arXiv subjects

Tao Liang

Publications and source records attributed to Tao Liang.

At least 19 recordsLinked to original sources

arXiv2026

Linear Growth and Nonlinear Stability of Two-Dimensional MHD Couette Flow with Vertical Dissipation

We study the two-dimensional incompressible magnetohydrodynamic system near the Couette equilibrium $\bigl((y,0)^{\rm T},(\beta,0)^{\rm T}\bigr)$ on $\mathbb T\times\mathbb R$, in the anisotropic regime where both viscosity and magnetic diffusivity act only in the vertical direction. For the inviscid linearized problem with $|\beta|>1/2$, we prove sharp linear-in-time growth of the vorticity and current density at the level of the time rate. In contrast, the horizontal components of the velocity and magnetic perturbations remain uniformly bounded, while the vertical components exhibit quantitative inviscid damping at the rate $\langle t\rangle^{-1}$. For the nonlinear problem, we introduce shear-adapted Fourier multipliers that simultaneously capture enhanced dissipation, critical-time effects, and echo-type resonant interactions. Under a suitable horizontal background magnetic field and a quantitative compatibility condition between the magnetic field strength, viscosity, and magnetic diffusivity, we establish global nonlinear stability for divergence-free perturbations satisfying $ \left\| \bigl( \mathbf v_{\rm in}-(y,0)^{\rm T}, \mathbf H_{\rm in}-(\beta,0)^{\rm T} \bigr) \right\|_{H^N} \leq \varepsilon_0\min\{\mu,\nu\}^{1/2}, N\geq4$. Moreover, the nonzero horizontal Fourier modes decay in a shear-adapted $H^N$ norm at the enhanced-dissipation rate $e^{-c\min\{\mu,\nu\}^{1/3}\,t},$ and the vertical velocity and magnetic components gain an additional inviscid-damping factor $\langle t\rangle^{-1}$.

math.AP↗

arXiv2026

Electron-beam Writing of Spectrally Uniform Green Single-photon Emitters in Hexagonal Boron Nitride

Scalable quantum photonic technologies require single-photon emitters whose positions and emission energies can be engineered simultaneously. Hexagonal boron nitride (hBN) is an attractive room-temperature host, but deterministic creation of spectrally reproducible emitters remains challenging. Here, we use a standard scanning electron microscope as a direct-writing tool to activate bright green single-photon emitters in hBN at predefined sites, without ion implantation or post-fabrication thermal annealing. The written emitters exhibit reproducible zero-phonon-line emission centered near 536 nm, room-temperature antibunching with g(2)(0) as low as 0.08, high brightness, strong linear polarization, and stable emission. Thickness-dependent activation, stacking experiments, cathodoluminescence spectroscopy, and first-principles calculations support a carbon-related defect complex as the most plausible origin of the emission. As a proof of nanophotonic compatibility, we further activate emitters in a nanoparticle-on-mirror plasmonic nanocavity and observe photoluminescence enhancement accompanied by shortened emission lifetimes. These results establish electron-beam direct writing as a practical route to site-selective, spectrally uniform green quantum emitters in hBN, offering a promising basis for integrated room-temperature quantum photonic architectures.

physics.optics↗

arXiv2026

Chain conformations in adsorbed layer during polymer capillary imbibition

We conducted molecular dynamics simulations to investigate chain conformations in adsorbed layers during polymer capillary imbibition. While the imbibition length adheres to the classical Lucas-Washburn equation, a notable deviation in mobile bead density emerges under strong confinement, consistent with \emph{in situ} dielectric spectroscopy experiments. The proportion of loop structures within adsorbed layers progressively increases during capillary infiltration, attributed to the relaxation of initially stretched chains toward equilibrium configurations. Furthermore, systematic analysis revealed that chain relaxation dynamics exhibit length-dependent retardation, especially under high confinement. The characteristic desorption time demonstrates chain-length dependence in quantitative agreement with scaling predictions.

cond-mat.soft↗

arXiv2026

Composition design of refractory compositionally complex alloys using machine learning models

