arXiv subjects

Donghao Li

Publications and source records attributed to Donghao Li.

At least 19 recordsLinked to original sources

arXiv2026

DeepSeek-V4.1-Flash: Pushing the Limits of KV Cache Compression

The widespread adoption of long-horizon agents has made model workloads increasingly input-heavy. Although prior work has substantially reduced the cost of long-context computation, prefill remains computationally expensive, and large KV caches continue to strain HBM and SSD capacity and data-transfer bandwidth. Together, these compute, storage, and bandwidth demands constitute the primary bottleneck to further lowering deployment costs. To address this challenge, we introduce DeepSeek-V4.1-Flash, a multimodal Mixture-of-Experts (MoE) model with 552B backbone parameters and support for contexts of up to one million tokens. With its Causal Encoder-Decoder (CED) architecture, the model activates 16B parameters per token during decode but only 8B parameters during prefill, substantially improving cost efficiency for agentic workloads. To push the limits of KV cache compression, DeepSeek-V4.1-Flash combines cross-layer KV cache reuse in Compressed Sparse Attention 2 (CSA2) with FP4 KV caching. These designs reduce its global KV cache footprint (always in HBM) to 890 bytes per token, roughly 1/4 of the corresponding footprint of DeepSeek-V4-Flash. Further, through a dedicated deployment optimization known as SWA Bounded Replay, DeepSeek-V4.1-Flash reduces its persistent KV cache footprint (always on SSD or in host memory) to roughly 1/8 of that of DeepSeek-V4-Flash. Despite its much smaller KV cache footprint, the model delivers substantially better performance than the baseline. In addition, we streamline the DeepSeek-V4 architecture and introduce several efficient architectural extensions. We pretrain DeepSeek-V4.1-Flash on a multimodal corpus comprising 45T tokens and conduct comprehensive post-training, yielding strong performance across diverse text-based and multimodal agentic scenarios. Model checkpoints are available at https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash.

cs.CL↗

arXiv2026

Study on Thickness and Temperature Dependence of Thermoelectric Properties in SnS Nanofilms

SnS as an environmentally friendly, cost-effective, and earth-abundant narrow-bandgap semiconductor material, has demonstrated significant application potential in the field of medium-temperature thermoelectric conversion. However, the thermoelectric performance of its bulk counterpart is inherently constrained by intrinsic point defects (e.g., vacancies) and the material's specific band structure. Low-dimensional engineering has emerged as a pivotal strategy for overcoming these limitations and enhancing thermoelectric performance. In this work, we systematically investigate the thermoelectric properties of SnS nanofilms with distinct thicknesses (82 nm, 199 nm, 616 nm, and 813 nm) across a temperature range of 300-600 K. Measurements were conducted using time-domain thermoreflectance (TDTR) and a dedicated thin-film thermoelectric parameter test system (ZEM-3). Our results confirm that low-dimensionalization effectively boosts the thermoelectric performance of SnS, with the thermoelectric figure of merit (ZT) displaying a pronounced dependence on both film thickness and temperature. All four SnS thin films exhibit thermoelectric performance that is markedly superior to that of bulk SnS. This enhancement is primarily attributed to the quantum confinement effect, energy filtering effect, and intensified phonon scattering, all of which are induced by the low-dimensional structural characteristics. This work provides not only experimental evidence and theoretical insights for the performance optimization of SnS nanofilms but also establishes a foundational framework for the development of high-efficiency, eco-friendly medium-temperature thermoelectric materials, thereby holding significant scientific value and practical implications.

cond-mat.mtrl-sci↗

arXiv2026

Descent and Brauer-Manin Obstructions on Deligne-Mumford Stacks

We generalize and compare local-global obstructions for algebraic stacks over number fields. For smooth separated Deligne-Mumford stacks of finite type with a quasi-projective coarse moduli space, we prove that the descent obstruction is contained in the Brauer--Manin obstruction. By lifting $\mathbb{G}_m$-gerbes, we obtain an inclusion between the corresponding composite obstructions. We also show that in this setting the descent obstruction coincides with both the étale Brauer-Manin obstruction and the iterated descent obstruction. The Brauer-Manin obstruction also coincides with the iterated Brauer-Manin obstruction.

