arXiv subjects

Jianming Liu

Publications and source records attributed to Jianming Liu.

17 recordsLinked to original sources

arXiv2026

HadAgent: Harness-Aware Decentralized Agentic AI Serving with Proof-of-Inference Blockchain Consensus

Proof-of-Work (PoW) blockchain consensus consumes vast computational resources without producing useful output, while the rapid growth of large language model (LLM) agents has created unprecedented demand for GPU computation. We present HadAgent, a decentralized agentic AI serving system that replaces hash-based mining with Proof-of-Inference (PoI), a consensus mechanism in which nodes earn block-creation rights by executing deterministic LLM inference tasks. Because verification requires only re-executing a single forward pass under identical conditions, cross-node verification operates at consensus speed. HadAgent organizes validated records into a three-lane block body with dedicated DATA, MODEL, and PROOF channels, each protected by an independent Merkle root for fine-grained tamper detection. A two-tier node architecture classifies secondary nodes as trusted or non-trusted based on historical behavior: trusted nodes serve inference results in real time through optimistic execution, while non-trusted nodes must undergo full consensus verification. A harness layer monitors node behavior through heartbeat probes, anomaly detection via deterministic recomputation, and automated trust management, creating a self-correcting feedback loop that isolates malicious or unreliable participants. Experiments on a prototype implementation demonstrate 100% detection rate and 0% false positive rate for tampered records, sub-millisecond validation latency for record and hub operations, and effective harness convergence that excludes adversarial nodes within two rounds while promoting honest nodes to trusted status within five rounds.

cs.DC↗

arXiv2026

Predicting oscillations in complex networks with delayed feedback

Oscillatory dynamics are common features of complex networks, often playing essential roles in regulating function. Across scales from gene regulatory networks to ecosystems, delayed feedback mechanisms are key drivers of system-scale oscillations. The analysis and prediction of such dynamics are highly challenging, however, due to the combination of high-dimensionality, non-linearity and delay. Here, we systematically investigate how structural complexity and delayed feedback jointly induce oscillatory dynamics in complex systems, and introduce an analytic framework comprising theoretical dimension reduction and data-driven prediction. We reveal that oscillations emerge from the interplay of structural complexity and delay, with reduced models uncovering their critical thresholds and showing that greater connectivity lowers the delay required for their onset. Our theory is empirically tested in an experiment on a programmable electronic circuit, where oscillations are observed once structural complexity and feedback delay exceeded the critical thresholds predicted by our theory. Finally, we deploy a reservoir computing pipeline to accurately predict the onset of oscillations directly from timeseries data. Our findings deepen understanding of oscillatory regulation and offer new avenues for predicting dynamics in complex networks.

cond-mat.dis-nn↗

arXiv2025

PDE-Agent: A toolchain-augmented multi-agent framework for PDE solving

Solving Partial Differential Equations (PDEs) is a cornerstone of engineering and scientific research. Traditional methods for PDE solving are cumbersome, relying on manual setup and domain expertise. While Physics-Informed Neural Network (PINNs) introduced end-to-end neural network-based solutions, and frameworks like DeepXDE further enhanced automation, these approaches still depend on expert knowledge and lack full autonomy. In this work, we frame PDE solving as tool invocation via LLM-driven agents and introduce PDE-Agent, the first toolchain-augmented multi-agent collaboration framework, inheriting the reasoning capacity of LLMs and the controllability of external tools and enabling automated PDE solving from natural language descriptions. PDE-Agent leverages the strengths of multi-agent and multi-tool collaboration through two key innovations: (1) A Prog-Act framework with graph memory for multi-agent collaboration, which enables effective dynamic planning and error correction via dual-loop mechanisms (localized fixes and global revisions). (2) A Resource-Pool integrated with a tool-parameter separation mechanism for multi-tool collaboration. This centralizes the management of runtime artifacts and resolves inter-tool dependency gaps in existing frameworks. To validate and evaluate this new paradigm for PDE solving , we develop PDE-Bench, a multi-type PDE Benchmark for agent-based tool collaborative solving, and propose multi-level metrics for assessing tool coordination. Evaluations verify that PDE-Agent exhibits superior applicability and performance in complex multi-step, cross-step dependent tasks. This new paradigm of toolchain-augmented multi-agent PDE solving will further advance future developments in automated scientific computing. Our source code and dataset will be made publicly available.

