arXiv subjects

Yi Jiang

Publications and source records attributed to Yi Jiang.

At least 37 records · Page 2Linked to original sources

arXiv2026

Quantum geometry and critical temperature enhancement in MgB$_2$ superconductivity

MgB$_2$, a phonon-mediated superconductor with record-high critical temperature $T_c\simeq 39$ K, is revisited to obtain a comprehensive theory of electrons, phonons, and their coupling with minimal ab initio input. We construct compact analytic models for the electronic structure, phonons, and electron-phonon coupling (EPC) of MgB$_2$. We show that strong in-plane B $sp^2$ bonding realizes an obstructed band structure whose natural description is a bond-centered kagome lattice, yielding small quasi-2D $σ$-band Fermi-surface cylinders and pronounced quantum-geometric effects. The phonon spectrum is found to closely track that of a graphene-like boron layer, but the heavy intercalated Mg atoms dominate the three acoustic branches and rigidly lift the boron modes into the optical sector, while the in-plane B-B bond-stretching mode exhibits a pronounced softening along $Γ$-A. By symmetry, this $Γ$-point bond-stretching mode is the only $Γ$ phonon that can couple to the $σ$ Fermi surface, explaining its dominant contribution to the EPC. Upon electron doping toward the doubly degenerate band edge of the $σ$ sheets, we find that a reduced density of states competes with enhanced EPC matrix elements. At light electron doping, ab initio calculations show that the EPC enhancement dominates, leading to an increase in $T_c$ (within the clean doping limit without disorder effects). Using the Gaussian approximation for the EPC tensor, we further show that this enhancement is overwhelmingly quantum geometric in origin, arising from a geometric EPC contribution of the small $σ$ Fermi surface peaked at $Γ$. Overall, our results provide a transparent, symmetry-based account of superconductivity in MgB$_2$ and suggest that quantum-geometric effects can be essential for shaping doping trends in phonon-mediated superconductors.

cond-mat.supr-con↗

arXiv2026

MOSS Transcribe Diarize Technical Report

Speaker-Attributed, Time-Stamped Transcription (SATS) aims to transcribe what is said and to precisely determine the timing of each speaker, which is particularly valuable for meeting transcription. Existing SATS systems rarely adopt an end-to-end formulation and are further constrained by limited context windows, weak long-range speaker memory, and the inability to output timestamps. To address these limitations, we present MOSS Transcribe Diarize, a unified multimodal large language model that jointly performs Speaker-Attributed, Time-Stamped Transcription in an end-to-end paradigm. Trained on extensive real wild data and equipped with a 128k context window for up to 90-minute inputs, MOSS Transcribe Diarize scales well and generalizes robustly. Across comprehensive evaluations, it outperforms state-of-the-art commercial systems on multiple public and in-house benchmarks.

cs.SD↗

arXiv2026

GlaKG: A Biomarker-Centric Fundus Knowledge Graph for Explainable Glaucoma Diagnosis and Risk Assessment

Glaucoma is a leading cause of irreversible blindness worldwide, yet most automated diagnosis systems rely on opaque deep-learning models that offer little clinical interpretability. We present GlaKG, a biomarker-centric fundus knowledge graph that integrates structural biomarkers, clinically grounded rules, and image features to produce traceable reasoning for glaucoma diagnosis and risk stratification. GlaKG encodes six entity types (Fundus Image, Optic Disc, Neural Rim, Pathology, Diagnosis, Risk Level), eight relation types, and 11 clinically validated rules into a unified graph, so that every prediction is accompanied by an explicit reasoning chain linking biomarker evidence to activated clinical rules. To keep knowledge-based reasoning strictly separate from label information, we adopt a post-processing fusion framework that combines ResNet50 image embeddings with a normalized KG reasoning-chain score via a tunable weight alpha, with all fitting confined to the training split. On a publicly available, AI-annotated fundus dataset, GlaKG reaches F1 = 0.9953 for binary glaucoma classification and 0.930 accuracy with 0.922 weighted F1 for four-class risk stratification; we report openly that the dataset's biomarker annotations are highly label-correlated, and therefore frame these figures as an upper bound attainable with clean structured biomarkers rather than as leakage-free image-only performance. Feature-importance analysis shows KG-derived and biomarker features contributing near-equally (51.1% vs. 48.9%), and the reasoning chain flags borderline cases by exposing low chain scores rather than failing silently. GlaKG's central contribution is therefore a clinically auditable reasoning framework that complements raw predictive performance by explicitly exposing the biomarker evidence and rule activations behind each decision.

