arXiv subjects

Cheng Liu

Publications and source records attributed to Cheng Liu.

At least 91 records · Page 5Linked to original sources

arXiv2025

Visualizing Nanodomain Superlattices in Halide Perovskites Giving Picosecond Quantum Transients

The high optoelectronic quality of halide perovskites lends them to be utilized in optoelectronic devices and recently in emerging quantum emission applications. Advancements in perovskite nanomaterials have led to the discovery of processes in which luminescence decay times are sub-100 picoseconds, stimulating the exploration of even faster radiative rates for advanced quantum applications, which have only been prominently realised in III-V materials grown through costly epitaxial growth methods. Here, we discovered ultrafast quantum transients of time scales ~2 picoseconds at low temperature in bulk formamidinium lead iodide films grown through scalable solution or vapour approaches. Using a multimodal strategy, combining ultrafast spectroscopy, optical and electron microscopy, we show that these transients originate from quantum tunnelling in nanodomain superlattices. The outcome of the transient decays, photoluminescence, mirrors the photoabsorption of the states, with an ultra-narrow linewidth at low temperature as low as <2 nm (~4 meV). Localized correlation of the emission and structure reveals that the nanodomain superlattices are formed by alternating ordered layers of corner sharing and face sharing octahedra. This discovery opens new applications leveraging intrinsic quantum properties and demonstrates powerful multimodal approaches for quantum investigations.

cond-mat.mtrl-sci↗

arXiv2025

S$^2$R: Teaching LLMs to Self-verify and Self-correct via Reinforcement Learning

Recent studies have demonstrated the effectiveness of LLM test-time scaling. However, existing approaches to incentivize LLMs' deep thinking abilities generally require large-scale data or significant training efforts. Meanwhile, it remains unclear how to improve the thinking abilities of less powerful base models. In this work, we introduce S$^2$R, an efficient framework that enhances LLM reasoning by teaching models to self-verify and self-correct during inference. Specifically, we first initialize LLMs with iterative self-verification and self-correction behaviors through supervised fine-tuning on carefully curated data. The self-verification and self-correction skills are then further strengthened by both outcome-level and process-level reinforcement learning, with minimized resource requirements, enabling the model to adaptively refine its reasoning process during inference. Our results demonstrate that, with only 3.1k self-verifying and self-correcting behavior initialization samples, Qwen2.5-math-7B achieves an accuracy improvement from 51.0\% to 81.6\%, outperforming models trained on an equivalent amount of long-CoT distilled data. Extensive experiments and analysis based on three base models across both in-domain and out-of-domain benchmarks validate the effectiveness of S$^2$R. Our code and data are available at https://github.com/NineAbyss/S2R.

cs.CL↗

arXiv2025

Controlling the superradiant phase transition in the quantum Rabi model

In the ultrastrong-coupling regime, the quantum Rabi model can exhibit quantum phase transition (QPT) when the ratio of the qubit transition frequency to the frequency of the cavity field approaches infinity. However, it is challenging to control the QPT in few-body systems because of the limited coupling strength and the A^2 term. Here, we propose a practical scheme to manipulate the QPT of the quantum Rabi model in the strong-coupling regime. By applying a periodic frequency modulation to the two-level system in a quantum Rabi model in the strong-coupling regime, an anisotropic quantum Rabi model with ultrastrong and tunable coupling strengths for rotating and counterrotating terms is obtained. The ground state and excitation energy of this model, in terms of the modulation parameters, are studied. We find that the QPT of the quantum Rabi model can be observed in the strong-coupling regime and externally controlled by the modulation.

quant-ph↗

arXiv2025

MakeAnything: Harnessing Diffusion Transformers for Multi-Domain Procedural Sequence Generation

A hallmark of human intelligence is the ability to create complex artifacts through structured multi-step processes. Generating procedural tutorials with AI is a longstanding but challenging goal, facing three key obstacles: (1) scarcity of multi-task procedural datasets, (2) maintaining logical continuity and visual consistency between steps, and (3) generalizing across multiple domains. To address these challenges, we propose a multi-domain dataset covering 21 tasks with over 24,000 procedural sequences. Building upon this foundation, we introduce MakeAnything, a framework based on the diffusion transformer (DIT), which leverages fine-tuning to activate the in-context capabilities of DIT for generating consistent procedural sequences. We introduce asymmetric low-rank adaptation (LoRA) for image generation, which balances generalization capabilities and task-specific performance by freezing encoder parameters while adaptively tuning decoder layers. Additionally, our ReCraft model enables image-to-process generation through spatiotemporal consistency constraints, allowing static images to be decomposed into plausible creation sequences. Extensive experiments demonstrate that MakeAnything surpasses existing methods, setting new performance benchmarks for procedural generation tasks.