Refractory compositionally complex alloys (RCCAs) are considered the next generation high-temperature materials. However, their high-dimensional composition spaces are too large to explore by traditional density functional theory or experimental means, making new RCCA discovery slow and cumbersome. This work has addressed these challenges with an integrated composition design framework that can efficiently and exhaustively explore the relationship between the compositions and two fundamental aspects: 1) the phase stability, including the target body-centered cubic (BCC) phase and its competing phases (hexagonal closed-pack (HCP) structures, Laves and B2 intermetallic phases), and 2) the mechanical properties. This framework is demonstrated with RCCAs within nine refractory metals (Ti, V, Cr, Zr, Nb, Mo, Hf, Ta, and W). Theory-guided machine learning (ML) models were employed to find the composition-mechanical property relationship of RCCAs, where the established theory is used to supplement the yield strength data at ultra-high temperature, and a forward sequential feature selection (SFS) is used to determine feature selection. The resulting ML model for temperature-dependent yield strength was found to have an R_squared value of 0.98 over the entire temperature range (from 0 to 2000 K). The impact of each constituent element on the six key properties is evaluated. The addition of Nb tends to stabilize the BCC phase and the addition of Ti improves the ductility of RCCAs. Combined with all methods involved in this framework, the on-demand designer allows the alloy designers to have all properties for any RCCA compositions and narrow down the composition space by applying custom screening criteria. The output from the predictor and screener provides valuable guidance for our experimental study of RCCAs and accelerates the pace of materials discovery.

cond-mat.mtrl-sci↗

arXiv2026

Premier: Personalized Preference Modulation with Learnable User Embedding in Text-to-Image Generation

Text-to-image generation has advanced rapidly, yet it still struggles to capture the nuanced user preferences. Existing approaches typically rely on multimodal large language models to infer user preferences, but the derived prompts or latent codes rarely reflect them faithfully, leading to suboptimal personalization. We present Premier, a novel preference modulation framework for personalized image generation. Premier represents each user's preference as a learnable embedding and introduces a preference adapter that fuses the user embedding with the text prompt. To enable accurate and fine-grained preference control, the fused preference embedding is further used to modulate the generative process. To enhance the distinctness of individual preference and improve alignment between outputs and user-specific styles, we incorporate a dispersion loss that enforces separation among user embeddings. When user data are scarce, new users are represented as linear combinations of existing preference embeddings learned during training, enabling effective generalization. Experiments show that Premier outperforms prior methods under the same history length, achieving stronger preference alignment and superior performance on text consistency, ViPer proxy metrics, and expert evaluations.

cs.CV↗

arXiv2026

LongCat-Flash-Thinking-2601 Technical Report

We introduce LongCat-Flash-Thinking-2601, a 560-billion-parameter open-source Mixture-of-Experts (MoE) reasoning model with superior agentic reasoning capability. LongCat-Flash-Thinking-2601 achieves state-of-the-art performance among open-source models on a wide range of agentic benchmarks, including agentic search, agentic tool use, and tool-integrated reasoning. Beyond benchmark performance, the model demonstrates strong generalization to complex tool interactions and robust behavior under noisy real-world environments. Its advanced capability stems from a unified training framework that combines domain-parallel expert training with subsequent fusion, together with an end-to-end co-design of data construction, environments, algorithms, and infrastructure spanning from pre-training to post-training. In particular, the model's strong generalization capability in complex tool-use are driven by our in-depth exploration of environment scaling and principled task construction. To optimize long-tailed, skewed generation and multi-turn agentic interactions, and to enable stable training across over 10,000 environments spanning more than 20 domains, we systematically extend our asynchronous reinforcement learning framework, DORA, for stable and efficient large-scale multi-environment training. Furthermore, recognizing that real-world tasks are inherently noisy, we conduct a systematic analysis and decomposition of real-world noise patterns, and design targeted training procedures to explicitly incorporate such imperfections into the training process, resulting in improved robustness for real-world applications. To further enhance performance on complex reasoning tasks, we introduce a Heavy Thinking mode that enables effective test-time scaling by jointly expanding reasoning depth and width through intensive parallel thinking.

cs.AI↗

arXiv2025

Stability threshold of the 2D Boussinesq system near Couette flow in an infinite channel

In this paper, we study the stability threshold of the two-dimensional Boussinesq equations around the Couette flow in an infinite channel $\mathbb{R} \times [-1, 1]$ under no-slip boundary conditions. We prove that the Couette flow is asymptotically stable under initial perturbations satisfying $\| \mathbf{v}^{\mathrm{in}} -(y,0)\|_{H^2} \le \varepsilon_0 \nu^{\frac12}$, and $\| \rho^{\mathrm{in}}-1 \|_{H^1} + \big\| |\partial_x|^{\frac13} \rho^{\mathrm{in}} \big\|_{H^1} \le \varepsilon_1 \nu^{\frac56}$. Compared with the work of Masmoudi, Zhai, and Zhao [J. Funct. Anal., 284 (2023), 109736], where the asymptotic stability of the 2D Navier-Stokes-Boussinesq system around Couette flow in a finite channel $\mathbb{T} \times [-1, 1]$ was established, our result improves the stability threshold for the temperature from $\nu^{\frac{11}{12}}$ to $\nu^{\frac56}$.