cs.AI↗

arXiv2025

Textual and Visual Guided Task Adaptation for Source-Free Cross-Domain Few-Shot Segmentation

Few-Shot Segmentation(FSS) aims to efficient segmentation of new objects with few labeled samples. However, its performance significantly degrades when domain discrepancies exist between training and deployment. Cross-Domain Few-Shot Segmentation(CD-FSS) is proposed to mitigate such performance degradation. Current CD-FSS methods primarily sought to develop segmentation models on a source domain capable of cross-domain generalization. However, driven by escalating concerns over data privacy and the imperative to minimize data transfer and training expenses, the development of source-free CD-FSS approaches has become essential. In this work, we propose a source-free CD-FSS method that leverages both textual and visual information to facilitate target domain task adaptation without requiring source domain data. Specifically, we first append Task-Specific Attention Adapters (TSAA) to the feature pyramid of a pretrained backbone, which adapt multi-level features extracted from the shared pre-trained backbone to the target task. Then, the parameters of the TSAA are trained through a Visual-Visual Embedding Alignment (VVEA) module and a Text-Visual Embedding Alignment (TVEA) module. The VVEA module utilizes global-local visual features to align image features across different views, while the TVEA module leverages textual priors from pre-aligned multi-modal features (e.g., from CLIP) to guide cross-modal adaptation. By combining the outputs of these modules through dense comparison operations and subsequent fusion via skip connections, our method produces refined prediction masks. Under both 1-shot and 5-shot settings, the proposed approach achieves average segmentation accuracy improvements of 2.18\% and 4.11\%, respectively, across four cross-domain datasets, significantly outperforming state-of-the-art CD-FSS methods. Code are available at https://github.com/ljm198134/TVGTANet.

cs.CV↗

arXiv2023

Few-shot Medical Image Segmentation via Cross-Reference Transformer

Deep learning models have become the mainstream method for medical image segmentation, but they require a large manually labeled dataset for training and are difficult to extend to unseen categories. Few-shot segmentation(FSS) has the potential to address these challenges by learning new categories from a small number of labeled samples. The majority of the current methods employ a prototype learning architecture, which involves expanding support prototype vectors and concatenating them with query features to conduct conditional segmentation. However, such framework potentially focuses more on query features while may neglect the correlation between support and query features. In this paper, we propose a novel self-supervised few shot medical image segmentation network with Cross-Reference Transformer, which addresses the lack of interaction between the support image and the query image. We first enhance the correlation features between the support set image and the query image using a bidirectional cross-attention module. Then, we employ a cross-reference mechanism to mine and enhance the similar parts of support features and query features in high-dimensional channels. Experimental results show that the proposed model achieves good results on both CT dataset and MRI dataset.

eess.IV↗

arXiv2020

Liutex-based Modified Navier-Stokes Equation

The Navier-Stokes (NS) partial differential equations, as the governing equation of fluid dynamics, are based on particles with zero volume. In such a condition, conservation law of moment of momentum is automatically satisfied, and thus NS equations only contain conservation of mass, momentum and energy. Because of the difficulty to get an analytical solution of NS equations, scientists develop finite element method (FEM) and finite volume method (FVM) to obtain approximate solutions. However, these methods are dependent on the size of finite volumes, which is not zero. Considering the size of the finite volume, conservation law of moment of momentum is no longer automatically satisfied and it is necessary to develop a new control equation to take effect of momentum into account. In this paper, new relations of reciprocal shear stresses are derived from conservation law of moment of momentum, and so are new constitutive relations and modified NS equations. The Liutex based NS equation may be the universal governing equations for both laminar and turbulent flow which is dominated by vortices and conservation of moment of momentum must be satisfied.

