arXiv subjects

Liang Luo

Publications and source records attributed to Liang Luo.

At least 19 recordsLinked to original sources

arXiv2026

Learning Spin Hamiltonians from Terahertz Two-Dimensional Coherent Spectroscopy

Effective Hamiltonians connect microscopic interactions to measurable collective behavior in quantum materials, but determining their parameters directly from experiment remains a challenging inverse problem. We introduce a supervised machine-learning framework that infers Hamiltonian parameters from nonlinear terahertz two-dimensional coherent spectra. A calibrated forward model generates spectra from candidate Hamiltonians, a common preprocessing pipeline maps simulated and experimental spectra into the same representation, and a neural network learns the inverse map from spectral fingerprints to microscopic parameters. We demonstrate the approach for rare-earth orthoferrites using a two-sublattice Landau--Lifshitz--Gilbert spin model with exchange, Dzyaloshinskii--Moriya interaction, anisotropies, and damping. Synthetic benchmarks show that nonlinear spectra encode parameters beyond those fixed by the linear response, with inference accuracy tracking the physical spectral sensitivity and robustness against noise improved by using multiple inter-pulse delays. Applied to experimental THz-2DCS data from Sm$_{0.4}$Er$_{0.6}$FeO$_3$, the inferred parameters yield physically reasonable forward simulations, while remaining discrepancies identify limitations of the reduced model. These results establish THz-2DCS as a data-rich platform for effective-Hamiltonian inference and model refinement, enabling experimentally driven identification of microscopic interactions while providing a foundation for understanding, predicting, and ultimately controlling the emergent properties of quantum materials.

physics.comp-ph↗

arXiv2026

MARCO: Click-Intent Decomposition for Calibrated Ads Conversion Prediction

Not all clicks are equal. Industrial ads ranking decouples conversion probability into click-through rate (CTR) and post-click conversion rate (CVR), yet treats every click as the same event. In reality, users provide a free, self-generated signal of intent through their physical UI interactions. Different click types on the same ad exhibit a 4-fold difference in actual conversion rates. By conflating these signals, the standard CVR model under-predicts high-intent clicks and over-predicts low-intent ones, which is a bias masked by near-perfect aggregate calibration. We propose MARCO (Multi-intent Ads Ranking Composition Optimization), a framework that resolves this bias by decomposing each click by intent. Using the logged click type as a free behavioral label, MARCO trains per-intent CVR heads on homogeneous populations, and at serving time composes their per-intent CVR estimates under a predicted distribution over intents. Theoretically, we prove that decomposition never raises population risk, give the exact headroom under squared loss and non-negativity under the deployed loss, and show through a routing-efficiency dial how much of it reaches serving. Because the population-optimal score is unchanged, any gain is a finite-capacity estimation and calibration effect that we validated both offline and online. For deployment at scale, we further cast multi-impression, multi-click attribution as credit assignment with a bias-variance tradeoff analogous to RL return estimation, showing last-impression, first-click attribution is the low-bias, low-variance, deterministic choice under production constraints, and derive three consistency conditions enforced end-to-end at scale. Deployed at binary intent granularity, MARCO corrects per-intent calibration to approximately 100%, lifts conversions per click by +2.80%, and drives +0.98% cumulative improvement in topline metrics.

cs.LG↗

arXiv2026

Catalytic Stabilization of Ergotropy and Backflow Suppression in Open Many-Body Quantum Batteries

Coherent energy backflow and non-Markovian oscillations limit energy retention and degrade extractable work (ergotropy) in open many-body quantum batteries. Here, we present a catalyst-mediated charging protocol for a collective spin-array quantum battery coupled to a laser-driven charger. Using the open-system Lindblad master equation, we examine the energy transfer dynamics when both charger and battery are symmetrically coupled to an off-resonant auxiliary catalytic mode. Numerical simulations reveal that while unassisted bipartite setups exhibit pronounced backflow oscillations and poor energy retention, catalytic mediation quenches transient oscillations and accelerates energy injection. The auxiliary system operates as an energy-invariant conduit, maintaining a constant energy expectation value $\langle H_C(t)\rangle \approx \langle H_C(0)\rangle$ and negligible transient population throughout the evolution. Microscopically, virtual excitations of the catalyst generate an effective complex inter-subsystem coupling $J_{\text{eff}}$, which induces an underdamped-to-overdamped dynamical crossover and introduces selective coherence damping. This mechanism prevents population depletion in the battery, stabilizing the population inversion and significantly increasing the asymptotic steady-state ergotropy with increasing battery size $N_B$. These findings clarify the dissipative dynamics of catalyst-mediated energy transfer and provide a practical scheme for improving storage stability in modern quantum hardware platforms.