math.AG↗

arXiv2026

Efficient Multi-objective Prompt Optimization via Pure-exploration Bandits

Prompt engineering has become central to eliciting the capabilities of large language models (LLMs). At its core lies prompt selection -- efficiently identifying the most effective prompts. However, most prior investigations overlook a key challenge: the inherently multi-faceted nature of prompt performance, which cannot be captured by a single metric. To fill this gap, we study the multi-objective prompt selection problem under two practical settings: Pareto prompt set recovery and best feasible prompt identification. Casting the problem into the pure-exploration bandits framework, we adapt provably efficient algorithms from multi-objective bandits and further introduce a novel design for best feasible arm identification in structured bandits, with theoretical guarantees on the identification error in the linear case. Extensive experiments across multiple LLMs show that the bandit-based approaches yield significant improvements over baselines, establishing a principled and efficient framework for multi-objective prompt optimization.

cs.LG↗

arXiv2026

Breaking the Computational Barrier: Provably Efficient Actor-Critic for Low-Rank MDPs

Reinforcement learning (RL) is a fundamental framework for sequential decision-making, in which an agent learns an optimal policy through interactions with an unknown environment. In settings with function approximation, many existing RL algorithms achieve favorable sample complexity, but often rely on computationally intractable oracles. In this paper, we use supervised learning as a computational proxy to establish a clear hierarchy of commonly adopted RL oracles under low-rank Markov Decision Processes (MDPs). This hierarchy shows that policy evaluation is the most computationally efficient oracle, provided that supervised learning can be efficiently solved. Motivated by this observation, we propose a novel optimistic actor-critic algorithm that relies solely on the policy evaluation oracle. We prove that our algorithm outperforms the existing sample complexity guarantees for low-rank MDPs while avoiding computationally expensive planning or optimization oracles commonly assumed in prior works. We further extend our theoretical results to approximately low-rank MDPs and demonstrate that this setting captures a broad class of real-world environments. Finally, we validate our theoretical results with experiments on several standard Gym environments.

cs.LG↗

arXiv2026

DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence

We present a preview version of DeepSeek-V4 series, including two strong Mixture-of-Experts (MoE) language models -- DeepSeek-V4-Pro with 1.6T parameters (49B activated) and DeepSeek-V4-Flash with 284B parameters (13B activated) -- both supporting a context length of one million tokens. DeepSeek-V4 series incorporate several key upgrades in architecture and optimization: (1) a hybrid attention architecture that combines Compressed Sparse Attention (CSA) and Heavily Compressed Attention (HCA) to improve long-context efficiency; (2) Manifold-Constrained Hyper-Connections (mHC) that enhance conventional residual connections; (3) and the Muon optimizer for faster convergence and greater training stability. We pre-train both models on more than 32T diverse and high-quality tokens, followed by a comprehensive post-training pipeline that unlocks and further enhances their capabilities. DeepSeek-V4-Pro-Max, the maximum reasoning effort mode of DeepSeek-V4-Pro, redefines the state-of-the-art for open models, outperforming its predecessors in core tasks. Meanwhile, DeepSeek-V4 series are highly efficient in long-context scenarios. In the one-million-token context setting, DeepSeek-V4-Pro requires only 27% of single-token inference FLOPs and 10% of KV cache compared with DeepSeek-V3.2. This enables us to routinely support one-million-token contexts, thereby making long-horizon tasks and further test-time scaling more feasible. The model checkpoints are available at https://huggingface.co/collections/deepseek-ai/deepseek-v4.

cs.CL↗

arXiv2026

SACRED: A Faithful Annotated Multimedia Multimodal Multilingual Dataset for Classifying Connectedness Types in Online Spirituality

In religion and theology studies, spirituality has garnered significant research attention for the reason that it not only transcends culture but offers unique experience to each individual. However, social scientists often rely on limited datasets, which are basically unavailable online. In this study, we collaborated with social scientists to develop a high-quality multimedia multi-modal datasets, \textbf{SACRED}, in which the faithfulness of classification is guaranteed. Using \textbf{SACRED}, we evaluated the performance of 13 popular LLMs as well as traditional rule-based and fine-tuned approaches. The result suggests DeepSeek-V3 model performs well in classifying such abstract concepts (i.e., 79.19\% accuracy in the Quora test set), and the GPT-4o-mini model surpassed the other models in the vision tasks (63.99\% F1 score). Purportedly, this is the first annotated multi-modal dataset from online spirituality communication. Our study also found a new type of connectedness which is valuable for communication science studies.

cs.CL↗

arXiv2025

AI-Accelerated Qubit Readout at the Single-Photon Level for Scalable Atomic Quantum Processors