physics.flu-dyn↗

arXiv2020

Neural Spatio-Temporal Beamformer for Target Speech Separation

Purely neural network (NN) based speech separation and enhancement methods, although can achieve good objective scores, inevitably cause nonlinear speech distortions that are harmful for the automatic speech recognition (ASR). On the other hand, the minimum variance distortionless response (MVDR) beamformer with NN-predicted masks, although can significantly reduce speech distortions, has limited noise reduction capability. In this paper, we propose a multi-tap MVDR beamformer with complex-valued masks for speech separation and enhancement. Compared to the state-of-the-art NN-mask based MVDR beamformer, the multi-tap MVDR beamformer exploits the inter-frame correlation in addition to the inter-microphone correlation that is already utilized in prior arts. Further improvements include the replacement of the real-valued masks with the complex-valued masks and the joint training of the complex-mask NN. The evaluation on our multi-modal multi-channel target speech separation and enhancement platform demonstrates that our proposed multi-tap MVDR beamformer improves both the ASR accuracy and the perceptual speech quality against prior arts.

eess.AS↗

arXiv2019

Letter: Modified normalized Rortex/vortex identification method

In this paper, a modified normalized Rortex/vortex identification method named Omega_R is presented to improve the original Omega_R method and cure the bulging phenomenon on the iso-surfaces caused by the original Omega_R method. Detailed mathematical explanations and the relationship with the Q criterion are described. In addition, the new developed formula does not require two original coordinate rotations, and the calculation is greatly simplified. The numerical results are demonstrated to show the effectiveness of the new modified normalized Rortex/vortex identification method.

physics.flu-dyn↗

arXiv2018

A novel Hermite RBF-based differential quadrature method for solving two-dimensional variable-order time fractional advection-diffusion equation with Neumann boundary condition

In this paper, a novel Hermite radial basis function-based differential quadrature method (H-RBF-DQ) is presented. This new method is designed to treat derivative boundary conditions accurately. The developed method is very different from the original Hermite RBF method. In order to illustrate the specific process of this method, although the method can be used to study most of partial differential equations, the numerical simulation of two-dimensional variable-order time fractional advection-diffusion equation is chosen as an example. For the general case of irregular geometry, the meshless local form of RBF-DQ was used and the multiquadric type of radial basis functions are selected for the computations. The method is validated by the documented test examples involving variable-order fractional modeling of air pollution. The numerical results demonstrate the robustness and the versatility of the proposed approach.

physics.comp-ph↗

arXiv2016

Proportionate Affine Projection Algorithms for Block-sparse System Identification

A new family of block-sparse proportionate affine projection algorithms (BS-PAPA) is proposed to improve the performance for block-sparse systems. This is motivated by the recent block-sparse proportionate normalized least mean square (BS-PNLMS) algorithm. It is demonstrated that the affine projection algorithm (APA), proportionate APA (PAPA), BS-PNLMS and PNLMS are all special cases of the proposed BS-PAPA algorithm. Meanwhile, an efficient implementation of the proposed BS-PAPA and block-sparse memory PAPA (BS-MPAPA) are also presented to reduce computational complexity. Simulation results demonstrate that the proposed BS-PAPA and BS-MPAPA algorithms outperform the APA, PAPA and MPAPA algorithms for block-sparse system identification in terms of both faster convergence speed and better tracking ability.

cs.IT↗

arXiv2015

Proportionate Adaptive Filtering for Block Sparse System Identification

In this paper, a new family of proportionate normalized least mean square (PNLMS) adaptive algorithms that improve the performance of identifying block-sparse systems is proposed. The main proposed algorithm, called block-sparse PNLMS (BS-PNLMS), is based on the optimization of a mixed l2,1 norm of the adaptive filter coefficients. It is demonstrated that both the NLMS and the traditional PNLMS are special cases of BS-PNLMS. Meanwhile, a block-sparse improved PNLMS (BS-IPNLMS) is also derived for both sparse and dispersive impulse responses. Simulation results demonstrate that the proposed BS-PNLMS and BS-IPNLMS algorithms outperformed the NLMS, PNLMS and IPNLMS algorithms with only a modest increase in computational complexity.