quant-ph↗

arXiv2026

NxN E-valuation: Hypothesis Certification via a Conformal CRT Null

We propose NxN E-valuation, a handy, e-value-based hypothesis-certification algorithm that lets a hypothesis be verified without building any case-specific certification procedure---such as constructing a dedicated null hypothesis---as long as a large enough dataset is available. The method is especially suited to LLM-based exploration systems, where LLMs are remarkably good at proposing hypotheses but suffer badly from hallucination; this hallucination prevents us from harvesting LLM outputs directly, and existing remedies each fall short. The most common solutions include letting the LLM verify or correct itself circular verification and held-out testing (where false hypotheses can still pass via spurious correlations), among other remedies detailed in the introduction. To resolve this, NxN E-valuation exploits the naturally existing large training set and lets different samples serve as null hypotheses for one another. This design directly realizes a conditional randomization test (CRT) that certifies each hypothesis. The approach can be a universally better replacement for at least LLM circular verification and held-out-data testing, provided the LLM's generations are hypotheses that apply to each individual sample.

cs.AI↗

arXiv2026

ROCS: Request-Oriented Compute Sharing for Efficient Large-Scale Recommendation

Modern recommendation models gain prediction quality by scaling feature-interaction and sequence modules, but production cost constraints cap how far systems can scale. In this work, we propose Request-Oriented Compute Sharing (ROCS), a modeling and inference paradigm that exploits a unique property of recommendation inference: each user request is evaluated against many candidates, while request-side features are shared across candidates. ROCS defers request-candidate interactions as late as possible, isolates candidate-dependent representations, and evaluates substantial portions of the model once per request rather than once per candidate, significantly improving inference efficiency while maintaining or improving prediction quality. To realize this paradigm, we develop Generalized Layer Masking (GLM) to enforce candidate isolation in feature-interaction architectures, and Deep Cross Attention (DCA) to extend request-oriented sharing to sequence architectures. To support efficient GPU deployment, we co-design In-Kernel Broadcast Optimization (IKBO) that significantly accelerates ROCS model execution. Experiments on public benchmarks show that ROCS consistently improves the quality-efficiency tradeoff across recommendation backbones. On production-scale workloads, ROCS achieves up to a 3x QPS improvement on retrieval models without quality degradation and a 0.5% relative LogLoss improvement with a 50% QPS gain on a short-form video ranking model. ROCS has been deployed across large-scale recommendation systems spanning ads and organic surfaces, retrieval and ranking stages, and more than two orders of magnitude in inference complexity, delivering significant online gains at reduced infrastructure cost.

cs.LG↗

arXiv2026

ReasonRec: A Reasoning-Augmented Multimodal Agent for Unified Recommendation

Recent advances in multimodal recommenders excel at feature fusion but remain opaque and inefficient decision-makers, lacking explicit reasoning and self-awareness of uncertainty. We introduce ReasonRec, a reasoning-augmented multimodal agent structured around a three-stage explicit reasoning pipeline. Specifically, we propose a reasoning-aware visual instruction tuning strategy that systematically transforms diverse recommendation tasks into unified CoT prompts, enabling the VLM to explicitly articulate intermediate decision steps. Additionally, our evidence-horizon curriculum progressively enhances the reasoning complexity to better handle cold-start and long-tail user scenarios, significantly boosting model generalization. Furthermore, the uncertainty-guided delegation mechanism empowers the agent to assess its own confidence, strategically allocating computational resources to optimize both recommendation accuracy and inference efficiency. Comprehensive experiments on four standard recommendation tasks across five real-world datasets demonstrate that ReasonRec achieves over 30% relative improvement in key ranking metrics compared to state-of-the-art multimodal recommenders. Crucially, ReasonRec substantially reduces inference latency by dynamically delegating up to 35% of queries to efficient sub-models without compromising accuracy. Extensive ablation studies further confirm that each proposed reasoning and planning mechanism individually contributes substantially to ReasonRec's overall effectiveness. Collectively, our results illustrate a clear pathway towards interpretable, adaptive, and efficient multimodal recommendation through explicit reasoning and agentic design.

