Searcharxiv⌕ Search

arXiv subjects

Ting Cao

Publications and source records attributed to Ting Cao.

At least 19 recordsLinked to original sources
arXiv2026

MakoXC: Rearchitecting DFT Exchange-Correlation with Matrix-Aligned and Knowledge-Organized Sparsity

Density Functional Theory (DFT) is indispensable for materials science and drug discovery, yet the exchange--correlation (XC) evaluation remains a major bottleneck due to its cubic scaling. Although linear-scaling methods exploit electronic nearsightedness to reduce asymptotic complexity, they produce irregular sparse workloads that hide implicit sparsity and prevent efficient use of modern AI accelerators. We present MakoXC, a modular matrix-aligned XC evaluation engine that rearchitects nearsightedness-induced sparsity into regular, accelerator-friendly computations. MakoXC co-designs three key techniques: (1) Matrix-Aligned Cells reorganize nearsightedness-induced interactions into dense, accelerator-aligned data clusters; (2) Sparsity-Guided Activation translates deeper implicit sparsity into numerically correct structured execution for practical linear scaling; and (3) Kernel-Fused Pipeline consolidates fragmented workloads into a unified, compute-intensive execution path that fully unleashes accelerator throughput. Extensive evaluations show that MakoXC achieves average speedups of 67.8$\times$ speedup over standard XC evaluation and 4.7$\times$ over state-of-the-art linear-scaling methods. When integrated into a production-grade commercial DFT package, MakoXC scales XC evaluation to ubiquitin (1,231 atoms, def2-SVP) on 64 GPUs, enabling the end-to-end DFT calculation to complete in under five minutes. By restructuring XC evaluation into a unified, structured computation, MakoXC demonstrates how scientific workloads can achieve genuine low complexity while maximizing parallel efficiency on AI accelerators.

cs.DC↗
arXiv2026

Zeva: In-Context Causal Learning for Generalizable Embodied Manipulation

Generalizable embodied manipulation remains difficult to achieve through pretraining alone, due to unseen physical conditions in the real world. We argue that robots need to learn from their own physical interactions on the fly during real-world deployment and use this knowledge to inform subsequent actions. We present Zeva, the first framework that enables in-context learning from a robot's own physical interaction experience while keeping the policy model frozen. Zeva employs a Causal Interaction Extractor to encode an executed action and its induced state change into a causal interaction signal, which is stored in a dual-timescale causal memory. For subsequent actions, relevant causal interaction signals are retrieved from memory and injected into the frozen policy model as context. Experiments in simulation and real-world manipulation demonstrate that Zeva achieves the best performance among the compared frontier VLAs and WAMs and, more importantly, enables self-evolution during deployment without gradient updates. Its success rate continues to improve as the robot accumulates interaction experience. Furthermore, the acquired interaction experience can generalize across tasks.

cs.RO↗
arXiv2026

Zetta $\zeta$: An Efficient Closed-Loop Embodied Harness for Self-Evolving Physical Intelligence

Embodied agents are increasingly used to close the gap left by end-to-end policy models. Yet the agentic path has not realized closed-loop learning in physical execution: existing harnesses remain largely open-loop, following fixed skills during rollout and reflecting only after an episode completes. Such post-hoc reflection cannot govern execution as it unfolds, because physical interaction requires decisions to track rapidly changing robot-environment states at a frequency beyond today's large agentic models. We present Zetta, a closed-loop embodied harness that evolves code-based runtime critics and recovery skills online while keeping the base policy frozen. Through three timescale-separated loops, Zetta provides action-frequency governance, rollout-level critic-recovery proposal, and validation-gated skill updates. Together with Z-Infra, a rollout infrastructure decoupling agent logic from heterogeneous execution resources, Zetta achieves state-of-the-art success on LIBERO-Pro and RoboCasa under our current rollout budget, reaching 90.8% and 93.6%, with an 11.1x inference speedup; success continues to scale with self-exploration experience; learned skills transfer zero-shot, and clear robotic "Aha Moments" emerge. These results show that closed-loop harness self-evolution opens a scaling path for reliable physical intelligence.

