Search arXiv⌕ Search

arXiv subjects

Shuai Dong

Publications and source records attributed to Shuai Dong.

At least 19 recordsLinked to original sources
arXiv2026

Colossal magnetoresistance in narrow-gap magnetic semiconductor Eu$_2$CdAs$_2$

Discovering antiferromagnetic materials with a sizable colossal magnetoresistance (CMR) effect and precisely tuning their responses remain significant challenges in materials science. This work reports the synthesis, phase transitions, and magnetotransport properties of the narrow-gap Zintl arsenide Eu$_2$CdAs$_2$. The compound undergoes antiferromagnetic ordering at $T_\mathrm{N} = 9$ K and exhibits semiconducting behavior with a low-temperature activation energy of 3.2 meV, achieving a large CMR effect of $-10^4\%$ at 9 T. Below $T_\mathrm{N}$, a distinct butterfly-shaped magnetoresistance hysteresis emerges without any macroscopic ferromagnetic order, revealing a transport anomaly likely dictated by AFM domain dynamics and/or the formation and percolation of localized magnetic polarons. Furthermore, a systematic survey of reported Eu-based Zintl phases uncovers a nearly exponential scaling trend between the maximum CMR magnitude and the low-temperature activation energy. This correlation demonstrates that bandgap engineering serves as an effective strategy for tailoring magnetotransport, establishing Eu-containing Zintl phases as a versatile platform for exploring and manipulating CMR effects.

cond-mat.mtrl-sci↗
arXiv2026

MIRA: A Multimodal Illustrated Deep Research Agent for Controllable Reports

Deep research agents are increasingly capable of searching the open web and synthesizing evidence, yet they remain poorly suited to controllable report generation in multimodal deep research. This task requires agents to handle interleaved textual and visual inputs, gather multimodal evidence through multi-step web search, and deliver comprehensive illustrated reports under user-specified requirements. Existing systems face two limitations. During reasoning, they often underuse images discovered along the search trajectory; during generation, they tend to produce fixed-form answers with limited control over report granularity, multimodal composition, and evidence grounding. We introduce MIRA, a Multimodal Illustrated Deep Research Agent for controllable reports. MIRA integrates textual and visual evidence throughout research and writing. At its core, MIRA-Harness translates user requirements into generation contracts and connects four stages: evidence collection, report submission, feedback-driven revision, and final delivery. Building on this harness, MIRA-Data constructs high-quality multi-hop multimodal research trajectories, and we post-train foundation models to operate within the workflow. Using Qwen3.6-35B-A3B, MIRA achieves state-of-the-art performance among open agents on 7 representative, established multimodal-search and deep-research-agent benchmarks, while producing user-aligned illustrated reports with traceable textual and visual evidence.

cs.CV↗
arXiv2026

What Should a Self-Teacher See? Privileged Context Design for On-Policy Self-Distillation

More privileged information does not always make a better teacher. We study this tension in on-policy self-distillation (OPSD), where a self-teacher scores the student's own rollouts under privileged context, conventionally a complete reference solution that bundles the final answer with one particular reasoning path. Holding the student view and training fixed within each scale, we compare that default against three abstractions compiled offline, a named strategy, a method-independent framing, and a problem category, and against an answer-only control that keeps the destination but removes the path. In the primary runs on competition mathematics, the best intermediate contexts improve the in-domain peak mean over the full solution by 1.4 points at 4B and 1.6 at 8B, while storing an order of magnitude fewer hint tokens. Comparisons across three seeds also show positive mean gains for the framing and category contexts at both scales. Answer-only conditioning remains competitive in the primary runs, within 0.2 points of the full solution at these scales. The preferred context varies with student scale and task. Initial teacher-student KL does not order downstream performance. What a self-teacher should see is therefore not everything it could, but the level of abstraction its student can still act on.