cs.IR↗

arXiv2026

UnpredictaBench: A Benchmark for Evaluating Distributional Randomness in LLMs

We introduce UnpredictaBench, an evaluation that tests the ability of large language models (LLMs) to capture true underlying distributions. As LLMs are increasingly used as substitutes for other entities (e.g., for humans in economic simulations), the tendency of many models to collapse towards a single plausible answer means a failure to capture the unpredictability of real systems. Recent work on improving output diversity is insufficient for this setting: simulation requires samples that are calibrated to a target distribution, not merely varied outputs. UnpredictaBench isolates a simplified but fundamental version of this problem: sampling outcomes from individual target distributions, including canonical statistical distributions, distributions induced by stochastic programs, and natural-language scenarios that describe random processes. We introduce 448 such problems together with KS@N, a general-purpose evaluation metric that quantifies how well a model outputs approximate black-box target distributions via the Kolmogorov-Smirnov statistical test. This is the rate at which we fail to reject model samples of size N against ground-truth samples, with larger N indicating greater difficulty. Tested across open and proprietary models, we find a large spread in distributional capabilities. For instance, when models generate samples of size 100 (KS@100, our standard metric), scores range from near 0 to over 20%. No model is able to achieve over 40% at KS@100, showing significant headroom in distributional sampling as a capability. Although adding reasoning can somewhat increase scores, we find no immediate solution for this issue. UnpredictaBench shows that even simple distributional simulation remains challenging, making it a necessary first step toward using LLMs as stand-ins for complex systems. Project website and resources are available at https://unpredictabenchmark.github.io/.

cs.CL↗

arXiv2026

LoKA: Low-precision Kernel Applications for Recommendation Models At Scale

Recent GPU generations deliver significantly higher FLOPs using lower-precision arithmetic, such as FP8. While successfully applied to large language models (LLMs), its adoption in large recommendation models (LRMs) has been limited. This is because LRMs are numerically sensitive, dominated by small matrix multiplications (GEMMs) followed by normalization, and trained in communication-intensive environments. Applying FP8 directly to LRMs often degrades model quality and prolongs training time. These challenges are inherent to LRM workloads and cannot be resolved merely by introducing better FP8 kernels. Instead, a system-model co-design approach is needed to successfully integrate FP8. We present LoKA (Low-precision Kernel Applications), a framework that makes FP8 practical for LRMs through three principles: profile under realistic distributions to know where low precision is safe, co-design model components with hardware to expand where it is safe, and orchestrate across kernel libraries to maximize the gains. Concretely, LoKA Probe is a statistically grounded, online benchmarking method that learns activation and weight statistics, and quantifies per-layer errors. This process pinpoints safe and unsafe, fast and slow sites for FP8 adoption. LoKA Mods is a set of reusable model adaptations that improve both numerical stability and execution efficiency with FP8. LoKA Dispatch is a runtime that leverages the statistical insights from LoKA Probe to select the fastest FP8 kernel that satisfies the accuracy requirements.

