(†)$ reduces to a $q$-difference linear or Riccati equation, or to an equation that can be transformed to a $q$-difference Riccati equation. In the autonomous case, explicit meromorphic solutions of $(†)$ are presented. Given that $(†)$ can be transformed into a difference equation, we proceed to discuss the growth of the composite function $f(ω(z))$, where $ω(z)$ is an entire function satisfying $ω(z+1)=qω(z)$, and demonstrate how the proposed difference Painlevé property, as discussed in the literature, applies for $q$-difference equations.

math.CV↗

arXiv2024

Zero distribution of finite order Bank--Laine functions

It is known that a Bank-Laine function $E$ is a product of two normalized solutions of the second order differential equation $f"+Af=0$ $(\dagger)$, where $A=A(z)$ is an entire function. By using Bergweiler and Eremenko's method of constructing transcendental entire function $A(z)$ by gluing certain meromorphic functions with infinitely many times, we show that, for each $λ\in[1,\infty)$ and each $δ\in[0,1]$, there exists a Bank--Laine function $E$ such that $E=f_1f_2$ with $f_1$ and $f_2$ being two entire functions such that $λ(f_1)=δλ$ and $λ(f_2)=λ$, respectively. We actually provide a simpler construction of the special Bank--Laine functions given by Bergweiler and Eremenko.

math.CV↗