cs.LG↗
arXiv2026

RL Starts before RL: On Policy Distillation for Better Reinforcement Learning

Reinforcement learning (RL) improves reasoning, but its performance depends on the policy from which training begins. We study on-policy distillation (OPD) as a preparation stage for RL and ask whether its benefits extend beyond improvements in the distilled model's initial accuracy. Under shared RL settings, students initialized with OPD reach higher final performance than those trained with direct RL or supervised fine-tuning followed by RL. This advantage can emerge even when OPD produces little immediate improvement in accuracy. Pre-RL Pass@k does not fully explain the benefit: similar or even higher values do not necessarily lead to better performance after RL. Behavioral analyses point to alignment with the teacher's distribution beyond top-1 agreement as a possible explanation. Such alignment may favor higher-quality reasoning paths while retaining alternatives that RL can further refine using outcome feedback. We further examine how trajectory sources and divergence objectives affect the value of distillation for subsequent RL. Standard reverse-KL OPD performs better before RL, but forward-KL OPD overtakes it afterward. Student rollouts outperform teacher rollouts under both objectives before and after RL. These findings highlight the importance of both objective choice and the states receiving supervision for subsequent RL. Our results support OPD as preparation for RL and favor forward KL when OPD is followed by RL in our comparison.

cs.LG↗
arXiv2026

Unconventional Magnetism, Sliding Ferroelectricity, and Magneto-Optical Kerr Effect in Multiferroic Bilayers

Antiferromagnetic (AFM) materials provide a platform to couple altermagnetic (AM) spin-splitting with the magneto-optical Kerr effect (MOKE), offering potential for next-generation quantum technologies. In this work, first-principles calculations, symmetry analysis, and kp modeling are employed to show that interlayer sliding in AFM multiferroic bilayers enables control of electronic, magnetic, and magneto-optical properties. This study reveals an intriguing dimension-driven AM crossover: the 2D paraelectric (PE) bilayer exhibits spin-degenerate bands protected by the [C2||Mc] spin-space symmetry, whereas the 3D counterpart manifests AM spin-splitting along kz \neq 0 paths. Furthermore, interlayer sliding breaks this Mc symmetry and stabilizes a ferroelectric (FE) state with compensated ferrimagnetism, where the Zeeman-like field is responsible for the nonrelativistic spin-splitting. In the FE phase, spin-orbit coupling (SOC) lifts accidental degeneracies and produces `alternating' spin-polarized bands through the interplay of Zeeman and Rashba effects. Crucially, spin polarization, ferrovalley polarization, and the Kerr angle can all be reversed by switching either sliding ferroelectricity or the Neel vector. Our findings reveal the rich coupling among electronic, magnetic, and optical orders in sliding multiferroics, illustrating new prospects for ultralow-power spintronic and optoelectronic devices.

cond-mat.mtrl-sci↗
arXiv2026

One- and two-dimensional cluster states for topological phase simulation and measurement-based quantum computation

Quantum entanglement is a fundamental resource for quantum information processing and serves as a critical benchmark for quantum hardware performance. Cluster states are a special class of entangled states that serve as universal resources for measurement-based quantum computation and possess an intrinsic symmetry-protected topological order, which confers robustness against symmetry-respecting noise. Here we report the scalable preparation and verification of genuine multipartite cluster states on the 105-qubit Zuchongzhi 3.1 superconducting processor. We achieve one-dimensional cluster states of up to 95 qubits and two-dimensional cluster states of up to 72 qubits. The symmetry-protected topological cluster states exhibit input-state-dependent robustness under symmetry-breaking perturbations due to an operational parity structure that enhances the performance of measurement-based quantum computation. Furthermore, we use our two-dimensional cluster states to implement the Deutsch-Jozsa algorithm within the measurement-based quantum computation framework, achieving higher output-state fidelity compared with traditional circuit-based models and a query efficiency advantage over classical approaches. Our work establishes a scalable platform that combines large-scale entanglement generation, symmetry-protected topological order and practical quantum algorithms to enable robust, fault-tolerant measurement-based quantum computation.

quant-ph↗
arXiv2026

SafeRI: Recognition and Intervention for Token-Level Safety Intervention in Large Vision Language Models

Existing safety alignment methods for vision-language models usually modify the model behavior globally: once the safety parameters are trained or loaded, they participate in both unsafe and already-safe generations. This always-on intervention can unnecessarily perturb the model's original reasoning path and degrade general multimodal capabilities. We argue that safety alignment should be an on-demand intervention rather than a permanent modification to every decoding trajectory. To this end, we propose a streaming recognition and gated LoRA framework for intrinsic VLM safety. During autoregressive generation, a lightweight recognizer estimates whether the current pre-token generation state is safe or unsafe. Its output updates the LoRA gate for the following decoding step; otherwise, generation follows the frozen-backbone policy. The LoRA module is trained from unsafe prefixes, transition statements, and safe continuations, so that it learns to redirect unsafe generations back to safe responses after activation. Experiments across multiple safety and general-purpose benchmarks demonstrate the effectiveness of our method in post-alignment settings.