cs.RO↗
arXiv2026

Tunable Multiband Geometry and Fractional Phases in Higher Vortexable Systems

Higher vortexability is often viewed as a route to topological flat bands with higher-Landau-level-like quantum geometry. Here we emphasize a complementary perspective: it provides a tunable multiband structure in which wave function geometry can be varied continuously while the band dispersion, degeneracy, and topology remain fixed. We perform systematic exact-diagonalization studies of many-body phases in fractionally filled higher vortexable moir\'e systems, retaining the full flat-band Hilbert space rather than projecting onto a single band. The multiband treatment reveals a cascade of Abelian and non-Abelian phases at zero magnetic field, including integer and fractional exciton insulators, Abelian fractional Chern insulators, Moore-Read and Read-Rezayi states. At fixed filling, different phases are connected through transitions or crossovers driven solely by changes in wave function geometry, highlighting quantum geometry itself as a direct tuning parameter between competing topological states. At fillings associated with Moore-Read and Read-Rezayi states, our calculations show that interband mixing shifts the optimal quantum geometry regime without suppressing non-Abelian topological order under screened Coulomb interaction. Our results establish higher vortexable moir\'e bands as a tunable platform for exploring geometry-driven multiband topological phases at zero magnetic field.

cond-mat.str-el↗
arXiv2026

Topological Charge-Transfer Excitons

Excitons possess internal structure absent from single-particle Bloch particles, allowing their band topology to emerge from the bound-state structure rather than being inherited from their constituents. This raises the question of how the internal structure of a bound state can provide a microscopic origin of exciton topology. Here we show that the real-space embedding of charge-transfer excitons can generate an intrinsic manifold of symmetry-related off-site composite orbitals whose coupling supports topological exciton bands. Lateral electron-hole separation embeds the localized exciton on the bond connecting its constituent sites rather than on either site. We demonstrate this mechanism in a honeycomb lattice, where three bond-centered charge-transfer exciton orbitals form a Kagome lattice. By solving the Bethe-Salpeter equation, we show that this emergent multi-orbital manifold supports a topological exciton flat band upon time-reversal symmetry breaking, even when the electron and hole bands are topologically trivial. The resulting band exhibits nearly uniformly distributed quantum geometry, favorable for interaction-driven bosonic states. Our results establish a general route toward topological bands of localized composite bound states and unconventional strongly correlated bosonic phases.

cond-mat.mes-hall↗
arXiv2026

Training Skills Like Parameters via Self-Supervised Semantic Diffusion

While Large Language Models (LLMs) demonstrate remarkable general instruction-following capabilities, they often fall short of human experts in highly specialized, open-ended domains such as creative screenwriting. Prior approaches typically adopt post-training, yet both supervised fine-tuning and reinforcement learning require weight access that closed-source frontier models do not offer, and demand heavy compute. Moreover, what is learned is tied to a single checkpoint and cannot be inspected by humans. Recent advancements in agentic continual learning instead attempt to bridge this gap by accumulating external textual skills. However, these methods heavily rely on costly human expert annotations or unreliable LLM-as-a-judge feedback for reflection. To overcome this bottleneck, we propose a novel, unsupervised self-evolving agent framework inspired by the corruption-and-reconstruction paradigm of diffusion models. Instead of relying on explicit external scoring, we leverage existing high-quality human artifacts to construct self-supervised signals. Training then follows the familiar loop of neural network training, forward, loss, and backward, with the loss coming from contrasting the agent's reconstruction against the human original. What is updated is not model weights but an external library of textual skills. We evaluate our framework on the challenging task of short drama screenwriting. Experimental results demonstrate that our method enables the agent to autonomously extract and internalize highly generalizable skills, significantly enhancing its domain-specific generation capabilities. Furthermore, this self-contrastive reflection paradigm offers a scalable pathway for agents to teach themselves the production of complex, high-quality human artifacts, without requiring external supervision.

cs.CL↗
arXiv2026

Ab initio time-dependent GW approach for nonequilibrium exciton-phonon coupled dynamics across momentum space

The dynamics of optical excitations in materials generally involves intertwined electron-hole (e-h) and electron-phonon (e-ph) interactions out of equilibrium. However, a full theoretical description of such nonequilibrium dynamics requires a systematic treatment of the coherent excitonic excitations and exciton-phonon interactions across the entire crystal momentum space in real time, which remains a major challenge and out of reach for first-principles approaches. Here, we present a new ab initio time-dependent adiabatic GW methodology that incorporates full finite-momentum e-h and e-ph couplings, enabling real-time simulations of the coherently coupled exciton-phonon dynamics. The excitonic excitations are naturally described by the equation of motion of the interacting single-particle density matrix, whereas their couplings to phonons are formulated within a linear-response framework, hence the simulations can be efficiently carried out within a primitive unit cell. We demonstrate the capabilities of this new approach by investigating the direct-to-indirect exciton transitions in monolayer WSe2 in a pump-probe setup of time-resolved and angle-resolved photoemission spectroscopy. Our results reveal that the phonon-mediated ultrafast intervalley dynamics of excitons of this system is within ~0.5 ps, manifested as in-gap photoemission intensity transfer from the K-valley to the Q-valley. This work establishes a comprehensive and practical nonequilibrium Green's function framework for accurately simulating nonequilibrium and coherent excitations involving coupled excitons and phonons from first principles.