cs.CV↗

arXiv2025

Sequential Multi-objective Multi-agent Reinforcement Learning Approach for Predictive Maintenance

Existing predictive maintenance (PdM) methods typically focus solely on whether to replace system components without considering the costs incurred by inspection. However, a well-considered approach should be able to minimize Remaining Useful Life (RUL) at engine replacement while maximizing inspection interval. To achieve this, multi-agent reinforcement learning (MARL) can be introduced. However, due to the sequential and mutually constraining nature of these 2 objectives, conventional MARL is not applicable. Therefore, this paper introduces a novel framework and develops a Sequential Multi-objective Multi-agent Proximal Policy Optimization (SMOMA-PPO) algorithm. Furthermore, to provide comprehensive and effective degradation information to RL agents, we also employed Gated Recurrent Unit, quantile regression, and probability distribution fitting to develop a GRU-based RUL Prediction (GRP) model. Experiments demonstrate that the GRP method significantly improves the accuracy of RUL predictions in the later stages of system operation compared to existing methods. When incorporating its output into SMOMA-PPO, we achieve at least a 15% reduction in average RUL without unscheduled replacements (UR), nearly a 10% increase in inspection interval, and an overall decrease in maintenance costs. Importantly, our approach offers a new perspective for addressing multi-objective maintenance planning with sequential constraints, effectively enhancing system reliability and reducing maintenance expenses.

eess.SY↗

arXiv2025

Low-temperature magnetic behaviour on the triangular lattice in hexagonal Ba$_3$Tb(BO$_3$)$_3$

The hexagonal polymorph of Ba$_3$Tb(BO$_3$)$_3$ contains Tb$^{3+}$ ions on a quasi-2D triangular lattice, resulting in geometric magnetic frustration. Powder samples of Ba$_3$Tb(BO$_3$)$_3$ have been investigated using specific heat, powder neutron diffraction (PND), inelastic neutron scattering (INS) and muon-spin relaxation spectroscopy ($μ$SR). No long-range magnetic ordering is observed down to the lowest measured temperatures of 75 mK in PND and specific heat data and 1.5 K in the $μ$SR data. Modelling the INS spectrum using a point charge model suggests that the ground state is a singlet with a low-lying doublet on each of the two crystallographically independent Tb$^{3+}$ sites and that both the Tb ions display weak XY single-ion anisotropy.

cond-mat.str-el↗

arXiv2025

Magnetoelastic coupling in the stretched diamond lattice of TbTaO$_4$

The magnetic structure of diamond-like lattice has been studied extensively in terms of the magnetic frustration. Here we report the distortion of stretched diamond lattice of Tb$^{3+}$ (4$f^8$) in M-TbTaO$_4$ on application of a magnetic field. We have investigated the structural and magnetic properties of M phase terbium tantalate M-TbTaO$_4$ as a function of temperature and magnetic field using magnetometry and powder neutron diffraction. Sharp $λ$-shape transitions in $d(χT)/dT$, $dM/dH$ and specific heat data confirm the previously reported three-dimensional (3D) antiferromagnetic ordering at $T_N \approx 2.25$ K. On application of a magnetic field the Néel temperature is found to decrease and variable field neutron diffraction experiments below $T_N$ at 1.6 K show an increase in both the bond and angle distortion of the stretched diamond lattice with magnetic field, indicating a potential magneto-elastic coupling effect. By combining our magnetometry, heat capacity and neutron diffraction results we generate a magnetic phase diagram for M-TbTaO$_4$ as a function of temperature and field.