Quantum state readout with minimal resources is crucial for scalable quantum information processing. As a leading platform, neutral atom arrays rely on atomic fluorescence imaging for qubit readout, requiring short exposure, low photon count schemes to mitigate heating and atom loss while enabling mid-circuit feedback. However, a fundamental challenge arises in the single-photon regime where severe overlap in state distributions causes conventional threshold discrimination to fail. Here, we report an AI-accelerated Bayesian inference method for fluorescence readout in neutral atom arrays. Our approach leverages Bayesian inference to achieve reliable state detection at the single-photon level under short exposure. Specifically, we introduce a weakly anchored Bayesian scheme that requires calibration of only one state, addressing asymmetric calibration challenges common across quantum platforms. Furthermore, acceleration is achieved via a permutation-invariant neural network, which yields a 100-fold speedup by compressing iterative inference into a single forward pass. The approach achieves relative readout fidelity above 99% and 98% for histogram overlaps of 61% and 72%, respectively, enabling reliable extraction of Rabi oscillations and Ramsey interference results unattainable with conventional threshold based methods. This framework supports scalable, real-time readout of large atom arrays and paves the way toward AI-enhanced quantum technology in computation and sensing.

quant-ph↗

arXiv2025

PromptEnhancer: A Simple Approach to Enhance Text-to-Image Models via Chain-of-Thought Prompt Rewriting

Recent advancements in text-to-image (T2I) diffusion models have demonstrated remarkable capabilities in generating high-fidelity images. However, these models often struggle to faithfully render complex user prompts, particularly in aspects like attribute binding, negation, and compositional relationships. This leads to a significant mismatch between user intent and the generated output. To address this challenge, we introduce PromptEnhancer, a novel and universal prompt rewriting framework that enhances any pretrained T2I model without requiring modifications to its weights. Unlike prior methods that rely on model-specific fine-tuning or implicit reward signals like image-reward scores, our framework decouples the rewriter from the generator. We achieve this by training a Chain-of-Thought (CoT) rewriter through reinforcement learning, guided by a dedicated reward model we term the AlignEvaluator. The AlignEvaluator is trained to provide explicit and fine-grained feedback based on a systematic taxonomy of 24 key points, which are derived from a comprehensive analysis of common T2I failure modes. By optimizing the CoT rewriter to maximize the reward from our AlignEvaluator, our framework learns to generate prompts that are more precisely interpreted by T2I models. Extensive experiments on the HunyuanImage 2.1 model demonstrate that PromptEnhancer significantly improves image-text alignment across a wide range of semantic and compositional challenges. Furthermore, we introduce a new, high-quality human preference benchmark to facilitate future research in this direction.

cs.CV↗

arXiv2025

Directly Aligning the Full Diffusion Trajectory with Fine-Grained Human Preference

Recent studies have demonstrated the effectiveness of directly aligning diffusion models with human preferences using differentiable reward. However, they exhibit two primary challenges: (1) they rely on multistep denoising with gradient computation for reward scoring, which is computationally expensive, thus restricting optimization to only a few diffusion steps; (2) they often need continuous offline adaptation of reward models in order to achieve desired aesthetic quality, such as photorealism or precise lighting effects. To address the limitation of multistep denoising, we propose Direct-Align, a method that predefines a noise prior to effectively recover original images from any time steps via interpolation, leveraging the equation that diffusion states are interpolations between noise and target images, which effectively avoids over-optimization in late timesteps. Furthermore, we introduce Semantic Relative Preference Optimization (SRPO), in which rewards are formulated as text-conditioned signals. This approach enables online adjustment of rewards in response to positive and negative prompt augmentation, thereby reducing the reliance on offline reward fine-tuning. By fine-tuning the FLUX model with optimized denoising and online reward adjustment, we improve its human-evaluated realism and aesthetic quality by over 3x.

cs.AI↗

arXiv2025

Augmenting Online RL with Offline Data is All You Need: A Unified Hybrid RL Algorithm Design and Analysis

This paper investigates a hybrid learning framework for reinforcement learning (RL) in which the agent can leverage both an offline dataset and online interactions to learn the optimal policy. We present a unified algorithm and analysis and show that augmenting confidence-based online RL algorithms with the offline dataset outperforms any pure online or offline algorithm alone and achieves state-of-the-art results under two learning metrics, i.e., sub-optimality gap and online learning regret. Specifically, we show that our algorithm achieves a sub-optimality gap $\tilde{O}(\sqrt{1/(N_0/\mathtt{C}(π^*|ρ)+N_1}) )$, where $\mathtt{C}(π^*|ρ)$ is a new concentrability coefficient, $N_0$ and $N_1$ are the numbers of offline and online samples, respectively. For regret minimization, we show that it achieves a constant $\tilde{O}( \sqrt{N_1/(N_0/\mathtt{C}(π^{-}|ρ)+N_1)} )$ speed-up compared to pure online learning, where $\mathtt{C}(π^-|ρ)$ is the concentrability coefficient over all sub-optimal policies. Our results also reveal an interesting separation on the desired coverage properties of the offline dataset for sub-optimality gap minimization and regret minimization. We further validate our theoretical findings in several experiments in special RL models such as linear contextual bandits and Markov decision processes (MDPs).