cond-mat.mtrl-sci↗
arXiv2026

Embodied.cpp: A Portable Inference Runtime of Embodied AI Models on Heterogeneous Robots

Embodied AI models now span vision-language-action (VLA) models and world-action models (WAMs), but practical deployment remains fragmented across model-specific Python stacks, backend assumptions, and robot-side glue code, especially on heterogeneous edge devices. Existing inference runtimes are designed mainly for request-response serving and therefore do not satisfy the runtime contract of embodied deployment: multi-rate execution inside closed-loop control, latency-first batch-1 inference on heterogeneous hardware, and extensible embodied interfaces beyond fixed token I/O. We present Embodied$.$cpp, a portable C++ inference runtime for embodied models. Based on an architectural analysis of representative VLA models and WAMs, Embodied$.$cpp captures a shared execution path and organizes it into five layers: input adapters, sequence builders, backbone execution, head plugins, and deployment adapters. The runtime provides modular multi-rate execution, latency-first fused inference, and extensible operator and I/O support, enabling deployment across heterogeneous devices, robots, and simulators through one backend abstraction. We evaluate Embodied$.$cpp on three VLA and two WAM models, using normalized comparisons across Python and C++ quantization configurations. Overall, Embodied$.$cpp achieves 1.05x-2.70x inference speedups and 7\%-77\% lower VRAM relative to Python baselines, while maintaining near-baseline success for most configurations. These results show that Embodied$.$cpp improves deployment efficiency while preserving high control quality across diverse embodied model architectures. Project Link: https://github.com/SEU-PAISys/Embodied.cpp

cs.RO↗
arXiv2026

Electrically Programmable Correlated Topology and Magnetism in a Moir\'e Trilayer

Strong electron-electron interactions underlie a wide range of quantum many-body phenomena, including magnetism, superconductivity, and charge fractionalization. A central goal is to achieve in situ control over lattice geometry, bandwidth, and band topology within a single platform. Here we realize such an electrically programmable quantum many-body system in an alternating twisted trilayer MoTe$_2$, where an out-of-plane displacement field continuously modifies the layer polarization, effective lattice, and topology of the moir\'e bands. At zero displacement field, the system realizes a triangular lattice hosting a correlated insulator at one hole per moir\'e unit cell ($\nu = -1$). Doping this state produces strongly asymmetric magnetic responses: double-exchange-like ferromagnetism for $|\nu| > 1$, and signatures of spin polarons and antiferromagnetism for $|\nu| < 1$. At large displacement field, interlayer hybridization reconstructs the electronic structure into a honeycomb lattice with a flat Chern band, supporting integer and fractional Chern insulators. Magneto-optical measurements further reveal the signatures of gap closure and Landau-level formation from a spin-polarized Fermi surface near the crossover between the two regimes. These results establish a unified, electrically tunable platform in which correlated magnetism and topological states emerge from a single controllable band structure.

cond-mat.mes-hall↗
arXiv2026

Switching Chern number by sliding and gating in alternately twisted tetralayer MoTe2

Switching the bulk Chern number in topological materials is of central importance for the design of topological electronic devices. Motivated by recent observations of integer and fractional quantum anomalous Hall effects in twisted transition metal dichalcogenides (tTMDs), we realize the switching of valley Chern number through sliding and gating in alternately twisted tetralayer (ATT) MoTe$_{2}$. Using large-scale density functional theory (DFT) calculations, we show that the Chern number of the first $K$-valley moir\'e band evolves from $+1$ to $-1$ under the interlayer sliding. Furthermore, an applied electric field can switch the valley Chern number from $-1$ to $+1$. Based on the developed continuum model, we reveal that these switching behaviors are caused by the sliding- and gate-dependent intralayer moir\'e potential distributions across the layers. Our results establish ATT MoTe$_{2}$ as a promising platform for engineering moir\'e band topologies through the design of moir\'e potentials with sliding in multilayer moir\'e systems.

cond-mat.mtrl-sci↗
arXiv2026

ActProbe: Action-Space Probe for Early Failure Detection of Generative Robot Policies