cs.LG↗

arXiv2026

Magnetic Brightening and Nanoscale Imaging of Spin-Polarized Helical Edge Modes

Efficient sub-10 nm electric transport remains a major challenge for nanoelectronics due to high losses and impedance mismatches in conventional Drude metals. Despite their promise of dissipationless, reflection-free conduction, topologically protected chiral edge modes remain little explored in their nanoscale spin polarized transport-particularly regarding real-space visualization, magnetic field tunability, and high-frequency edge conductivity. Here, we report magnetic brightening and nanoscale visualization of highly spin-polarizable infrared helical edge states using cryogenic magneto-infrared scattering-type scanning near-field optical microscopy (cm-IR-sSNOM). Our measurements reveal magnetic field-induced near-field conductivity at step edges, uncovering quantum spin Hall spin-splitting modes with enhanced infrared polarizability and slightly narrowed near-field profiles. In addition, the infrared edge electrodynamic response scales nearly linearly with atomic layer number, providing compelling evidence that magnetic-field-induced gaps do not disrupt individual-layer edge states at energies of around 100 meV. These results sharply contrast with microwave and DC transport, where even small magnetically induced gaps decrease edge conduction. Magnetically tunable, topologically robust high-frequency edge modes open a pathway toward ultralow-loss nanoscale interconnects and quantum logic architectures for next-generation microelectronics, spintronics and quantum information science.

cond-mat.mes-hall↗

arXiv2026

FreeScale: Distributed Training for Sequence Recommendation Models with Minimal Scaling Cost

Modern industrial Deep Learning Recommendation Models typically extract user preferences through the analysis of sequential interaction histories, subsequently generating predictions based on these derived interests. The inherent heterogeneity in data characteristics frequently result in substantial under-utilization of computational resources during large-scale training, primarily due to computational bubbles caused by severe stragglers and slow blocking communications. This paper introduces FreeScale, a solution designed to (1) mitigate the straggler problem through meticulously load balanced input samples (2) minimize the blocking communication by overlapping prioritized embedding communications with computations (3) resolve the GPU resource competition during computation and communication overlapping by communicating through SM-Free techniques. Empirical evaluation demonstrates that FreeScale achieves up to 90.3% reduction in computational bubbles when applied to real-world workloads running on 256 H100 GPUs.

cs.LG↗

arXiv2026

Ultrafast Magneto-Pressure Spectroscopy and Control of Correlated Phases in a Trilayer Nickelate

Ultrafast spectroscopy under simultaneous high pressure and magnetic field provides a versatile approach for investigating pressure-driven electronic instabilities and correlated phases, and for probing potential bulk superconducting behavior under extreme conditions. However, such an experimental platform has yet to be implemented, standing as a roadblock to a fuller understanding of nonequilibrium superconductivity and vortex-controlled quasi-particle (QP) dynamics. Here, we bridge this capability gap by developing high pressure (up to 40 GPa), high magnetic field (up to 7 T), cryogenic (down to 5 K) femtosecond spectroscopy, and using it to probe magneto-pressure evolution of quasiparticle dynamics in the trilayer nickelate $\mathrm{Pr}_4\mathrm{Ni}_3\mathrm{O}_{10}$. We observe pronounced critical slowing down of QP relaxation at the charge-density-wave transition, which collapses under applied pressure. At higher pressures, the relaxation instead lengthens at low temperature, consistent with incipient superconducting correlations. However, the negligibel magnetic-field-dependence up to 7~T and absence of vortex-induced pre-bottleneck dynamics--robust signatures observed in our controlled bulk superconducting samples--indicates that any superconducting state under the present pressure conditions is likely non-bulk, filamentary, or strongly inhomogeneous. The magneto-pressure ultrafast capability opens a new avenue for resolving outstanding questions surrounding pressure-induced superconductivity and intertwined orders in correlated quantum materials.

cond-mat.supr-con↗

arXiv2026

Extreme Terahertz Nonlinear Phononics by Coherence-Imprinted Control of Hybrid Order