cs.IT↗

arXiv2015

Block Sparse Memory Improved Proportionate Affine Projection Sign Algorithm

A block sparse memory improved proportionate affine projection sign algorithm (BS-MIP-APSA) is proposed for block sparse system identification under impulsive noise. The new BS-MIP-APSA not only inherits the performance improvement for block-sparse system identification, but also achieves robustness to impulsive noise and the efficiency of the memory improved proportionate affine projection sign algorithm (MIP-APSA). Simulations indicate that it can provide both faster convergence rate and better tracking ability under impulsive interference for block sparse system identification as compared to APSA and MIP-APSA.

eess.SY↗

arXiv2015

An Improved Variable Step-size Zero-point Attracting Projection Algorithm

This paper proposes an improved variable step-size (VSS) scheme for zero-point attracting projection (ZAP) algorithm. The proposed VSS is proportional to the sparseness difference between filter coefficients and the true impulse response. Meanwhile, it works for both sparse and non-sparse system identification, and simulation results demonstrate that the proposed algorithm could provide both faster convergence rate and better tracking ability than previous ones.

cs.OH↗

arXiv2014

The application of spectrum standardization method for carbon analysis in coal using laser-induced breakdown spectroscopy

Measurements of carbon content in coal using laser-induced breakdown spectroscopy (LIBS) is limited by its low measurement precision and accuracy. A spectrum standardization method was proposed to achieve both reproducible and accurate results for the quantitative analysis of carbon content in coal with LIBS. The proposed method utilized the molecular carbon emissions to compensate the diminution of atomic carbon emission caused by matrix effect. The compensated carbon line intensities were further converted into an assumed standard state with fixed plasma temperature, electron density, and total number density of elemental carbon, which is proportional to its concentration in the coal samples. In addition, in order to obtained better compensation for total carbon number density fluctuations, an iterative algorithm was applied, which is different from our previous standardization calculations. The modified spectrum standardization model was applied to the measurement of carbon content in 24 bituminous coal samples. The results demonstrated that the proposed method had superior performance over the generally applied normalization methods. The average relative standard deviation, the coefficient of determination, the root-mean-square error of prediction, and the average maximum relative error for the modified model were 3.44%, 0.83, 2.71%, and 12.61%, respectively, while the corresponding values for the normalization with segmental spectrum area were 6.00%, 0.75, 3.77%, and 15.40%, respectively, showing an overwhelming improvement.

physics.optics↗

arXiv2013

A New Variable Step-size Zero-point Attracting Projection Algorithm

This paper proposes a new variable step-size (VSS) scheme for the recently introduced zero-point attracting projection (ZAP) algorithm. The proposed variable step-size ZAPs are based on the gradient of the estimated filter coefficients sparseness that is approximated by the difference between the sparseness measure of current filter coefficients and an averaged sparseness measure. Simulation results demonstrate that the proposed approach provides both faster convergence rate and better tracking ability than previous ones.

cs.OH↗

arXiv2013

A Three-class ROC for Evaluating Doubletalk Detectors in Acoustic Echo Cancellation

Doubletalk detector (DTD) is essential to keep adaptive filter from diverging in the presence of near-end speech in acoustic echo cancellation (AEC), and there was a receiver operating characteristic (ROC) to characterize DTD performance. However, the traditional ROC for evaluating DTD used a static time-invariant room acoustic impulse response and could not evaluate DTDs which distinguish echo path change from doubletalk. We solve these problems by extending the traditional binary detection ROC to three class, and simulations show the efficiency of the proposed method.

cs.IT↗

arXiv2013

An Improved Variable Step-size Affine Projection Sign Algorithm for Echo Cancellation

This paper proposes an improved variable step-size (VSS) algorithm for the recently introduced affine projection sign algorithm (APSA) based on the recovery of the near-end signal energy in the error signal. Simulation results demonstrate that, compared to the previous VSS for APSA, the proposed approach provides both more robustness to impulse interference and better tracking ability of echo path change.

cs.OH↗