arXiv subjects

Di Cao

Publications and source records attributed to Di Cao.

14 recordsLinked to original sources

arXiv2026

Omni Interaction Agent Technical Report

In this work, we present Gander, an end-to-end model that unifies omni perception, realtime interaction, and agentic capabilities within a single framework. In contrast to turn-based conventional paradigms, Gander continuously receives streaming inputs across multiple modalities, including video, speech, and text, enabling natural full-duplex interaction in both everyday conversations and complex workflow-oriented agent scenarios. Users can interrupt the model at any time, while the model can also proactively provide intermediate feedback or ask follow up questions. To natively support these capabilities, Gander adopts two key architectural designs: 1) It employs a Cerebellum-Brain collaborative framework, in which the Cerebellum is responsible for realtime interaction and omni conversational capabilities, while the Brain handles complex reasoning and higher-level agentic tasks. The two components interact continuously through tool calling and the agent orchestration runtime. 2) The Cerebellum is built upon a streaming Thinker-Talker architecture, user inputs and model outputs are further flattened into an ordered token stream at the chunk level, providing a unified representation for low latency, continuous interaction. We conduct comprehensive evaluations of Gander across four dimensions: conversational ability, omni understanding, interactive capability, and agentic intelligence. Internal human evaluations demonstrate that Gander maintains the natural and expressive spoken dialogue capabilities of SOTA open source models while achieving competitive performance in omni interaction. Gander also demonstrates robustness in challenging real-world scenarios, including background noise interference, multi-party interactions, and backchannel communication. We release Gander together with its models, code, and data to facilitate further research and development in the community.

eess.AS↗

arXiv2026

X$^3$-OPD: Distilling Reasoning into Large Audio-Language Models via On-Policy Alignment

While large audio-language models have achieved remarkable progress in auditory perception, they still lag behind text-based large language models in deep logical reasoning, primarily due to the scarcity of high-quality audio reasoning data. To bridge this gap, we propose X$^3$-OPD, a cross-modal on-policy distillation framework that transfers reasoning capabilities from a powerful text teacher to an audio-language student. During training, the student generates reasoning trajectories conditioned on its own acoustic perception, while the teacher provides token-level guidance using matched textual inputs and verified answers. We further construct a three-tier symmetric corpus covering textual reasoning rendered into speech, audio-event reasoning grounded in complex acoustic scenes, and spoken-dialogue reasoning involving paralinguistic cues. This design extends cross-modal distillation beyond textually recoverable content to reasoning grounded in non-linguistic events, prosody, and conversational context. Experiments on MMSU, MMAU, BIG Bench Audio, and MMAR demonstrate that X$^3$-OPD substantially improves audio-grounded reasoning and chain-of-thought quality while largely preserving the model's existing capabilities under domain shift.

cs.LG↗

arXiv2026

X-OPD: Cross-Modal On-Policy Distillation for Capability Alignment in Speech LLMs

While the shift from cascaded dialogue systems to end-to-end (E2E) speech Large Language Models (LLMs) improves latency and paralinguistic modeling, E2E models often exhibit a significant performance degradation compared to their text-based counterparts. The standard Supervised Fine-Tuning (SFT) and Reinforcement Learning (RL) training methods fail to close this gap. To address this, we propose X-OPD, a novel Cross-Modal On-Policy Distillation framework designed to systematically align the capabilities of Speech LLMs to their text-based counterparts. X-OPD enables the Speech LLM to explore its own distribution via on-policy rollouts, where a text-based teacher model evaluates these trajectories and provides token-level feedback, effectively distilling teacher's capabilities into student's multi-modal representations. Extensive experiments across multiple benchmarks demonstrate that X-OPD significantly narrows the gap in complex tasks while preserving the model's inherent capabilities.

eess.AS↗

arXiv2026

Alteration of the Brains Microbiome and Neuroinflammation Associated with Ventricular Catheters