cs.CV↗

arXiv2026

LoRAShield: Data-Free Editing Alignment for Secure Personalized LoRA Sharing

The proliferation of Low-Rank Adaptation (LoRA) models has democratized personalized text-to-image generation, enabling users to share lightweight models (e.g., personal portraits) on platforms like Civitai and Liblib. However, this "share-and-play" ecosystem introduces critical risks: benign LoRAs can be weaponized by adversaries to generate harmful content (e.g., political, defamatory imagery), undermining creator rights and platform safety. Existing defenses like concept-erasure methods focus on full diffusion models (DMs), neglecting LoRA's unique role as a modular adapter and its vulnerability to adversarial prompt engineering. To bridge this gap, we propose LoRAShield, the first data-free editing framework for securing LoRA models against misuse. Our platform-driven approach dynamically edits and realigns LoRA's weight subspace via adversarial optimization and semantic augmentation. Experimental results demonstrate that LoRAShield achieves remarkable effectiveness, efficiency, and robustness in blocking malicious generations without sacrificing the functionality of the benign task. By shifting the defense to platforms, LoRAShield enables secure, scalable sharing of personalized models, a critical step toward trustworthy generative ecosystems.

cs.CR↗

arXiv2026

New Crystal Structures Hide in Plain Sight: A Stress Test for AI-Guided Materials Discovery

New types of crystal structures are discovered only rarely, and the artificial intelligence (AI) models now reshaping materials discovery have so far produced new chemical compositions within known structural families rather than genuinely new structures. We report GdNiSn4 and LuNiSn4, intermetallics that adopt a previously unreported structure type, found not by computation but by exploratory synthesis. Single-crystal diffraction shows that the structure is an intergrowth of two known structural units. We then use this system as a benchmark for two leading generative models, MatterGen and DiffCSP++. For DiffCSP++, the benchmark is performed in its crystallographically constrained setting, using the required space-group and Wyckoff-position inputs. Under our sampling budget, neither model recovers the experimentally reported monoclinic structure within the structural-matching tolerance. The generated structures are evaluated without further structural relaxation using the nonmagnetic analog LuNiSn4, where we rule out 4f magnetism as the cause. Because the new structure is built from familiar building blocks, it should be derivable. We argue that encoding chemical reasoning, such as the stacking of known motifs, is a concrete path toward AI that can discover structurally novel materials.

cond-mat.mtrl-sci↗

arXiv2026

MOSS-Audio Technical Report

MOSS-Audio is a unified audio-language model for speech, environmental sound, and music understanding, supporting audio captioning, time-aware question answering, timestamped transcription, and audio-grounded reasoning. MOSS-Audio couples a dedicated audio encoder with a modality adapter and a large language model: the encoder produces 12.5 Hz temporal representations, the adapter projects them into the decoder space, and the decoder generates autoregressive text outputs. Two design choices are central to the system: DeepStack cross-layer feature injection, which exposes the decoder to acoustic information from multiple encoder depths, and time markers, which provide explicit temporal cues by inserting timestamp markers into the audio-token stream. At the data level, we design an event-preserving audio annotation pipeline that segments raw audio at coherent event boundaries, applies branch-specific annotation to speech, music, and general audio, and merges the results into unified captions for pretraining. The intermediate branch-specific captions are further retained to support the construction of task-oriented SFT data. The model is pretrained on large-scale audio-language data, with time-aware objectives incorporated to support temporal grounding, and then undergoes multi-stage post-training to enhance instruction following and audio-grounded reasoning. We release 4B and 8B variants in both Instruct and Thinking configurations. MOSS-Audio achieves strong performance across general audio understanding, speech captioning, ASR, and timestamped ASR, positioning it as a promising understanding foundation for future voice agents.

cs.SD↗

arXiv2026

Veda: Scalable Video Diffusion via Distilled Sparse Attention

Scaling Diffusion Transformers to generate high-resolution, long videos is constrained by the quadratic cost of self-attention, and existing sparse attention methods degrade under high sparsity. We show empirically that generation quality is determined not by the sparsity ratio itself, but by how well the sparse mask aligns with the tile-wise geometry of full attention. Based on this insight, we propose Veda, a distilled sparse attention framework that formulates tile selection as an explicit reconstruction problem from full attention. Veda integrates statistics-aware tile scoring with head-aware tiling to reduce estimation error and structural mismatch, enabling aggressive sparsity. A hardware-efficient tile-skipping kernel converts theoretical sparsity into practical wall-clock speedups. Experiments on large video diffusion models, including Waver and Wan2.1, demonstrate substantial acceleration with no noticeable degradation in generation quality. To generate 720P 10-second videos on Waver-T2V-12B, Veda achieves a 5.1$\times$ end-to-end speedup and a 10.5$\times$ self-attention speedup, reducing attention overhead from 92% to 50%. Notably, the gains increase with sequence length, indicating that Veda scales favorably with spatiotemporal resolution across models.