cond-mat.str-el↗

arXiv2025

LUT-DLA: Lookup Table as Efficient Extreme Low-Bit Deep Learning Accelerator

The emergence of neural network capabilities invariably leads to a significant surge in computational demands due to expanding model sizes and increased computational complexity. To reduce model size and lower inference costs, recent research has focused on simplifying models and designing hardware accelerators using low-bit quantization. However, due to numerical representation limits, scalar quantization cannot reduce bit width lower than 1-bit, diminishing its benefits. To break through these limitations, we introduce LUT-DLA, a Look-Up Table (LUT) Deep Learning Accelerator Framework that utilizes vector quantization to convert neural network models into LUTs, achieving extreme low-bit quantization. The LUT-DLA framework facilitates efficient and cost-effective hardware accelerator designs and supports the LUTBoost algorithm, which helps to transform various DNN models into LUT-based models via multistage training, drastically cutting both computational and hardware overhead. Additionally, through co-design space exploration, LUT-DLA assesses the impact of various model and hardware parameters to fine-tune hardware configurations for different application scenarios, optimizing performance and efficiency. Our comprehensive experiments show that LUT-DLA achieves improvements in power efficiency and area efficiency with gains of $1.4$~$7.0\times$ and $1.5$~$146.1\times$, respectively, while maintaining only a modest accuracy drop. For CNNs, accuracy decreases by $0.1\%$~$3.1\%$ using the $L_2$ distance similarity, $0.1\%$~$3.4\%$ with the $L_1$ distance similarity, and $0.1\%$~$3.8\%$ when employing the Chebyshev distance similarity. For transformer-based models, the accuracy drop ranges from $1.4\%$ to $3.0\%$.

cs.AR↗

arXiv2025

Rational map associated with the Sigmoid Beverton-Holt model on the projective line over $\mathbb{Q}_p$

We describe the dynamical structure of the $p$-adic rational dynamical systems associated with the Sigmoid Beverton-Holt model on the projective line over the field $\mathbb{Q}_p$ of $p$-adic numbers. Our methods are minimal decomposition of $p$-adic polynomials with coefficients in $\mathbb{Z}_p$ established by Fan and Liao and the chaotic description of $p$-adic repellers of Fan, Liao, Wang and Zhou.

math.DS↗

arXiv2024

Graph Coarsening via Supervised Granular-Ball for Scalable Graph Neural Network Training

Graph Neural Networks (GNNs) have demonstrated significant achievements in processing graph data, yet scalability remains a substantial challenge. To address this, numerous graph coarsening methods have been developed. However, most existing coarsening methods are training-dependent, leading to lower efficiency, and they all require a predefined coarsening rate, lacking an adaptive approach. In this paper, we employ granular-ball computing to effectively compress graph data. We construct a coarsened graph network by iteratively splitting the graph into granular-balls based on a purity threshold and using these granular-balls as super vertices. This granulation process significantly reduces the size of the original graph, thereby greatly enhancing the training efficiency and scalability of GNNs. Additionally, our algorithm can adaptively perform splitting without requiring a predefined coarsening rate. Experimental results demonstrate that our method achieves accuracy comparable to training on the original graph. Noise injection experiments further indicate that our method exhibits robust performance. Moreover, our approach can reduce the graph size by up to 20 times without compromising test accuracy, substantially enhancing the scalability of GNNs.

cs.LG↗

arXiv2024

Minute-cadence Observations of the LAMOST Fields with the TMTS: VI. Absolute Physical Parameters of Contact Binaries

With the development of wide-field surveys, a large amount of data on short-period W UMa contact binaries have been obtained. Continuous and uninterrupted light curves as well as high-resolution spectroscopic data are crucial in determining the absolute physical parameters. Targets with both TMTS light curves and LAMOST medium-resolution spectra were selected. The absolute physical parameters were inferred with the W-D code for ten systems, all of them are W-type shallow or medium contact binaries. The O'Connell effect observed in the light curves can be explained by adding a spot on the primary or secondary component in the models. According to O-C analysis, the orbital periods exhibit a long-term increasing or decreasing trend, amongst which J0132, J1300, and J1402 show periodic variations that may be attributed to the presence of a third body or magnetic activity cycles. Spectral subtraction analysis revealed that the equivalent width of H$α$ indicates strong magnetic activity in J0047, J0305, J0638, and J1402. Among the 10 selected binary systems, except for J0132 and J0913, the more massive components are found to be main-sequence stars while the less massive components have evolved off the main sequence. In J0132, both components are in the main sequence, whereas both components of J0913 lie above the terminal-age main sequence. Based on the relationship between orbital angular momentum and total mass for these two systems, as well as their low fill-out factors, it is possible that these two systems are newly formed contact binaries, having recently evolved from the detached configuration.