math.AP↗

arXiv2025

PreferThinker: Reasoning-based Personalized Image Preference Assessment

Personalized image preference assessment aims to evaluate an individual user's image preferences by relying only on a small set of reference images as prior information. Existing methods mainly focus on general preference assessment, training models with large-scale data to tackle well-defined tasks such as text-image alignment. However, these approaches struggle to handle personalized preference because user-specific data are scarce and not easily scalable, and individual tastes are often diverse and complex. To overcome these challenges, we introduce a common preference profile that serves as a bridge across users, allowing large-scale user data to be leveraged for training profile prediction and capturing complex personalized preferences. Building on this idea, we propose a reasoning-based personalized image preference assessment framework that follows a \textit{predict-then-assess} paradigm: it first predicts a user's preference profile from reference images, and then provides interpretable, multi-dimensional scores and assessments of candidate images based on the predicted profile. To support this, we first construct a large-scale Chain-of-Thought (CoT)-style personalized assessment dataset annotated with diverse user preference profiles and high-quality CoT-style reasoning, enabling explicit supervision of structured reasoning. Next, we adopt a two-stage training strategy: a cold-start supervised fine-tuning phase to empower the model with structured reasoning capabilities, followed by reinforcement learning to incentivize the model to explore more reasonable assessment paths and enhance generalization. Furthermore, we propose a similarity-aware prediction reward to encourage better prediction of the user's preference profile, which facilitates more reasonable assessments exploration. Extensive experiments demonstrate the superiority of the proposed method.

cs.AI↗

arXiv2025

Introducing LongCat-Flash-Thinking: A Technical Report

We present LongCat-Flash-Thinking, an efficient 560-billion-parameter open-source Mixture-of-Experts (MoE) reasoning model. Its advanced capabilities are cultivated through a meticulously crafted training process, beginning with long Chain-of-Thought (CoT) data cold-start and culminating in large-scale Reinforcement Learning (RL). We first employ a well-designed cold-start training strategy, which significantly enhances the reasoning potential and equips the model with specialized skills in both formal and agentic reasoning. Then, a core innovation is our domain-parallel training scheme, which decouples optimization across distinct domains (e.g., STEM, Code, Agentic) and subsequently fuses the resulting expert models into a single, nearly Pareto-optimal model. This entire process is powered by our Dynamic ORchestration for Asynchronous rollout (DORA) system, a large-scale RL framework that delivers a greater than threefold training speedup over synchronous methods on tens of thousands of accelerators. As a result, LongCat-Flash-Thinking achieves state-of-the-art performance among open-source models on a suite of complex reasoning tasks. The model exhibits exceptional efficiency in agentic reasoning, reducing average token consumption by 64.5% (from 19, 653 to 6, 965) on AIME-25, without degrading task accuracy. We release LongCat-Flash-Thinking to promote further advances in reasoning systems and agentic AI research.

cs.AI↗

arXiv2025

DiffHash: Text-Guided Targeted Attack via Diffusion Models against Deep Hashing Image Retrieval

Deep hashing models have been widely adopted to tackle the challenges of large-scale image retrieval. However, these approaches face serious security risks due to their vulnerability to adversarial examples. Despite the increasing exploration of targeted attacks on deep hashing models, existing approaches still suffer from a lack of multimodal guidance, reliance on labeling information and dependence on pixel-level operations for attacks. To address these limitations, we proposed DiffHash, a novel diffusion-based targeted attack for deep hashing. Unlike traditional pixel-based attacks that directly modify specific pixels and lack multimodal guidance, our approach focuses on optimizing the latent representations of images, guided by text information generated by a Large Language Model (LLM) for the target image. Furthermore, we designed a multi-space hash alignment network to align the high-dimension image space and text space to the low-dimension binary hash space. During reconstruction, we also incorporated text-guided attention mechanisms to refine adversarial examples, ensuring them aligned with the target semantics while maintaining visual plausibility. Extensive experiments have demonstrated that our method outperforms state-of-the-art (SOTA) targeted attack methods, achieving better black-box transferability and offering more excellent stability across datasets.