cs.LG↗

arXiv2025

Hunyuan-Game: Industrial-grade Intelligent Game Creation Model

Intelligent game creation represents a transformative advancement in game development, utilizing generative artificial intelligence to dynamically generate and enhance game content. Despite notable progress in generative models, the comprehensive synthesis of high-quality game assets, including both images and videos, remains a challenging frontier. To create high-fidelity game content that simultaneously aligns with player preferences and significantly boosts designer efficiency, we present Hunyuan-Game, an innovative project designed to revolutionize intelligent game production. Hunyuan-Game encompasses two primary branches: image generation and video generation. The image generation component is built upon a vast dataset comprising billions of game images, leading to the development of a group of customized image generation models tailored for game scenarios: (1) General Text-to-Image Generation. (2) Game Visual Effects Generation, involving text-to-effect and reference image-based game visual effect generation. (3) Transparent Image Generation for characters, scenes, and game visual effects. (4) Game Character Generation based on sketches, black-and-white images, and white models. The video generation component is built upon a comprehensive dataset of millions of game and anime videos, leading to the development of five core algorithmic models, each targeting critical pain points in game development and having robust adaptation to diverse game video scenarios: (1) Image-to-Video Generation. (2) 360 A/T Pose Avatar Video Synthesis. (3) Dynamic Illustration Generation. (4) Generative Video Super-Resolution. (5) Interactive Game Video Generation. These image and video generation models not only exhibit high-level aesthetic expression but also deeply integrate domain-specific knowledge, establishing a systematic understanding of diverse game and anime art styles.

cs.CV↗

arXiv2025

In vacuum metasurface for optical microtrap array

Optical tweezer arrays of laser-cooled and individual controlled particles have revolutionized the atomic, molecular and optical physics, and they afford exquisite capabilities for applications in quantum simulation of many-body physics, quantum computation and quantum sensing. Underlying this development is the technical maturity of generating scalable optical beams, enabled by active components and high numerical aperture objective. However, such a complex combination of bulk optics outside the vacuum chamber is very sensitive to any vibration and drift. Here we demonstrate the generation of 3*3 static tweezer array with a single chip-scale multifunctional metasurface element in vacuum, replacing the meter-long free space optics. Fluorescence counts on the camera validates the successfully trapping of the atomic ensemble array. Further, we discuss the strategy to achieve low scattering and crosstalk, where a metasurface design featuring dual-wavelength independent control is included. Our results, together with other recent development in integrated photonics for cold atoms, could pave the way for compact and portable quantum sensors and simulators in platforms of neutral atom arrays.

physics.optics↗

arXiv2025

Chirality-induced quantum nonreciprocity

Chirality, nonreciprocity, and quantum correlations are at the center of a wide range of intriguing effects and applications across natural sciences and emerging quantum technologies. However, the direct link combining these three essential concepts has remained unknown till now. Here, we establish a chiral non-Hermitian platform with flying atoms and demonstrate chirality-induced nonreciprocal bipartite quantum correlations between two channels: Quantum correlation emerges when two spatially separated light beams of the same polarization propagate in opposite directions in the atomic cloud, and it becomes zero when they travel in the same direction. Thus, just by flipping the propagation direction of one of the beams while keeping its polarization the same as the other beam, we can create or annihilate quantum correlations between two channels. We also show that this nonreciprocal quantum correlation can be extended to multi-color sidebands with Floquet engineering. Our findings may pave the road for realizing one-way quantum effects, such as nonreciprocal squeezing or entanglement, with a variety of chiral devices, for the emerging applications of e.g., directional quantum network or nonreciprocal quantum metrology.

quant-ph↗

arXiv2025

FedMIA: An Effective Membership Inference Attack Exploiting "All for One" Principle in Federated Learning