astro-ph.SR↗

arXiv2024

Receiver-device-independent quantum secure direct communication

Quantum secure direct communication (QSDC) enables the message sender to directly send secure messages to the receiver through the quantum channel without keys. Device-independent (DI) and measurement-device-independent (MDI) QSDC protocols can enhance QSDC's practical security in theory. DI QSDC requires extremely high global detection efficiency and has quite low secure communication distance. DI and MDI QSDC both require high-quality entanglement. Current entanglement sources prepare entangled photon pairs with low efficiency, largely reducing their practical communication efficiency. In the paper, we propose a single-photon-based receiver-device-independent (RDI) QSDC protocol. It only relies on the trusted single-photon source, which is nearly on-demand under current technology, and treats all the receiving devices in both communication parties as ``black-boxes''. The parties ensure the message security only from the observed statistics. We develop a numerical method to simulate its performance in practical noisy communication situation. RDI QSDC provides the same security level as MDI QSDC. Compared with DI and MDI QSDC, RDI QSDC has some advantages. First, it uses the single-photon source and single-photon measurement, which makes it obtain the practical communication efficiency about 3415 times of that in DI QSDC and easy to implement. The whole protocol is feasible with current technology. Second, it has higher photon loss robustness and noise tolerance than DI QSDC, which enables it to have a secure communication distance about 26 times of that in DI QSDC. Based on above features, the RDI QSDC protocol makes it possible to achieve highly-secure and high-efficient QSDC in the near future.

quant-ph↗

arXiv2024

The LAMOST Spectroscopic Survey of Supergiants in M31 and M33

We present systematic identifications of supergiants of M31/M33 based on massive LAMOST spectroscopic survey. Radial velocities of nearly 5000 photometrically selected M31/M33 supergiant candidates have been properly derived from the qualified spectra released in LAMOST DR10. By comparing their radial velocities with those predicted from the rotation curves of M31, as well as utilizing {\it Gaia} astrometric measurements to exclude foreground contaminations, 199 supergiant members in M31, including 168 `Rank1' and 31 `Rank2', have been successfully identified. This sample contains 62 blue supergiants (BSGs, all `Rank1'), 134 yellow supergiants (YSGs, 103 `Rank1' and 31 `Rank2') and 3 red supergiants (RSGs, all `Rank1'). For M33, we identify 84 supergiant members (56 `Rank1' and 28 `Rank2'), which includes 28 BSGs (all `Rank1'), 53 YSGs (25 `Rank1' and 28 `Rank2') and 3 RSGs (all `Rank1'). So far, this is one of the largest supergiant sample of M31/M33 with full optical wavelength coverage (3700 \textless $λ$ \textless 9100 Å). This sample is valuable for understanding the star formation and stellar evolution under different environments.

astro-ph.GA↗

arXiv2024

Enabling Energy-Efficient Deployment of Large Language Models on Memristor Crossbar: A Synergy of Large and Small

Large language models (LLMs) have garnered substantial attention due to their promising applications in diverse domains. Nevertheless, the increasing size of LLMs comes with a significant surge in the computational requirements for training and deployment. Memristor crossbars have emerged as a promising solution, which demonstrated a small footprint and remarkably high energy efficiency in computer vision (CV) models. Memristors possess higher density compared to conventional memory technologies, making them highly suitable for effectively managing the extreme model size associated with LLMs. However, deploying LLMs on memristor crossbars faces three major challenges. Firstly, the size of LLMs increases rapidly, already surpassing the capabilities of state-of-the-art memristor chips. Secondly, LLMs often incorporate multi-head attention blocks, which involve non-weight stationary multiplications that traditional memristor crossbars cannot support. Third, while memristor crossbars excel at performing linear operations, they are not capable of executing complex nonlinear operations in LLM such as softmax and layer normalization. To address these challenges, we present a novel architecture for the memristor crossbar that enables the deployment of state-of-the-art LLM on a single chip or package, eliminating the energy and time inefficiencies associated with off-chip communication. Our testing on BERT_Large showed negligible accuracy loss. Compared to traditional memristor crossbars, our architecture achieves enhancements of up to 39X in area overhead and 18X in energy consumption. Compared to modern TPU/GPU systems, our architecture demonstrates at least a 68X reduction in the area-delay product and a significant 69% energy consumption reduction.

cs.AI↗

arXiv2024

Human Stone Toolmaking Action Grammar (HSTAG): A Challenging Benchmark for Fine-grained Motor Behavior Recognition