cs.IR↗

arXiv2025

Improved stability threshold for 2D Navier-Stokes Couette flow in an infinite channel

We study the nonlinear stability of the two-dimensional Navier-Stokes equations around the Couette shear flow in the channel domain $\mathbb{R}\times[-1,1]$ subject to Navier slip boundary conditions. We establish a quantitative stability threshold for perturbations of the initial vorticity $\omega_{in}$, showing that stability holds for perturbations of order $\nu^{1/2}$ measured in an anisotropic Sobolev space. This sharpens the recent work of Arbon and Bedrossian [Comm. Math. Phys., 406 (2025), Paper No. 129] who proved stability under the threshold $\nu^{1/2}(1+\ln(1/\nu))^{-1/2}$. Our result removes the logarithmic loss and identifies the natural scaling $\nu^{1/2}$ as the critical size of perturbations for nonlinear stability in this setting.

math.AP↗

arXiv2025

Enhancing Reward Models for High-quality Image Generation: Beyond Text-Image Alignment

Contemporary image generation systems have achieved high fidelity and superior aesthetic quality beyond basic text-image alignment. However, existing evaluation frameworks have failed to evolve in parallel. This study reveals that human preference reward models fine-tuned based on CLIP and BLIP architectures have inherent flaws: they inappropriately assign low scores to images with rich details and high aesthetic value, creating a significant discrepancy with actual human aesthetic preferences. To address this issue, we design a novel evaluation score, ICT (Image-Contained-Text) score, that achieves and surpasses the objectives of text-image alignment by assessing the degree to which images represent textual content. Building upon this foundation, we further train an HP (High-Preference) score model using solely the image modality to enhance image aesthetics and detail quality while maintaining text-image alignment. Experiments demonstrate that the proposed evaluation model improves scoring accuracy by over 10\% compared to existing methods, and achieves significant results in optimizing state-of-the-art text-to-image models. This research provides theoretical and empirical support for evolving image generation technology toward higher-order human aesthetic preferences. Code is available at https://github.com/BarretBa/ICTHP.

cs.CV↗

arXiv2025

Large global solutions to the Oldroyd-B model with dissipation

In the first part of this work, we investigate the Cauchy problem for the $d$-dimensional incompressible Oldroyd-B model with dissipation in the stress tensor equation. By developing a weighted Chemin-Lerner framework combined with a refined energy argument, we prove the existence and uniqueness of global solutions for the system under a mild constraint on the initial velocity field, while allowing a broad class of large initial data for the stress tensor. Notably, our analysis accommodates general divergence-free initial stress tensors ( $\mathrm{div}\tau_0=0$) and significantly relaxes the requirements on initial velocities compared to classical fluid models. This stands in sharp contrast to the finite-time singularity formation observed in the incompressible Euler equations, even for small initial data, thereby highlighting the intrinsic stabilizing role of the stress tensor in polymeric fluid dynamics. The second part of this paper focuses on the small-data regime. Through a systematic exploitation of the perturbative structure of the system, we establish global well-posedness and quantify the long-time behavior of solutions in Sobolev spaces $H^3(\mathbb{T}^d)$. Specifically, we derive exponential decay rates for perturbations, demonstrating how the dissipative mechanisms inherent to the Oldroyd-B model govern the asymptotic stability of the system.

math.AP↗

arXiv2025

Stability threshold for two-dimensional Boussinesq systems near-Couette shear flow in a finite channel

In this paper, we investigate the stability threshold problem of the two-dimensional Navier-Stokes Boussinesq(NSB) equations in a finite channel $ \T \times [-1,1]$, focusing on the stability around the near Couette shear flow $ (U(y), 0)$, assuming the Navier slip boundary conditions are satisfied. In particular, when the initial data for the vorticity resides in an anisotropic Sobolev space of size $ O(\min \{ \mu^{\frac{1}{2}}, \nu^{\frac{1}{2}}\})$, and the initial perturbation of the temperature resides in an anisotropic Sobolev space of size $ O(\min \{ \mu, \nu\})$, we derive the nonlinear enhanced dissipation effect and the inviscid damping effect for the NSB system.