Coherent control of quantum materials has progressed along two major fronts: nonlinear phononics, which reshapes lattices to induce emergent states, and Floquet engineering, which tailors electronic band reconstruction via time-periodic driving. Both mechanisms face fundamental limitations at terahertz (THz) frequencies: phononic nonlinearities are intrinsically weak in standard lattices, while electronic Floquet states are often constrained by rapid decoherence upon light-off and by a scarcity of coherence-resolved, multi-correlation probes beyond (quasi-)stationary band structures. Here we report an extreme THz nonlinear-phononics mechanism in $\text{Ta}_\text{2}\text{NiSe}_\text{5}$, where a highly susceptible non-equilibrium electronic correlation bath dramatically amplifies lattice nonlinearities under coherent driving. Utilizing THz two-dimensional spectroscopy as a coherence-tomography tool, we resolve an exceptionally rich landscape of approximately 30 distinct multi-order quantum pathways, including high-harmonic phonon generation, multi-quantum coherences, and multi-wave anharmonic cross-mode mixing. The density and complexity of this extreme manifold establishes a new benchmark for THz nonlinear phononics, as the multi-order quantum pathways surpass the limits of conventional lattice responses. These high-order signals collapse above ~100~K, defining an electronic correlation scale of a coherence-imprinted hybrid electronic-phonon order that governs the sustainability of high-order quantum correlations and nonlinear pathways beyond linear and equilibrium responses. Our results establish a route for correlation-boosted, phonon-anchored periodic Hamiltonian engineering and for certifying such periodically-driven states via multi-correlation coherence tomography.

cond-mat.str-el↗

arXiv2026

SOLARIS: Speculative Offloading of Latent-bAsed Representation for Inference Scaling

Recent advances in recommendation scaling laws have led to foundation models of unprecedented complexity. While these models offer superior performance, their computational demands make real-time serving impractical, often forcing practitioners to rely on knowledge distillation-compromising serving quality for efficiency. To address this challenge, we present SOLARIS (Speculative Offloading of Latent-bAsed Representation for Inference Scaling), a novel framework inspired by speculative decoding. SOLARIS proactively precomputes user-item interaction embeddings by predicting which user-item pairs are likely to appear in future requests, and asynchronously generating their foundation model representations ahead of time. This approach decouples the costly foundation model inference from the latency-critical serving path, enabling real-time knowledge transfer from models previously considered too expensive for online use. Deployed across Meta's advertising system serving billions of daily requests, SOLARIS achieves 0.67% revenue-driving top-line metrics gain, demonstrating its effectiveness at scale.

cs.LG↗

arXiv2026

Lamb-shift-mediated energy transfer in open quantum batteries

Open quantum batteries (QBs) operate under unavoidable system--environment interactions, where both dissipation and coherent frequency renormalization can affect their dynamics. While dissipative effects have been extensively studied, the role of environment-induced frequency shifts, such as the Lamb shift, remains less explored. Here, we investigate a driven open QB consisting of two coherently coupled quantum harmonic oscillators representing the charger and the battery. By incorporating dissipation and Lamb-shift corrections within a Lindblad master equation, we show that the Lamb shift renormalizes the system eigenfrequencies and modifies the resonance condition with the external drive. We further demonstrate that the resulting frequency renormalization leads to a mode-selective energy-transfer process, producing a redistribution of energy between the charger and the battery. This behavior is characterized through a supermode decomposition of the coupled system, which reveals how the environment-induced frequency shift alters the dominant energy-transfer channel. Our results clarify the role of coherent environmental effects in open quantum batteries and provide a physical framework for understanding work-extraction dynamics beyond purely dissipative descriptions.

quant-ph↗

arXiv2026

Implicit Turn-Wise Policy Optimization for Proactive User-LLM Interaction

Multi-turn human-AI collaboration is fundamental to deploying interactive services such as adaptive tutoring, conversational recommendation, and professional consultation. However, optimizing these interactions via reinforcement learning is hindered by the sparsity of verifiable intermediate rewards and the high stochasticity of user responses. To address these challenges, we introduce Implicit Turn-wise Policy Optimization (ITPO). ITPO leverages an implicit process reward model to derive fine-grained, turn-wise process rewards from sparse outcome signals. Unlike volatile token-level rewards, these turn-level signals exhibit superior robustness and may utilize a normalization mechanism to further enhance training stability. We evaluate ITPO across three representative multi-turn collaborative tasks: math tutoring, document writing, and medical recommendation. Empirical results demonstrate that ITPO, when combined with PPO, GRPO, or RLOO, consistently achieves improved convergence than existing baselines. Elaborate trajectory analysis confirms that ITPO infers turn-wise preferences that are semantically aligned with human judgment. Code is publicly available at https://github.com/Graph-COM/ITPO.