cs.CV↗
arXiv2026

Beyond Visual Memory: Mechanistic Diagnostics of Latent Visual Reasoning

Recent latent visual reasoning methods achieve substantial gains by inserting continuous latent tokens into multimodal language models. These gains are commonly attributed to the tokens encoding visual evidence; recent analyses, however, reveal a paradox: the tokens are loosely tied to the image and contribute little to the answer. Critically, these analyses treat latent tokens as a single unit, obscuring the source of the gains. We therefore decompose latent tokens into three testable components: latent slots, boundary markers, and format, and develop a state-of-the-art method as a probe under favorable conditions. Across six method-stage settings and four perception-heavy benchmarks, latent slots fail every prediction of the visual-memory account. Strikingly, retaining only the boundary markers preserves 78 to 100% of the gain in several settings, while the model attends to the image more narrowly at latent positions than at answer positions. These results do not support slot contents as recoverable visual memory; much of the benefit is instead associated with marker and format control, together with visual-attention routing. At matched accuracy, methods can still rely on markedly different mechanisms shaped by training supervision. Latent visual reasoning thus needs evaluation not only by accuracy but by what the model actually relies on.

cs.CV↗
arXiv2026

Autferroics-based true random number generators with enhanced performance

Physical entropy-driven true random number generators are essential for emerging probabilistic computing paradigms, but conventional implementations based on magnetic tunneling junctions have reached their performance plateaus limited by inherent tradeoffs and weak tunability. Here, autferroics, a sister branch of multiferroics, is proposed to construct true random number generators. Benefiting from its unique energy landscape due to strong seesaw-type magnetoelectricity, the performance of true random number generation can be significantly enhanced in autferroic tunneling junctions, verified by passing standard statistical tests. Furthermore, autferroics-based random number generators can exhibit multi-field tunability and intrinsic multi-state randomness, opening an avenue for efficient hardware realization of the complex-number arithmetic and simulation of probability distributions of quantum mixed states within stochastic circuits.

cond-mat.mtrl-sci↗
arXiv2026

SafeCap: Improving LVLM Safety with Image Captioning Reinforcement Learning

Large vision-language models (LVLMs) remain vulnerable to jailbreak attacks that exploit visual inputs to bypass safety alignment inherited from their language backbones. We propose SafeCap, a reinforcement-learning framework that aligns LVLMs through learned self-captioning. SafeCap trains a policy model to first generate a safety-relevant image caption and then produce a final answer; the caption is further optimized by whether it enables a frozen LLM to reach a safety-aligned decision. This caption-mediated objective encourages the policy to expose visual cues relevant to safe response generation rather than relying solely on direct refusal supervision. Across five multimodal safety benchmarks and six vision-utility benchmarks, SafeCap substantially improves aggregate safety performance under its intended DirectCap protocol, with gains of 3.7-19.0 points in safety average across four model settings while maintaining comparable or improved vision utility. Under controlled comparisons on matched backbones and data, SafeCap outperforms safety SFT, DPO, and SafeGRPO, demonstrating the effectiveness of caption-mediated reinforcement learning for multimodal safety alignment.

cs.CV↗
arXiv2026

Ferroelectric switchable altermagnetic-like compensated ferrimagnets with charge ordering

Unconventional collinear magnets with almost zero magnetization but prominent nonrelativistic spin-splitting, such as altermagnets, can inherit the advantages of both ferromagnets and antiferromagnets. By incorporating more degrees of freedom such as ferroelectricity and charge ordering, these unconventional magnets can be even more interesting and functionalized. With this design principle, the Fe$_3$O$_5$ monolayer is predicted to exhibit a hybrid spin-splitting mechanism, with the superposition of the altermagnetic-like $k$-path alternating splitting and ferrimagnet-like Zeeman splitting. Benefiting from the hidden magnetoelectricity based on the spin-charge coupling, such spin-splitting can be fully switched by an electric field. Its conductivity is highly spin-polarized, with a polarization ratio above $99\%$, comparable to half-metals but with zero magnetization.

cond-mat.mtrl-sci↗
arXiv2026

Sizable Ligand-Mediated Bond-Dependent Interactions in a Spin-1 Triangular Antiferromagnet NiI$_2$