cs.CV↗

arXiv2026

Engineering topological flat bands in $Γ$-valley moiré systems with Ising-type SOC: twisted 1T-ZrS$_2$ and 1T-SnSe$_2$

Twisted moiré superlattices hosting topological flat bands provide a platform to explore the interplay between topology and correlations. Here we investigate topological band structures in $Γ$-valley moiré systems based on 1T-ZrS$_2$ and 1T-SnSe$_2$. Using large-scale ab initio calculations and continuum modelling, we demonstrate that both materials exhibit an approximate spin-$U(1)$ symmetry and host isolated topological moiré valence bands, including quantum spin Hall and high spin Chern states. By constructing a hierarchy of $Γ$-valley moiré continuum models, we show that isolated moiré bands carry a trivial $C_3$ symmetry indicator when the low-energy physics is described by a single effective orbital and a single layer-hybridized branch, either bonding or antibonding. Topological bands therefore arise from inter-branch and/or inter-orbital coupling. Moreover, we determine interaction-driven phase diagrams using Hartree--Fock and exact diagonalization, finding various phases tunable by twist angle, interaction strength, and displacement field. We identify specific conditions under which fractional Chern insulators are favored. Together with previous work showing that the moiré conduction bands of 1T-ZrS$_2$ and 1T-SnSe$_2$ realize $M$-valley twisting and host quasi-one-dimensional physics, our results establish these systems as ideal platforms for strongly correlated moiré physics and provide a systematic framework for understanding topological band structures in $Γ$-valley moiré materials.

cond-mat.mtrl-sci↗

arXiv2026

Planar master integrals for two-loop NLO electroweak light-fermion contributions to $g g \rightarrow Z H$

For the two-loop next-to-leading-order electroweak (NLO EW) corrections to $gg \rightarrow ZH$, the light-fermion contributions can be classified into eight distinct topologies. Using the canonical differential-equations method, we perform an analytic computation of the master integrals (MIs) associated with the four planar topologies. Canonical bases are constructed using the Magnus-expansion method, and the resulting alphabets consist of algebraic symbol letters involving nontrivial radicals. We develop a systematic framework for identifying the radical structures of the canonical MIs, enabling their organization into suitable subsystems and, whenever possible, their representation in terms of Goncharov polylogarithms (GPLs) up to $\mathcal{O}(ε^4)$. Only a few MIs at $\mathcal{O}(ε^3)$ and $\mathcal{O}(ε^4)$ are instead represented as one-fold integrals over GPLs, due to the presence of nested square roots that obstruct the simultaneous rationalization of all radicals.

hep-ph↗

arXiv2026

SAVOIR: Learning Social Savoir-Faire via Shapley-based Reward Attribution

Social intelligence, the ability to navigate complex interpersonal interactions, presents a fundamental challenge for language agents. Training such agents via reinforcement learning requires solving the credit assignment problem: determining how individual utterances contribute to multi-turn dialogue outcomes. Existing approaches directly employ language models to distribute episode-level rewards, yielding attributions that are retrospective and lack theoretical grounding. We propose SAVOIR (ShApley Value fOr SocIal RL), a novel principled framework grounded in cooperative game theory. Our approach combines two complementary principles: expected utility shifts evaluation from retrospective attribution to prospective valuation, capturing an utterance's strategic potential for enabling favorable future trajectories; Shapley values ensure fair credit distribution with axiomatic guarantees of efficiency, symmetry, and marginality. Experiments on the SOTOPIA benchmark demonstrate that SAVOIR achieves new state-of-the-art performance across all evaluation settings, with our 7B model matching or exceeding proprietary models including GPT-4o and Claude-3.5-Sonnet. Notably, even large reasoning models consistently underperform, suggesting social intelligence requires qualitatively different capabilities than analytical reasoning.

cs.AI↗

arXiv2026

Stratagem: Learning Transferable Reasoning via Trajectory-Modulated Game Self-Play