Background and Objectives: Proximal catheter obstruction is the leading cause of ventriculoperitoneal shunt failure, yet the biological triggers of peri-catheter inflammation and tissue ingrowth remain poorly defined. Evidence of bacterial ribosomal RNA in human brain tissue suggests that low-biomass microbial exposure may influence the inflammatory microenvironment surrounding implants. This study examined if microbial signal is detectable in unaltered brain tissue and if catheter implantation produces microbial shifts relevant to shunt dysfunction. Methods: Twenty-nine female mice were assigned to unaltered control (UC), trauma control (TC), plain silicone catheter (PSC), or antibiotic-impregnated catheter (AIC) groups. Brain and cecum tissues were harvested at postoperative days 7 and 28 for 16S rRNA sequencing. Microbial composition and predicted functional pathways were analyzed. A separate cohort underwent longitudinal MRI to assess edema, glial scar formation, and macrophage-associated susceptibility signal. Results: Low-level microbial signal was detected in unaltered brain tissue. Catheter implantation induced material-dependent shifts in brain-associated microbial composition. PSC was associated with enrichment of pro-inflammatory taxa, whereas AIC favored immune-regulatory taxa. Predicted short-chain fatty acid biosynthesis was highest in AIC and lowest in PSC, while predicted lipopolysaccharide biosynthesis trended higher in PSC. MRI showed similar edema resolution but higher macrophage-associated susceptibility signal in PSC animals. Conclusion: Intracranial catheter implantation produces material-dependent shifts in low-biomass brain-associated microbial signal that parallel differential neuroimmune activation. These findings suggest catheter material may shape a biologically relevant peri-catheter niche with implications for chronic gliosis and proximal shunt obstruction.

q-bio.GN↗

arXiv2026

Source-Free Domain Adaptation for Geospatial Point Cloud Semantic Segmentation

Semantic segmentation of 3D geospatial point clouds is fundamental to remote sensing applications, yet domain shifts caused by regional and acquisition-related variations often degrade model performance. Although domain adaptation can mitigate such shifts, existing methods typically require access to source-domain data, which is often infeasible due to privacy concerns and regulatory policies. To address this, we propose LoGo (Local-Global Dual-Consensus), a novel source-free unsupervised domain adaptation (SFUDA) framework requiring only a pretrained model and unlabeled target data. At the local level, we introduce a class-balanced prototype estimation module that ensures that robust feature prototypes can be generated even for sample-scarce tail classes, effectively mitigating the feature collapse caused by long-tailed distributions. At the global level, we introduce an optimal transport-based global distribution alignment module that formulates pseudo-label assignment as a global optimization problem, effectively correcting the over-dominance of head classes inherent in local greedy assignments, and thereby preventing model predictions from being severely biased towards majority classes. Finally, we propose a dual-consistency pseudo-label filtering mechanism that retains only high-confidence pseudo-labels where local multi-augmented ensemble predictions align with global optimal transport assignments for self-training. Extensive experiments on two challenging benchmarks, encompassing cross-scene and cross-sensor settings, demonstrate that LoGo consistently outperforms existing state-of-the-art methods. The source code is available at https://github.com/GYproject/LoGo-SFUDA.

cs.CV↗

arXiv2025

A Graph Neural Network with Auxiliary Task Learning for Missing PMU Data Reconstruction

In wide-area measurement systems (WAMS), phasor measurement unit (PMU) measurement is prone to data missingness due to hardware failures, communication delays, and cyber-attacks. Existing data-driven methods are limited by inadaptability to concept drift in power systems, poor robustness under high missing rates, and reliance on the unrealistic assumption of full system observability. Thus, this paper proposes an auxiliary task learning (ATL) method for reconstructing missing PMU data. First, a K-hop graph neural network (GNN) is proposed to enable direct learning on the subgraph consisting of PMU nodes, overcoming the limitation of the incompletely observable system. Then, an auxiliary learning framework consisting of two complementary graph networks is designed for accurate reconstruction: a spatial-temporal GNN extracts spatial-temporal dependencies from PMU data to reconstruct missing values, and another auxiliary GNN utilizes the low-rank property of PMU data to achieve unsupervised online learning. In this way, the low-rank properties of the PMU data are dynamically leveraged across the architecture to ensure robustness and self-adaptation. Numerical results demonstrate the superior offline and online performance of the proposed method under high missing rates and incomplete observability.

eess.SY↗

arXiv2024

RealVul: Can We Detect Vulnerabilities in Web Applications with LLM?