The bond-dependent anisotropic Kitaev interactions are the key for the Kitaev model, which has attracted intense interest for its potential to host quantum-spin-liquid states and fractional excitations. However, experimental realizations of such interactions remain scarce. Here, we investigate the magnetic excitations of NiI$_2$, a van der Waals magnet with spin $S=1$. By combining inelastic neutron scattering, magnetization measurements, magnetic structure analysis, first-principles calculations, and linear-spin-wave simulations, we identify a minimal model that features substantial Kitaev and off-diagonal $Γ$ interactions, which together stabilize the canted magnetic ground state and open a gap in the spin-wave spectrum. Notably, these interactions arise from strong spin-orbit coupling on the ligand ions, despite the quenched orbital moment of the magnetic Ni$^{2+}$ ions. Our results provide compelling experimental evidence for the ligand-driven Kitaev mechanism. This demonstrates a concrete pathway to generating strong bond-dependent anisotropy in systems where the magnetic ions themselves have weak spin-orbit coupling, thereby substantially broadening the range of potential Kitaev materials.

cond-mat.str-el↗
arXiv2026

Phase-Factor-Controlled Interaction and Bonding between a Chiral Bobber and a Skyrmion String in the Conical Phase

Skyrmion interactions govern the formation of skyrmion lattices, clusters, and particle-like growth patterns. In contrast, the interaction between a chiral bobber and a skyrmion string remains largely unexplored, despite the role of bobbers as intermediate states in skyrmion-string formation and annihilation. Here we show that this interaction is intrinsically phase dependent in the conical phase. Using three-dimensional micromagnetic simulations, we compute the locally relaxed constrained energy landscape as a function of the bobber--skyrmion separation $R$ and the surface phase factor $ϕ_0$. We find that changing $ϕ_0$ qualitatively reshapes the interaction, producing repulsive, attractive, and bonding-like regimes that cannot be reduced to a conventional distance-dependent potential. Real-space analysis shows that this behavior originates from phase-dependent reconstruction of nonaxisymmetric outer distortion shells. The phase-controlled interaction persists over a finite field range and follows the expected top--bottom phase relation of surface-sensitive conical textures. These results identify the conical phase direction, often hidden in projected or thickness-averaged descriptions, as a previously underappreciated degree of freedom in the interaction between skyrmion strings and finite-length chiral textures, as demonstrated here for chiral bobbers.

cond-mat.mes-hall↗
arXiv2026

DOPD: Dual On-policy Distillation

On-policy distillation (OPD) offers superior capacity transfer by supervising student-sampled trajectories with dense token-level signals. To furnish high-quality supervision sources and thereby elevate the performance frontier of distillation, an intuitive direction is to infuse privileged information to either teacher or student itself. However, this additional input induces a potential failure mode we dub privilege illusion: a pattern that conflates the transferable capability gap that students are meant to close, and the information asymmetry gap that can only be mimicked but never replicated. This issue is further amplified by the inherent non-uniformity of token-level supervision, where only a small subset of tokens carries pivotal capability-bearing signals. To this end, we propose DOPD, an advantage-aware dual distillation paradigm that dynamically routes token-level supervision between privileged teacher and privileged student policies based on their advantage gap and relative probabilities. Each token receives supervision of different strength, objective, and strategy from either teacher or student itself, which transfers credible capability while simultaneously receiving auxiliary signals, to alleviate privilege illusion. Extensive experiments on both large language model (LLM) and vision-language model (VLM) settings demonstrate that DOPD consistently outperforms Vanilla OPD and other counterparts. Further results on stability, robustness, continual learning, and out-of-distribution tasks validate its superiority.

cs.AI↗
arXiv2026

Room-Temperature Noncolinear Ferroelectricity in van der Waals WO$_2$Cl$_2$ with a Wide Bandgap

Low-dimensional ferroelectrics are attractive for their promising prospects in nanoelectronics. Compared with widely-used ferroelectric perovskites, most low-dimensional ferroelectrics exhibit several inborn weaknesses such as small bandgaps (mostly <2 eV, i.e. semiconductors-like) or faint polarizations (e.g. $<1$ $μ$C/cm$^2$ for sliding ferroelectrics even if their bandgaps can be large). Here we experimentally demonstrate the room-temperature ferroelectricity of van der Waals WO$_2$Cl$_2$ . The well-tested d0 rule inherited from ferroelectric perovskites leads to a large dipole (~3 eÅ) from the off-center displacement of W$^6+$ ion and a wide bandgap of 2.80 eV. Its ferroelectricity is proved by multiple characterizations including second harmonic generation, piezoresponse force microscopy, and ferroelectric hysteresis loops. More interestingly, the exotic noncollinear dipole order is directly observed at the atomic level by integrated differential phase contrast scanning transmission electron microscopy. Our work paves an alternative route for low-dimensional ferroelectrics to pursue excellent ferroelectric performance and distinct physics of polarity.