Generative robot policies fail unpredictably at deployment: they hesitate at critical moments, drift off-task, or commit to unrecoverable actions. Existing online failure detectors either require white-box access to policy internals or add runtime overhead through resampling and observation-side signals. Our empirical analysis shows that emitted action chunks themselves already carry strong predictive signal for impending failures in generative robot policies. Motivated by this observation, we introduce ActProbe, a lightweight, pure action-space detector that uses two compact signals available from a single forward pass: Temporal Consistency Error (TCE) between consecutive action chunks and Action Chunk Magnitude (ACM) of the current chunk. ActProbe maps these signals to per-step failure probabilities with a task-conditioned LSTM-MLP architecture. Across a diverse suite of generative robot policies and benchmarks, ActProbe raises alerts before failures become visually recognizable, improving the accuracy (F1)-timeliness Pareto frontier of failure detection by an average hypervolume gain of +12.7% over both internal- and external-feature baselines, with a +9.0% early-detection ROC-AUC lead on unseen tasks. ActProbe further transfers to deployment, predicting failures on unseen real-robot pick tasks and accelerating RL fine-tuning (PPO) with 2.9x fewer environment interactions.

cs.RO↗
arXiv2026

Joint Agent Memory and Exploration Learning via Novelty Signals

In open-ended environments, exploration is fundamental for autonomous agents, yet current language model agents struggle with this. Effective exploration requires memory, but retaining raw interaction histories is computationally expensive over long trajectories. While latent memory offers a solution to compress interaction histories, its training lacks reliable supervisory signals. We introduce \textbf{J}oint \textbf{A}gent \textbf{M}emory and \textbf{E}xploration \textbf{L}earning (\textbf{JAMEL}), a framework that trains agentic memory and exploration policy together through novelty-driven interaction. We observe that memory and exploration form a mutually dependent loop: sustained exploration requires memory to distinguish exhausted behaviors from unseen ones, while novelty-seeking interaction provides the supervision needed to make memory useful for future exploration. By utilizing deterministic and persistent novelty signals such as code coverage in the GUI domain, we provide natural, annotation-free supervision for the memory module. Empirical evaluations demonstrate that \ours successfully generalizes to unseen environments. Its exploration capability outperforms open-weight baselines and rivals the exploration depth of a closed-source model while reducing token consumption. Our code and model are open-sourced at https://github.com/MobileLLM/JAMEL.

cs.AI↗
arXiv2026

GRIP-VLM: Group-Relative Importance Pruning for Efficient Vision-Language Models

In Vision-Language Models (VLMs), processing a massive number of visual tokens incurs prohibitive computational overhead. While recent training-aware pruning methods attempt to selectively discard redundant tokens, they largely rely on continuous-gradient relaxations. However, visual token pruning is inherently a discrete, non-convex combinatorial problem; consequently, these continuous approximations frequently trap the optimization in sub-optimal local minima, especially under aggressive compression budgets. To overcome this fundamental bottleneck, we propose GRIP-VLM, a Group-Relative Importance Pruning framework driven by Reinforcement Learning. Rather than relying on smooth-gradient assumptions, GRIP-VLM formulates pruning as a Markov Decision Process, employing a Group Relative Policy Optimization (GRPO) paradigm anchored by supervised warm-up to directly explore the discrete selection space. Integrated with a budget-aware scorer, our lightweight agent dynamically evaluates per-token importance and adapts to arbitrary compression ratios without retraining. Extensive experiments across diverse multimodal benchmarks demonstrate that GRIP-VLM consistently outperforms heuristic and supervised-learning baselines, achieving a superior Pareto frontier and delivering up to a 15\% inference speedup at equal accuracy.

cs.CV↗
arXiv2026

EmbodiSkill: Skill-Aware Reflection for Self-Evolving Embodied Agents

Embodied agents can benefit from skills that guide object search, action execution, and state changes across diverse environments. Since embodied environments vary across layouts, object states, and other execution factors, these skills must self-evolve from trajectories generated during task execution. However, existing skill self-evolution methods are mainly developed in digital environments and often convert trajectories into coarse skill updates. Directly applying this paradigm to embodied settings is problematic, because a failed task execution may reflect not only incorrect skill content, but also an execution lapse in which the agent fails to follow valid guidance. We propose EmbodiSkill, a training-free framework for embodied skill self-evolution through skill-aware reflection and targeted revision. EmbodiSkill interprets each trajectory with respect to the current skill, uses skill-changing evidence to update the skill body, and uses execution-lapse evidence to preserve and emphasize valid guidance. Experiments on ALFWorld and EmbodiedBench show that EmbodiSkill consistently improves embodied task success. On ALFWorld, EmbodiSkill enables a frozen Qwen3.5-27B executor to reach 93.28% task success, outperforming GPT-5.2 used as a direct agent without skills by 31.58%. These results show that skill-aware self-evolution helps embodied agents accumulate reusable procedural knowledge from their own trajectories.