The latest advancements in large language models (LLMs) have sparked interest in their potential for software vulnerability detection. However, there is currently a lack of research specifically focused on vulnerabilities in the PHP language, and challenges in extracting samples and processing persist, hindering the model's ability to effectively capture the characteristics of specific vulnerabilities. In this paper, we present RealVul, the first LLM-based framework designed for PHP vulnerability detection, addressing these issues. By vulnerability candidate detection methods and employing techniques such as normalization, we can isolate potential vulnerability triggers while streamlining the code and eliminating unnecessary semantic information, enabling the model to better understand and learn from the generated vulnerability samples. We also address the issue of insufficient PHP vulnerability samples by improving data synthesis methods. To evaluate RealVul's performance, we conduct an extensive analysis using five distinct code LLMs on vulnerability data from 180 PHP projects. The results demonstrate a significant improvement in both effectiveness and generalization compared to existing methods, effectively boosting the vulnerability detection capabilities of these models.

cs.CR↗

arXiv2024

EaTVul: ChatGPT-based Evasion Attack Against Software Vulnerability Detection

Recently, deep learning has demonstrated promising results in enhancing the accuracy of vulnerability detection and identifying vulnerabilities in software. However, these techniques are still vulnerable to attacks. Adversarial examples can exploit vulnerabilities within deep neural networks, posing a significant threat to system security. This study showcases the susceptibility of deep learning models to adversarial attacks, which can achieve 100% attack success rate (refer to Table 5). The proposed method, EaTVul, encompasses six stages: identification of important samples using support vector machines, identification of important features using the attention mechanism, generation of adversarial data based on these features using ChatGPT, preparation of an adversarial attack pool, selection of seed data using a fuzzy genetic algorithm, and the execution of an evasion attack. Extensive experiments demonstrate the effectiveness of EaTVul, achieving an attack success rate of more than 83% when the snippet size is greater than 2. Furthermore, in most cases with a snippet size of 4, EaTVul achieves a 100% attack success rate. The findings of this research emphasize the necessity of robust defenses against adversarial attacks in software vulnerability detection.

cs.CR↗

arXiv2024

LightCAM: A Fast and Light Implementation of Context-Aware Masking based D-TDNN for Speaker Verification

Traditional Time Delay Neural Networks (TDNN) have achieved state-of-the-art performance at the cost of high computational complexity and slower inference speed, making them difficult to implement in an industrial environment. The Densely Connected Time Delay Neural Network (D-TDNN) with Context Aware Masking (CAM) module has proven to be an efficient structure to reduce complexity while maintaining system performance. In this paper, we propose a fast and lightweight model, LightCAM, which further adopts a depthwise separable convolution module (DSM) and uses multi-scale feature aggregation (MFA) for feature fusion at different levels. Extensive experiments are conducted on VoxCeleb dataset, the comparative results show that it has achieved an EER of 0.83 and MinDCF of 0.0891 in VoxCeleb1-O, which outperforms the other mainstream speaker verification methods. In addition, complexity analysis further demonstrates that the proposed architecture has lower computational cost and faster inference speed.

cs.CL↗

arXiv2022

Optimal Estimation with Sensor Delay

Given a plant subject to delayed sensor measurement, there are several approaches to compensate for the delay. An obvious approach is to address this problem in state space, where the $n$-dimensional plant state is augmented by an $N$-dimensional (Pad\'e) approximation to the delay, affording (optimal) state estimate feedback vis-\`a-vis the separation principle. Using this framework, we show: (1) Feedback of the estimated plant states partially inverts the delay; (2) The optimal (Kalman) estimator decomposes into $N$ (Pad\'e) uncontrollable states, and the remaining $n$ eigenvalues are the solution to a reduced-order Kalman filter problem. Further, we show that the tradeoff of estimation error (of the full state estimator) between plant disturbance and measurement noise, only depends on the reduced-order Kalman filter (that can be constructed independently of the delay); (3) A subtly modified version of this state-estimation-based control scheme bears close resemblance to a Smith predictor. This modified state-space approach shares several limitations with its Smith predictor analog (including the inability to stabilize most unstable plants), limitations that are alleviated when using the unmodified state estimation framework.

math.OC↗

arXiv2020

Model-Free Voltage Regulation of Unbalanced Distribution Network Based on Surrogate Model and Deep Reinforcement Learning