Action recognition has witnessed the development of a growing number of novel algorithms and datasets in the past decade. However, the majority of public benchmarks were constructed around activities of daily living and annotated at a rather coarse-grained level, which lacks diversity in domain-specific datasets, especially for rarely seen domains. In this paper, we introduced Human Stone Toolmaking Action Grammar (HSTAG), a meticulously annotated video dataset showcasing previously undocumented stone toolmaking behaviors, which can be used for investigating the applications of advanced artificial intelligence techniques in understanding a rapid succession of complex interactions between two hand-held objects. HSTAG consists of 18,739 video clips that record 4.5 hours of experts' activities in stone toolmaking. Its unique features include (i) brief action durations and frequent transitions, mirroring the rapid changes inherent in many motor behaviors; (ii) multiple angles of view and switches among multiple tools, increasing intra-class variability; (iii) unbalanced class distributions and high similarity among different action sequences, adding difficulty in capturing distinct patterns for each action. Several mainstream action recognition models are used to conduct experimental analysis, which showcases the challenges and uniqueness of HSTAG https://nyu.databrary.org/volume/1697.

cs.CV↗

arXiv2024

Ordering-Based Causal Discovery for Linear and Nonlinear Relations

Identifying causal relations from purely observational data typically requires additional assumptions on relations and/or noise. Most current methods restrict their analysis to datasets that are assumed to have pure linear or nonlinear relations, which is often not reflective of real-world datasets that contain a combination of both. This paper presents CaPS, an ordering-based causal discovery algorithm that effectively handles linear and nonlinear relations. CaPS introduces a novel identification criterion for topological ordering and incorporates the concept of "parent score" during the post-processing optimization stage. These scores quantify the strength of the average causal effect, helping to accelerate the pruning process and correct inaccurate predictions in the pruning step. Experimental results demonstrate that our proposed solutions outperform state-of-the-art baselines on synthetic data with varying ratios of linear and nonlinear relations. The results obtained from real-world data also support the competitiveness of CaPS. Code and datasets are available at https://github.com/E2real/CaPS.

cs.LG↗

arXiv2024

Remaining Useful Life Prediction: A Study on Multidimensional Industrial Signal Processing and Efficient Transfer Learning Based on Large Language Models

Remaining useful life (RUL) prediction is crucial for maintaining modern industrial systems, where equipment reliability and operational safety are paramount. Traditional methods, based on small-scale deep learning or physical/statistical models, often struggle with complex, multidimensional sensor data and varying operating conditions, limiting their generalization capabilities. To address these challenges, this paper introduces an innovative regression framework utilizing large language models (LLMs) for RUL prediction. By leveraging the modeling power of LLMs pre-trained on corpus data, the proposed model can effectively capture complex temporal dependencies and improve prediction accuracy. Extensive experiments on the Turbofan engine's RUL prediction task show that the proposed model surpasses state-of-the-art (SOTA) methods on the challenging FD002 and FD004 subsets and achieves near-SOTA results on the other subsets. Notably, different from previous research, our framework uses the same sliding window length and all sensor signals for all subsets, demonstrating strong consistency and generalization. Moreover, transfer learning experiments reveal that with minimal target domain data for fine-tuning, the model outperforms SOTA methods trained on full target domain data. This research highlights the significant potential of LLMs in industrial signal processing and RUL prediction, offering a forward-looking solution for health management in future intelligent industrial systems.

cs.LG↗

arXiv2024

Whale Detection Enhancement through Synthetic Satellite Images

With a number of marine populations in rapid decline, collecting and analyzing data about marine populations has become increasingly important to develop effective conservation policies for a wide range of marine animals, including whales. Modern computer vision algorithms allow us to detect whales in images in a wide range of domains, further speeding up and enhancing the monitoring process. However, these algorithms heavily rely on large training datasets, which are challenging and time-consuming to collect particularly in marine or aquatic environments. Recent advances in AI however have made it possible to synthetically create datasets for training machine learning algorithms, thus enabling new solutions that were not possible before. In this work, we present a solution - SeaDroneSim2 benchmark suite, which addresses this challenge by generating aerial, and satellite synthetic image datasets to improve the detection of whales and reduce the effort required for training data collection. We show that we can achieve a 15% performance boost on whale detection compared to using the real data alone for training, by augmenting a 10% real data. We open source both the code of the simulation platform SeaDroneSim2 and the dataset generated through it.

cs.CV↗