Games offer a compelling paradigm for developing general reasoning capabilities in language models, as they naturally demand strategic planning, probabilistic inference, and adaptive decision-making. However, existing self-play approaches rely solely on terminal game outcomes, providing no mechanism to distinguish transferable reasoning patterns from game-specific heuristics. We present STRATAGEM, which addresses two fundamental barriers to reasoning transfer: domain specificity, where learned patterns remain anchored in game semantics, and contextual stasis, where static game contexts fail to cultivate progressive reasoning. STRATAGEM selectively reinforces trajectories exhibiting abstract, domain-agnostic reasoning through a Reasoning Transferability Coefficient, while incentivizing adaptive reasoning development via a Reasoning Evolution Reward. Experiments across mathematical reasoning, general reasoning, and code generation benchmarks demonstrate substantial improvements, with particularly strong gains on competition-level mathematics where multi-step reasoning is critical. Ablation studies and human evaluation confirm that both components contribute to transferable reasoning.

cs.AI↗

arXiv2026

M100: An Orchestrated Dataflow Architecture Powering General AI Computing

As deep learning-based AI technologies gain momentum, the demand for general-purpose AI computing architectures continues to grow. While GPGPU-based architectures offer versatility for diverse AI workloads, they often fall short in efficiency and cost-effectiveness. Various Domain-Specific Architectures (DSAs) excel at particular AI tasks but struggle to extend across broader applications or adapt to the rapidly evolving AI landscape. M100 is Li Auto's response: a performant, cost-effective architecture for AI inference in Autonomous Driving (AD), Large Language Models (LLMs), and intelligent human interactions, domains crucial to today's most competitive automobile platforms. M100 employs a dataflow parallel architecture, where compiler-architecture co-design orchestrates not only computation but, more critically, data movement across time and space. Leveraging dataflow computing efficiency, our hardware-software co-design improves system performance while reducing hardware complexity and cost. M100 largely eliminates caching: tensor computations are driven by compiler- and runtime-managed data streams flowing between computing elements and on/off-chip memories, yielding greater efficiency and scalability than cache-based systems. Another key principle was selecting the right operational granularity for scheduling, issuing, and execution across compiler, firmware, and hardware. Recognizing commonalities in AI workloads, we chose the tensor as the fundamental data element. M100 demonstrates general AI computing capability across diverse inference applications, including UniAD (for AD) and LLaMA (for LLMs). Benchmarks show M100 outperforms GPGPU architectures in AD applications with higher utilization, representing a promising direction for future general AI computing.

cs.LG↗

arXiv2026

USCNet: Transformer-Based Multimodal Fusion with Segmentation Guidance for Urolithiasis Classification

Kidney stone disease ranks among the most prevalent conditions in urology, and understanding the composition of these stones is essential for creating personalized treatment plans and preventing recurrence. Current methods for analyzing kidney stones depend on postoperative specimens, which prevents rapid classification before surgery. To overcome this limitation, we introduce a new approach called the Urinary Stone Segmentation and Classification Network (USCNet). This innovative method allows for precise preoperative classification of kidney stones by integrating Computed Tomography (CT) images with clinical data from Electronic Health Records (EHR). USCNet employs a Transformer-based multimodal fusion framework with CT-EHR attention and segmentation-guided attention modules for accurate classification. Moreover, a dynamic loss function is introduced to effectively balance the dual objectives of segmentation and classification. Experiments on an in-house kidney stone dataset show that USCNet demonstrates outstanding performance across all evaluation metrics, with its classification efficacy significantly surpassing existing mainstream methods. This study presents a promising solution for the precise preoperative classification of kidney stones, offering substantial clinical benefits. The source code has been made publicly available: https://github.com/ZhangSongqi0506/KidneyStone.

cs.CV↗

arXiv2026

Doubly charged Higgs production within the Higgs triplet model at future electron-positron colliders

We investigate in detail the discovery potential of the doubly charged Higgs boson at the Compact Linear Collider in $e^-e^-$, $e^-γ$, $γγ$, and $e^+e^-$ collision modes, within the Higgs triplet model at two extreme benchmark points as representatives of the Yukawa-like and gauge-like regions. In the Yukawa-like region, the most promising production mechanism is the single production via $e^-e^-$ and $e^-γ$ collisions. Given the subsequent decay of the doubly charged Higgs into a same-sign lepton pair, CLIC can achieve statistical significance well beyond the discovery threshold, within the parameter space permitted by experimental constraints. In the gauge-like region, with the $\ell^{\pm}\ell^{\pm} + \geq 3j$ final state, CLIC exhibits robust discovery potential for the doubly charged Higgs boson, up to a mass of approximately $1.2~\mathrm{TeV}$. We also investigate the search for doubly charged Higgs at the HL-LHC. Our results demonstrate that CLIC possesses greater advantages and offers superior discovery potential for the doubly charged Higgs boson, compared to the HL-LHC.

hep-ph↗

arXiv2026

Cubic Discrete Diffusion: Discrete Visual Generation on High-Dimensional Representation Tokens