Accurate knowledge of the distribution system topology and parameters is required to achieve good voltage controls, but this is difficult to obtain in practice. This paper develops a model-free approach based on the surrogate model and deep reinforcement learning (DRL). We have also extended it to deal with unbalanced three-phase scenarios. The key idea is to learn a surrogate model to capture the relationship between the power injections and voltage fluctuation of each node from historical data instead of using the original inaccurate model affected by errors and uncertainties. This allows us to integrate the DRL with the learned surrogate model. In particular, DRL is applied to learn the optimal control strategy from the experiences obtained by continuous interactions with the surrogate model. The integrated framework contains training three networks, i.e., surrogate model, actor, and critic networks, which fully leverage the strong nonlinear fitting ability of deep learning and DRL for online decision making. Several single-phase approaches have also been extended to deal with three-phase unbalance scenarios and the simulation results on the IEEE 123-bus system show that our proposed method can achieve similar performance as those that use accurate physical models.

eess.SY↗

arXiv2020

Distributed Voltage Regulation of Active Distribution System Based on Enhanced Multi-agent Deep Reinforcement Learning

This paper proposes a data-driven distributed voltage control approach based on the spectrum clustering and the enhanced multi-agent deep reinforcement learning (MADRL) algorithm. Via the unsupervised clustering, the whole distribution system can be decomposed into several sub-networks according to the voltage and reactive power sensitivity. Then, the distributed control problem of each sub-network is modeled as Markov games and solved by the enhanced MADRL algorithm, where each sub-network is modeled as an adaptive agent. Deep neural networks are used in each agent to approximate the policy function and the action value function. All agents are centrally trained to learn the optimal coordinated voltage regulation strategy while executed in a distributed manner to make decisions based on only local information. The proposed method can significantly reduce the requirements of communications and knowledge of system parameters. It also effectively deals with uncertainties and can provide online coordinated control based on the latest local information. Comparison results with other existing model-based and data-driven methods on IEEE 33-bus and 123-bus systems demonstrate the effectiveness and benefits of the proposed approach.

eess.SY↗

arXiv2017

Microstructure and properties of Cu-Sn-Zn-TiO2 Nano-composite coatings on mild steel

Cu-Sn-Zn coatings have been widely used in industry for their unique properties, such as good conductivity, high corrosion resistance and excellent solderability. To further improve the mechanical performance of Cu-Sn-Zn coatings, powder-enhanced method was applied in the current study and Cu-Sn-Zn-TiO2 nano-composite coatings with different TiO2 concentration were fabricated. The microstructure of Cu-Sn-Zn-TiO2 nano-composite coatings were investigated by X-ray diffraction (XRD) and Scanning Electron Microscopy (SEM). The mechanical properties of coatings including microhardness and wear resistance were studied. The results indicate that the incorporation of TiO2 nanoparticle can significantly influence the properties of Cu-Sn-Zn coatings. The microhardness of Cu-Sn-Zn coating was increased to 383 HV from 330 HV with 1g/L TiO2 addition. Also, the corrosion resistance of coating was enhanced. The effects of TiO2 nanoparticle concentration on the microstructure, mechanical properties and corrosion resistance of Cu-Sn-Zn-TiO2 nano-composite coatings were discussed.

physics.app-ph↗

arXiv2012

Probing the structure of the outflow in the tidal disruption flare Sw J1644+57 with long-term radio emission

The recently discovered high-energy transient Sw J1644+57 is thought to arise from the tidal disruption of a passing star by a dormant massive black hole. The long-term, bright radio emission of Sw J1644+57 is believed to result from the synchrotron emission of the blast wave produced by an outflow expanding into the surrounding medium. Using the detailed multi-epoch radio spectral data, we are able to determine the total number of radiating electrons in the outflow at different times, and further the evolution of the cross section of the outflow with time. We find that the outflow gradually transits from a conical jet to a cylindrical one at later times. The transition may be due to collimation of the outflow by the pressure of the shocked jet cocoon that forms while the outflow is propagating in the ambient medium. Since cylindrical jets usually exist in AGNs and extragalactic jets, this may provide independent evidence that Sw J1644+57 signals the onset of an AGN.

astro-ph.HE↗