Federated Learning (FL) is a promising approach for training machine learning models on decentralized data while preserving privacy. However, privacy risks, particularly Membership Inference Attacks (MIAs), which aim to determine whether a specific data point belongs to a target client's training set, remain a significant concern. Existing methods for implementing MIAs in FL primarily analyze updates from the target client, focusing on metrics such as loss, gradient norm, and gradient difference. However, these methods fail to leverage updates from non-target clients, potentially underutilizing available information. In this paper, we first formulate a one-tailed likelihood-ratio hypothesis test based on the likelihood of updates from non-target clients. Building upon this formulation, we introduce a three-step Membership Inference Attack (MIA) method, called FedMIA, which follows the "all for one"--leveraging updates from all clients across multiple communication rounds to enhance MIA effectiveness. Both theoretical analysis and extensive experimental results demonstrate that FedMIA outperforms existing MIAs in both classification and generative tasks. Additionally, it can be integrated as an extension to existing methods and is robust against various defense strategies, Non-IID data, and different federated structures. Our code is available in https://github.com/Liar-Mask/FedMIA.

cs.LG↗

arXiv2025

A Shared Low-Rank Adaptation Approach to Personalized RLHF

Reinforcement Learning from Human Feedback (RLHF) has emerged as a pivotal technique for aligning artificial intelligence systems with human values, achieving remarkable success in fine-tuning large language models. However, existing RLHF frameworks often assume that human preferences are relatively homogeneous and can be captured by a single, unified reward model. This assumption overlooks the inherent diversity and heterogeneity across individuals, limiting the adaptability of RLHF to personalized scenarios and risking misalignments that can diminish user satisfaction and trust in AI systems. In this paper, we address these challenges by introducing Low-Rank Adaptation (LoRA) into the personalized RLHF framework. We apply LoRA in the the aggregated parameter space of all personalized reward functions, thereby enabling efficient learning of personalized reward models from potentially limited local datasets. Our approach exploits potential shared structures among the local ground-truth reward models while allowing for individual adaptation, without relying on restrictive assumptions about shared representations as in prior works. We further establish sample complexity guarantees for our method. Theoretical analysis demonstrates the effectiveness of the proposed approach in capturing both shared and individual-specific structures within heterogeneous human preferences, addressing the dual challenge of personalization requirements and practical data constraints. Experimental results on real-world datasets corroborate the efficiency of our algorithm in the personalized RLHF setting.

cs.LG↗

arXiv2024

Charm physics with overlap fermions on 2+1-flavor domain wall fermion configurations

Decay constants of pseudoscalar mesons $D$, $D_s$, $η_c$ and vector mesons $D^*$, $D_s^*$, $J/ψ$ are determined from $N_f=2+1$ lattice QCD at a lattice spacing $a\sim0.08$ fm. For vector mesons, the decay constants defined by tensor currents are given in the $\overline{\rm MS}$ scheme at $2$ GeV. The calculation is performed on domain wall fermion configurations generated by the RBC-UKQCD Collaborations and the overlap fermion action is used for the valence quarks. Comparing the current results with our previous ones at a coarser lattice spacing $a\sim0.11$ fm gives us a better understanding of the discretization error. We obtain $f_{D_s^*}^T(\overline{\rm MS},\text{2 GeV})/f_{D_s^*}=0.907(20)$ with a better precision than our previous result. Combining our $f_{D_s^*}=277(11)$ MeV with the total width of $D_s^*$ determined in a recent work gives a branching fraction $4.26(52)\times10^{-5}$ for $D_s^*$ leptonic decay.

hep-lat↗

arXiv2024

PrivLM-Bench: A Multi-level Privacy Evaluation Benchmark for Language Models

The rapid development of language models (LMs) brings unprecedented accessibility and usage for both models and users. On the one hand, powerful LMs achieve state-of-the-art performance over numerous downstream NLP tasks. On the other hand, more and more attention is paid to unrestricted model accesses that may bring malicious privacy risks of data leakage. To address these issues, many recent works propose privacy-preserving language models (PPLMs) with differential privacy (DP). Unfortunately, different DP implementations make it challenging for a fair comparison among existing PPLMs. In this paper, we present PrivLM-Bench, a multi-perspective privacy evaluation benchmark to empirically and intuitively quantify the privacy leakage of LMs. Instead of only reporting DP parameters, PrivLM-Bench sheds light on the neglected inference data privacy during actual usage. PrivLM-Bench first clearly defines multi-faceted privacy objectives. Then, PrivLM-Bench constructs a unified pipeline to perform private fine-tuning. Lastly, PrivLM-Bench performs existing privacy attacks on LMs with pre-defined privacy objectives as the empirical evaluation results. The empirical attack results are used to fairly and intuitively evaluate the privacy leakage of various PPLMs. We conduct extensive experiments on three datasets of GLUE for mainstream LMs.

cs.CL↗