Visual generation with discrete tokens has gained significant attention as it enables a unified token prediction paradigm shared with language models, promising seamless multimodal architectures. However, current discrete generation methods remain limited to low-dimensional latent tokens (typically 8-32 dims), sacrificing the semantic richness essential for understanding. While high-dimensional pretrained representations (768-1024 dims) could bridge this gap, their discrete generation poses fundamental challenges. In this paper, we present Cubic Discrete Diffusion (CubiD), the first discrete generation model for high-dimensional representations. CubiD performs fine-grained masking throughout the high-dimensional discrete representation -- any dimension at any position can be masked and predicted from partial observations. This enables the model to learn rich correlations both within and across spatial positions, with the number of generation steps fixed at $T$ regardless of feature dimensionality, where $T \ll hwd$. On ImageNet-256, CubiD achieves state-of-the-art discrete generation with strong scaling behavior from 900M to 3.7B parameters. Crucially, we validate that these discretized tokens preserve original representation capabilities, demonstrating that the same discrete tokens can effectively serve both understanding and generation tasks. We hope this work will inspire future research toward unified multimodal architectures. Code is available at: https://github.com/YuqingWang1029/CubiD.

cs.CV↗

arXiv2026

Autonomous Algorithm Discovery for Ptychography via Evolutionary LLM Reasoning

Ptychography is a computational imaging technique widely used for high-resolution materials characterization, but high-quality reconstructions often require the use of regularization functions that largely remain manually designed. We introduce Ptychi-Evolve, an autonomous framework that uses large language models (LLMs) to discover and evolve novel regularization algorithms. The framework combines LLM-driven code generation with evolutionary mechanisms, including semantically-guided crossover and mutation. Experiments on three challenging datasets (X-ray integrated circuits, low-dose electron microscopy of apoferritin, and multislice imaging with crosstalk artifacts) demonstrate that discovered regularizers outperform conventional reconstructions, achieving up to +0.26 SSIM and +8.3~dB PSNR improvements. Besides, Ptychi-Evolve records algorithm lineage and evolution metadata, enabling interpretable and reproducible analysis of discovered regularizers.

cs.CE↗

arXiv2026

BASE: Burst-Adaptive Autoscaling via Stacked Ensembles for SLO Assurance and Cost Efficiency

Autoscaling is a technology that automatically scales resources for applications without human intervention to ensure runtime Quality of Service (QoS) while reducing costs. However, user-facing cloud applications serve dynamic workloads that often exhibit variability and contain bursts, posing challenges to autoscaling in maintaining QoS within Service-Level Objectives (SLOs). Conservative strategies risk over-provisioning, while aggressive ones may cause SLO violations, making it more challenging to design effective autoscaling. This paper introduces BASE, a burst-adaptive autoscaling framework that leverages a stacked ensemble of machine learning models to mitigate SLO violations and reduce costs for containerized services and applications operating under time-varying workloads. BASE incorporates a novel prediction-based burst detection mechanism that distinguishes between predictable workload spikes and actual uncertain bursts. When bursts are detected, BASE appropriately overestimates them and allocates resources accordingly to address the rapid growth in resource demand. On the other hand, BASE employs reinforcement learning to rectify potential inaccuracies in resource estimation, enabling more precise resource allocation during non-burst periods. Experiments across ten real-world workloads demonstrate BASE's effectiveness, achieving a significant reduction in SLO violations with lower resource costs compared to other prominent methods.

cs.SE↗

arXiv2026

From Sands to Mansions: Towards Automated Cyberattack Emulation with Classical Planning and Large Language Models

Evolving attacker capabilities demand realistic and continuously updated cyberattack emulation for threat-informed defense and security benchmarking. Towards automated attack emulation, this paper defines modular attack actions and a linking model to organize and chain heterogeneous attack tools into causality-preserving cyberattacks. Building on this foundation, we introduce Aurora: an automated cyberattack emulation system powered by symbolic planning and large language models (LLMs). Aurora crafts actionable, causality-preserving attack chains tailored to Cyber Threat Intelligence (CTI) reports and target environments, and automatically executes these emulations. Using Aurora, we generated an extensive cyberattack emulation dataset from 250 attack reports, 15 times larger than the leading expert-crafted dataset. Our evaluation shows that Aurora significantly outperforms existing methods in creating actionable, diverse, and realistic attack chains. We release the dataset and use it to evaluate three state-of-the-art intrusion detection systems, whose performance differed notably from results on older datasets, highlighting the need for up-to-date, automated attack emulation.

cs.CR↗