cond-mat.mtrl-sci↗
arXiv2026

Flash-GRPO: Efficient Alignment for Video Diffusion via One-Step Policy Optimization

Group Relative Policy Optimization has emerged as essential for aligning video diffusion models with human preferences, but faces a critical computational bottleneck: training a 14B parametered model typically demands hundreds of GPU days per experiment. Existing efficiency methods reduce costs through sliding window subsampling training timesteps, but fundamentally compromise optimization, exhibiting severe instability and failing to reach full trajectory performance. We present Flash-GRPO, a single-step training framework that outperforms full trajectory training in alignment quality under low computational budgets while substantially improving training efficiency. Flash-GRPO addresses two critical challenges: iso-temporal grouping eliminates timestep-confounded variance by enforcing prompt-wise temporal consistency, decoupling policy performance from timestep difficulty; temporal gradient rectification neutralizes the time-dependent scaling factor that causes vastly inconsistent gradient magnitudes across timesteps. Experiments on 1.3B to 14B parameter models validate Flash-GRPO's effectiveness, demonstrating substantial training acceleration with consistent stability and state-of-the-art alignment quality.

cs.CV↗
arXiv2026

GEAR-VLA: Learning Geometry-Aware Action Representations for Generalizable Robotic Manipulation

Vision-Language-Action (VLA) models achieve strong benchmark performance but still struggle in real-world deployment with unseen objects, background shifts, and different robot embodiments. We argue that this stems from the lack of a unified geometry-aware manipulation representation, leaving existing VLAs vulnerable to low-level trajectory supervision, misaligned 3D features, and embodiment differences. To address this, we propose GEAR-VLA, a VLA framework for learning unified geometry-aware action representations for generalizable robotic manipulation. GEAR-VLA adopts coarse-to-fine action learning, where multi-source embodied pretraining equips the VLM with embodied reasoning and discrete action understanding before latent action tokens connect action semantics to a gradient-decoupled DiT continuous action expert. It further performs semantic-aligned 3D integration by aligning a trainable 3D spatial backbone with the VLA representation while freezing the original VLM-aligned visual pathway. To share this representation across robots, GEAR-VLA uses embodiment canonicalization, where embodiment-aware states and embodiment-invariant actions confine robot differences to the low-level interface. Extensive simulation and real-world experiments demonstrate strong generalization: GEAR-VLA achieves state-of-the-art performance on LIBERO, zero-shot LIBERO-Plus, and RoboTwin 2.0, reaches 85.9% success on AgileX and 81.0% on the pretraining-unseen LDT-01 embodiment, and obtains 90.1% success on a 6,360-trial universal grasping benchmark with 212 unseen objects. Code and models will be released at https://github.com/babynabeauty/GEAR-VLA.

cs.RO↗
arXiv2026

NeuDW-CIM: a 65-nm 0.8-pJ/Sop Reconfigurable Neuromorphic Compute-in-Memory Macro with Nonlinear Dendrites and K-Winners

This work presents NeuDW-CIM, a highly efficient neuromorphic Compute-in-Memory (CIM) macro for Spiking Neural Networks (SNNs) implemented in 65 nm CMOS. The design introduces a custom twin 9T bit-cell for ternary in-puts/weights and a reconfigurable non-linear In-Memory ADC (IMA). The macro supports two specialized modes: 1) Nonlinear Dendrite (NLD) mode, which utilizes reconfigurable IMA to emulate biological dendritic functions, achieving measured accuracies of 97.2% on N-MNIST and 95.5% on DVS Gesture; and 2) Top-K Winner (KWN) mode, featuring an early-stopping mechanism that reduces IMA conversion latency by 30% and digital LIF latency by 10x. Benefiting from the sparse update in KWN mode, NeuDW-CIM achieves a measured energy efficiency (EE) of 0.8 pJ/SOP (1.6x improvement).

cs.AR↗