math.AP↗

arXiv2025

Route Sparse Autoencoder to Interpret Large Language Models

Mechanistic interpretability of large language models (LLMs) aims to uncover the internal processes of information propagation and reasoning. Sparse autoencoders (SAEs) have demonstrated promise in this domain by extracting interpretable and monosemantic features. However, prior works primarily focus on feature extraction from a single layer, failing to effectively capture activations that span multiple layers. In this paper, we introduce Route Sparse Autoencoder (RouteSAE), a new framework that integrates a routing mechanism with a shared SAE to efficiently extract features from multiple layers. It dynamically assigns weights to activations from different layers, incurring minimal parameter overhead while achieving high interpretability and flexibility for targeted feature manipulation. We evaluate RouteSAE through extensive experiments on Llama-3.2-1B-Instruct. Specifically, under the same sparsity constraint of 64, RouteSAE extracts 22.5% more features than baseline SAEs while achieving a 22.3% higher interpretability score. These results underscore the potential of RouteSAE as a scalable and effective method for LLM interpretability, with applications in feature discovery and model intervention. Our codes are available at https://github.com/swei2001/RouteSAEs.

cs.LG↗

arXiv2025

PFDial: A Structured Dialogue Instruction Fine-tuning Method Based on UML Flowcharts

Process-driven dialogue systems, which operate under strict predefined process constraints, are essential in customer service and equipment maintenance scenarios. Although Large Language Models (LLMs) have shown remarkable progress in dialogue and reasoning, they still struggle to solve these strictly constrained dialogue tasks. To address this challenge, we construct Process Flow Dialogue (PFDial) dataset, which contains 12,705 high-quality Chinese dialogue instructions derived from 440 flowcharts containing 5,055 process nodes. Based on PlantUML specification, each UML flowchart is converted into atomic dialogue units i.e., structured five-tuples. Experimental results demonstrate that a 7B model trained with merely 800 samples, and a 0.5B model trained on total data both can surpass 90% accuracy. Additionally, the 8B model can surpass GPT-4o up to 43.88% with an average of 11.00%. We further evaluate models' performance on challenging backward transitions in process flows and conduct an in-depth analysis of various dataset formats to reveal their impact on model performance in handling decision and sequential branches. The data is released in https://github.com/KongLongGeFDU/PFDial.

cs.CL↗

arXiv2024

Neuron-Level Sequential Editing for Large Language Models

This work explores sequential model editing in large language models (LLMs), a critical task that involves modifying internal knowledge within LLMs continuously through multi-round editing, each incorporating updates or corrections to adjust the model outputs without the need for costly retraining. Existing model editing methods, especially those that alter model parameters, typically focus on single-round editing and often face significant challenges in sequential model editing-most notably issues of model forgetting and failure. To address these challenges, we introduce a new model editing method, namely \textbf{N}euron-level \textbf{S}equential \textbf{E}diting (NSE), tailored for supporting sequential model editing. Specifically, we optimize the target layer's hidden states using the model's original weights to prevent model failure. Furthermore, we iteratively select neurons in multiple layers for editing based on their activation values to mitigate model forgetting. Our empirical experiments demonstrate that NSE significantly outperforms current modifying parameters model editing methods, marking a substantial advancement in the field of sequential model editing. Our code is released on \url{https://github.com/jianghoucheng/NSE}.

cs.CL↗

arXiv2024

Global well-posedness and large time behavior of solutions to the compressible Oldroyd-B model without stress diffusion

We consider the Cauchy problem ($\mathbb{R}^d, d=2,3$) and the initial boundary values problem ($\mathbb{T}^d, d=2,3$)associated to the compressible Oldroyd-B model which is first derived by Barrett, Lu and S\"{u}li [Existence of large-data finite-energy global weak solutions to a compressible Oldroyd-B model, Commun. Math. Sci., 15 (2017), 1265--1323] through micro-macro-analysis of the compressible Navier-Stokes-Fokker-Planck system.Due to lack of stress diffusion, the problems considered here are very difficult. Exploiting tools from harmonic analysis,notably the Littlewood Paley theory,we first establish the global well-posedness and time-decay rates for solutions of the model with small initial data in Besov spaces with critical regularity.Then, through deeply exploring and fully utilizing the structure of the perturbation system,we obtain the global well-posedness and exponential decay rates for solutions of the model with small initial data in the Soboles spaces $H^3(\mathbb{T}^d)$.Our obtained results improve considerably the recent results by Lu, Pokorn\'{y} [Anal. Theory Appl., 36 (2020), 348--372],Wang, Wen [Math. Models Methods Appl. Sci., 30 (2020), 139--179],and Liu, Lu, Wen [SIAM J. Math. Anal., 53 (2021), 6216--6242].

math.AP↗