cs.LG↗

arXiv2026

AuroraRL: Fast, Fault-Tolerant, and Cost-Efficient Reinforcement Learning over Decentralized Network

LLM reinforcement learning (RL) requires frequent synchronization of large model parameters between the trainer and distributed rollout actors. High-throughput RL post-training therefore relies on dedicated RDMA HPC/cloud clusters, an infrastructure cost most organizations cannot absorb. A natural alternative is to aggregate loosely-coupled GPUs over standard Ethernet and WAN links, but this commodity connectivity cannot sustain full-weight broadcasts: synchronizing an 8B model can take over 100~seconds on bandwidth-limited links, while rollout generation typically takes tens of seconds. Toward making RL practical in this regime, we observe that RL fine-tuning yields highly sparse per-step updates, with only around 1\% of parameter elements changing. On top of this insight, we present AuroraRL, a novel high-performance RL training system that preserves bit-exact updates without dropping or quantizing information, designed for commodity-networked, loosely-coupled GPU resources. AuroraRL represents each step as a sparse delta checkpoint, pipelines delta extraction with multi-stream transmission, overlaps transfer with rollout generation, and coordinates heterogeneous workers with throughput- and bandwidth-aware scheduling plus lease-based fault tolerance. Across Qwen3 4B--14B models deployed in up to four geographic regions, AuroraRL shrinks per-step weight transfer by 79$\times$ on Qwen3-8B, delivers 1.3--9.5$\times$ higher throughput than dense-broadcast baselines (PrimeRL-Full, async-tolerant, multi-stream variants), and brings end-to-end training within 8.91\% of an ideal RDMA single-datacenter baseline, while transparently tolerating common failures and preserving training accuracy. By leveraging on-demand, cross-cloud GPUs over commodity links, AuroraRL delivers 1.21--1.59$\times$ higher tokens per dollar than reserved RDMA clusters at comparable throughput.

cs.DC↗

arXiv2025

Meta Lattice: Model Space Redesign for Cost-Effective Industry-Scale Ads Recommendations

The rapidly evolving landscape of products, surfaces, policies, and regulations poses significant challenges for deploying state-of-the-art recommendation models at industry scale, primarily due to data fragmentation across domains and escalating infrastructure costs that hinder sustained quality improvements. To address this challenge, we propose Lattice, a recommendation framework centered around model space redesign that extends Multi-Domain, Multi-Objective (MDMO) learning beyond models and learning objectives. Lattice addresses these challenges through a comprehensive model space redesign that combines cross-domain knowledge sharing, data consolidation, model unification, distillation, and system optimizations to achieve significant improvements in both quality and cost-efficiency. Our deployment of Lattice at Meta has resulted in 10% revenue-driving top-line metrics gain, 11.5% user satisfaction improvement, 6% boost in conversion rate, with 20% capacity saving.

cs.IR↗

arXiv2025

A Semi-Convergent Stage-Wise Framework with Provable Global Convergence for Adaptive Total Variation Regularization

Image restoration requires a careful balance between noise suppression and structure preservation. While first-order total variation (TV) regularization effectively preserves edges, it often introduces staircase artifacts, whereas higher-order TV removes such artifacts but oversmooths fine details. To reconcile these competing effects, we propose a semi-convergent stage-wise framework that sequentially integrates first- and higher-order TV regularizers within an iterative restoration process implemented via ADMM. Each stage exhibits semi-convergence behavior, i.e., the iterates initially approach the ground truth before being degraded by over-regularization. By monitoring this evolution, the algorithm adaptively selects the locally optimal iterate (e.g., with the highest PSNR) and propagates it as the initial point for the next stage. This select-and-propagate mechanism effectively transfers local semi-convergence into a globally convergent iterative process. We establish theoretical guarantees showing that the sequence of stage-wise iterates is bounded, the objective values decrease monotonically. Extensive numerical experiments on denoising and deblurring benchmarks confirm that the proposed method achieves superior quantitative and perceptual performance compared with conventional first-, higher-order, hybrid TV methods, and learning based methods, while maintaining theoretical interpretability and algorithmic simplicity.

math.NA↗