cs.AI↗
arXiv2026

MemCompiler: Compile, Don't Inject -- State-Conditioned Memory for Embodied Agents

Existing memory systems for embodied agents typically inject retrieved memory as static context at episode start, a paradigm we term Ahead-of-time Monolithic Memory Injection (AMMI). However, this static design quickly becomes misaligned with the agent's evolving state and may degrade lightweight executors below the no-memory baseline. To address this, we propose MemCompiler, which reframes memory utilization as State-Conditioned Memory Compilation. A learned Memory Compiler reads a structured Brief State capturing the agent's current execution state and dynamically selects and compiles only relevant memory into executable guidance. This guidance is delivered through a text channel and a latent Soft-Mem channel that preserves perceptual information not expressible in text. Across Alf World, EmbodiedBench, and ScienceWorld, MemCompiler consistently improves over no-memory across open-source backbones (up to +129%), matches or approaches frontier closed-source systems, and reduces per-step latency by 60%, demonstrating that state-aware memory compilation improves both effectiveness and efficiency.

cs.RO↗
arXiv2026

Nonadiabatic Theory of Phonon Magnetic Moments in Insulators and Metals

We develop a nonadiabatic theory of phonon magnetic moments applicable to both insulators and metals. By relating the phonon magnetic moment to the force-velocity response of ions in a magnetic field, we derive a gauge-invariant expression using a gauge-covariant Wigner expansion. The formalism naturally separates Fermi-sea and Fermi-surface contributions and captures the full dependence on phonon frequency. In gapped systems, our theory reduces to previous adiabatic expressions in the low-frequency limit. Beyond this limit, it reveals additional contributions arising from resonant interband processes and the Fermi surface. Applying our theory to Pb$_{1-x}$Sn$_x$Te, we find that the Fermi-surface contribution substantially enhances the phonon magnetic moment, reproducing the same order of magnitude as the experimental observation. Our results provide a unified framework for describing phonon magnetic moments beyond the adiabatic regime.

cond-mat.mes-hall↗
arXiv2026

Quantum Geometric Quadrupole of Cooper Pairs

The size of Cooper pairs defines a fundamental length scale of superconductivity, conventionally set by band dispersion and the superconducting gap. This picture breaks down in flat bands, where quenched dispersion makes quantum geometry essential. Here we develop a general framework based on the Cooper pair quadrupole moment, whose trace gives the pair size. The framework holds for both dispersive and flat-band cases, and provides a unified description of the geometric origin of this length scale. In particular, when time-reversal symmetry is broken, Berry curvature enters through the phase structure of the pair wavefunction and gives an essential contribution absent from previous quantum-metric theories. Together, Berry curvature and quantum metric impose a geometric lower bound on the pair size. Applying this framework to rhombohedral graphene, we find that the Berry-curvature-induced contribution can dominate and yields pair sizes comparable to experimentally inferred coherence lengths. These results identify Berry curvature as a central geometric ingredient controlling the microscopic length scale of superconductivity.

cond-mat.supr-con↗
arXiv2026

Unfolding an Atomistic World: Atomistic Simulation of Reactor Pressure Vessel Steel Across Year-and-Meter Scales

Lifetime prediction of reactor pressure vessel (RPV) steel requires bridging atomistic degradation mechanisms with service-scale spatial and temporal regimes, from Angstroms and picoseconds to meters and decades. Existing engineering-scale models provide long-range reach but rely on fitted degradation laws, while recent atomistic kinetic Monte Carlo (AKMC) advances still fail to achieve year-and-meter-scale coverage. We present AtomWorld, an atomistic world-modeling framework for RPV steel lifetime simulation co-designed with leadership-scale supercomputing through three tightly coupled layers: (1) algorithm: AtomWorld recasts classical AKMC as an atomistic world model that learns consequence-aware state transitions over the ab initio energy landscape; (2) HPC: it co-designs this formulation with modern supercomputers, yielding a compute-dense, synchronization-light, and communication-efficient execution pipeline; and (3) application: it extends atomistic world modeling to engineering-scale simulation through a physically grounded voxel-parallel framework, offering a scalable pathway from local atomistic dynamics to engineering-scale degradation evolution. We demonstrate a paradigm shift in atomistic simulation: AtomWorld enables atomistic simulation of RPV steel across year-and-meter scales for the first time, extending direct atomistic modeling to ten-quintillion-atom systems and achieving a time-to-solution of 1.71 days for one simulated service year. These capabilities are sustained across five leadership supercomputers with 92-97% scaling efficiency and peak performance up to 1.27 EFLOP/s, corresponding to 48% of the Lineshine peak FP64 performance.

cs.DC↗